StyleObservationObs_CategoryMapping **Category Mapping**:\n- Each Discourse category = folder in repo\n- Each thread = subfolder\n- Each post = transaction in thread folder
StyleObservationObs_MCPPrimitives **Lower-Level Tools** (internal):\n- Atomic operations\n- Used by MCP tools and GitHub Actions\n- Examples: `snapshot/extract`, `tx/commit`, `story/draft`
StyleObservationObs_MemoryDefinition **Memory**: Ephemeral user content (chat message, Discourse post, uploaded doc)
StyleObservationObs_TwoStepSave 1. User ideates in chat with AI\n2. AI offers to post to Discourse for team visibility\n3. Discourse post includes `@aswritten` mention
StyleObservationObservation_AswrittenBrand I'm happy with the shift of the company name to
StyleObservationObservation_CompiledTarget content is a compiled target. The application is a compiled target
StyleObservationObservation_ContextSquishing you are going to get behind the scenes,
StyleObservationObservation_CrossFunctionalCascade As soon as a salesperson, you know, gains additional context about a client, the engineer immediately benefits in the actual hands-on coding and vice versa. When they make an architectural decision or implement the feature that now becomes
StyleObservationObservation_CultivationTrap Current state incentivizes **cultivating long-running personal AI instances**
StyleObservationObservation_DeepestLever the deepest lever we've ever had that reaches from CEO strategist every single, you know, strategic sales customer service role all the way into the core of the development process and back
StyleObservationObservation_ExistentialCrisis This is the existential crisis of modern business
StyleObservationObservation_GitNativeTriad git native, i.e. versionable, branchable, reviewable
StyleObservationObservation_IntentionalMemory One is to make memory intentional. So instead of this context squishing autosummarization, we have an agent or context that chooses to remember stuff, offers that to the user, and then saves that to the git repo intentionally
StyleObservationObservation_IPAsCode intellectual property as code
StyleObservationObservation_MultiplayerShift The movement is from instead of individuals talking to an AI where each individual is going to get a different answer for a request to execute a process that should be understood within the company domain [...] to a multiplayer collaborativ
StyleObservationObservation_NoNetworkProblem There is actually no network problem here. This is just as useful for one person, but it is even more useful for two people
StyleObservationObservation_StateSecrets every single employee in your company, CEO included, is telling state secrets to AI
StyleObservationRhetoricalQuestion Quoted direct speech from demo context used as evidence of UX failure — first-person embedded testimony
StyleObservationStyle_Antithesis_ChannelNotIdentity Antithesis: 'The product is not a phone number; it's the phone number distribution channel.' Scarlet uses direct negation followed by positive restatement — a rhetorical antithesis that sharpens the correction by holding both frames in a si
StyleObservationStyle_DirectCorrection Document opens with bold-formatted declarative correction: '**This is incorrect.**' and '**Scarlet has tried to correct this before.**' The use of bold for emphasis on correction statements, combined with past-tense accountability ('It keep
StyleObservationStyle_EpistemicPrecision Document consistently distinguishes epistemic states: 'leaning' vs 'decision', 'exploration' vs 'current worldview', 'not superseded' vs 'superseded'. Phrases like 'The leaning is recorded; it is not yet a decision' and 'This distinction mu
StyleObservationStyle_VerbatimQuotePattern Document structures argument through verbatim block-quotes from Scarlet, marked with '> \"...' convention. This pattern signals epistemic authority: the founder's exact words are the primary evidence, not paraphrase. The document author (Cl
StyleObservationStyleObs_1 the treatment of an organization as an AI organism, right? Whose DNA, if you will, is expressed
StyleObservationStyleObs_10 explicit, reviewable, and testable
StyleObservationStyleObs_11 Tricolon closing the conviction section: 'Foundation-level', 'Boulder', 'Boulder' — three conviction assessments in parallel structure, escalating uncertainty.
StyleObservationStyleObs_12 Code-switching between formal analytical register and informal conversational register (direct quotes from Scarlet in lowercase, no punctuation). Deliberate contrast.
StyleObservationStyleObs_2 just to do, you know,
StyleObservationStyleObs_3 recompiles everything and gets us, you know, closer and closer to where we want to go
StyleObservationStyleObs_4 They serve as telltales. That's the right word. They serve as telltales
StyleObservationStyleObs_5 The answer is no. The answer is that all of your, you know, an AI will always fill
StyleObservationStyleObs_6 where we have an organization
StyleObservationStyleObs_7 you can connect that to anything that you could automate with Zapier
StyleObservationStyleObs_9 The past 20-25 years optimized for human-readable documentation; AI-native paradigm optimizes for machine-queryable worldview
StyleObservationStyleObs_ActiveDecision We are building collective memory as a complete product - a collaborative worldview layer
StyleObservationStyleObs_ActiveVoice I'm able to steer models successfully
StyleObservationStyleObs_ActiveVoice_1 creating high quality output with AI requires extensive context
StyleObservationStyleObs_ActiveVoice1 This validates the MCP-agnostic architecture claim
StyleObservationStyleObs_ActiveVoiceRule Active voice required throughout. Cut modal-passive constructions: 'had to explain' -> 'explained', 'needed to figure out' -> 'figured out', 'tried to communicate' -> 'said'.
StyleObservationStyleObs_ADRStructure Standard ADR format: Context → Decision → Rationale → Consequences → Status
StyleObservationStyleObs_AgentSim_Antithesis Heavy antithesis structure throughout: 'What a collaborator said' vs. 'What doesn't transfer', 'procedural vs. human-driven', 'decay vs. depreciation', 'Park's flat importance vs. aswritten's provenance-carrying conviction'. The document is
StyleObservationStyleObs_AgentSim_BlockquoteRegister mixed register: structured markdown H2 sections with bold numbered parallels (memo genre) vs. colloquial first-person in blockquoted speech from a collaborator's talk transcript; filler-heavy conversational speech contrasts with tight analy
StyleObservationStyleObs_AgentSim_ProvisionalFlag The document opens with an explicit provisional flag in bold: 'Provisional — from research/agent-memory-architecture exploration branch, not yet promoted to main.' This is a deliberate safety rail (per Claim_ProvisionalFlagSafetyRail) that
StyleObservationStyleObs_AgentSim_VerbatimBlockquotes Extensive verbatim blockquotes from a collaborator's talk transcript, preserving her colloquial register ('um uh', 'kind of was surprising', 'stupid like I never missed closing the door'). The quoted speech is filler-heavy and conversationa
StyleObservationStyleObs_AhaMoment 'big aha moment' — colloquial product-discovery idiom; signals informal register and authentic voice in meeting notes.
StyleObservationStyleObs_AICompanyizedEmphasis All-caps used for semantic emphasis on key architectural constraints: WHOLE, TOWARD, AWAY. Signals non-negotiable design intent rather than stylistic preference.
StyleObservationStyleObs_AllCapsEmphasis ALL CAPS 'IS' used for maximum rhetorical emphasis on the identity claim. Appears three times across the document as a refrain.
StyleObservationStyleObs_AllCapsIS ALL-CAPS 'IS' used for emphatic assertion. Signals maximum conviction. Rhetorical device: typographic stress replacing spoken emphasis. Characteristic of the author's voice for high-conviction claims.
StyleObservationStyleObs_AltitudeTaxonomy Three altitudes of the same primitive
StyleObservationStyleObs_AmazingProductNotProduct It's both an amazing product and not a product at all
StyleObservationStyleObs_AmbientQuote You don't have to open a new tool. You're just using Claude Code... And in the middle of your current work, it's like, you should save that
StyleObservationStyleObs_Analogy_1 the code becomes the compilation output
StyleObservationStyleObs_Analogy_DuplicateBrain the constant joke in the enterprise customer, literally—I'm not kidding—is 'if only we could get AI to duplicate the customer's chief architect's brain.'
StyleObservationStyleObs_analogy_git collective memory creates a living organizational worldview that evolves through intentional memory-saving and branches like code
StyleObservationStyleObs_Analogy_Matrix 'I know Kung Fu' Matrix chips — but at the organizational level.
StyleObservationStyleObs_Analogy_ResearchPaper Citation analogy: 'like a research paper' makes the citation behavior concrete and familiar. Directly answers the hallucination hook by invoking academic accountability.
StyleObservationStyleObs_Analogy_Resonant AI identity as this sort of resonant body that has a shape to it
StyleObservationStyleObs_AnalogyBookVsModule Extended analogy: book (static snapshot) vs. module (alive, auto-updating). Rhetorical device used to crystallize the living module pitch for content creators.
StyleObservationStyleObs_AnalogyReadableTest a readable test framework becomes the critical human review interface
StyleObservationStyleObs_Anaphora every single home, but every single venture, every single backend, every surface
StyleObservationStyleObs_anaphora_all All automatically. All from the same source of truth. All version-controlled with Git.
StyleObservationStyleObs_Anaphora_OldNew Anaphora: OLD/NEW repeated structure for language shift section creates parallel contrast rhythm
StyleObservationStyleObs_Antithesis They solve memory for individuals. We solve memory for organizations. They create lock-in. We create freedom.
StyleObservationStyleObs_Antithesis_BDF Antithesis used to sharpen the bounded-domain argument: 'bounded' vs 'lossy', 'intentional' vs 'always doing it', 'higher fidelity' vs 'inherently lossy'. The rhetorical structure mirrors the strategy doc's antithesis-heavy style.
StyleObservationStyleObs_Antithesis_ConstellationVsSingle The document's central rhetorical move is antithesis: 'citations are constellations, not single claims' — a binary opposition that structures the entire argument. The title itself encodes the antithesis.
StyleObservationStyleObs_Antithesis_ContextVsDirection Antithesis: more context vs. direction — the sharpest competitive one-liner in the document
StyleObservationStyleObs_Antithesis_MCP JSON was 1.94x the token cost with no information gain. LLMs parse TriG natively.
StyleObservationStyleObs_Antithesis_MechanismVsPain Explicit antithesis in the meta-analysis: mechanism (knowledge retention) vs. pain (hallucination). Structural contrast drives the pitch redesign decision.
StyleObservationStyleObs_antithesis_not_is **NOT**: Three separate AI tools for three personas\n**IS**: One worldview that unifies how AI assists everyone
StyleObservationStyleObs_Antithesis_OrgVsPersonal Structural antithesis: 'what the org should do' vs 'how I can most effectively do it as a person' — the core tension that drives the entire document.
StyleObservationStyleObs_Antithesis_SmokesVsBattery Antithesis-heavy structure throughout: smokes vs. battery, mechanism vs. quality, appearance of being finished vs. the control mattering more. The core principle is stated as a paired antithesis: 'Smokes prove the mechanism; the battery pro
StyleObservationStyleObs_Antithesis_SponsorChampion Antithesis-heavy structure throughout: sponsor vs. champion, excited vs. rams it through, POC sequential vs. parallel, Team credit card vs. Enterprise MSA. The document's core argument is built on a series of binary contrasts that sharpen t
StyleObservationStyleObs_Antithesis_TriG Subgraph format is TriG, not structured JSON.
StyleObservationStyleObs_Antithesis_V0 The visitor doesn't watch a demo; they touch the product immediately.
StyleObservationStyleObs_Antithesis_VC bootstrapping but that are not lifestyle businesses are actually like very ambitious long term businesses
StyleObservationStyleObs_Antithesis_WarmCold The memory is structured around a series of antitheses: warm-start vs cold-start, async vs synchronous, GitHub vs no-GitHub, single user vs team, main vs branch. This antithesis-heavy structure is characteristic of Scarlet's design thinking
StyleObservationStyleObs_AntithesisBeforeAfter_Distillation The memory uses consistent before/after antithesis structure for every rename decision: old term → new term. This is the dominant rhetorical device throughout.
StyleObservationStyleObs_AntithesisConvictionTemporality Sustained antithesis between conviction (how settled) and temporality (when applicable) — the structural spine of the entire document. Pairs like 'canon-in-conviction but canon-soon in temporality' and 'decision-level conviction / roadmap-l
StyleObservationStyleObs_AntithesisNotBuilding achieved by not building
StyleObservationStyleObs_AntithesisOverplayed_20260521 antithesis structure: 'overplayed / not an explicit anti-pattern' vs prior claim; deliberate vs accidental framing
StyleObservationStyleObs_AntithesisReplaces Repeated antithesis pattern: 'What it replaces' → corrected to 'What it changes'; 'experts are bottlenecks' → corrected to 'experts are sources'. The document uses antithesis to mark strategic reframes — the old framing is named before the
StyleObservationStyleObs_AntithesisRhetoric antithesis-as-reframe; the deli story pivots Vouch's positioning through explicit antithesis: 'away from metaphors of power, control, and security towards respect, comfort, and invitation' — a three-beat antithetical structure that performs
StyleObservationStyleObs_AntithesisStructure Consistent antithetical pairing throughout: 'perspective / implementation', 'why / what', 'intent / value', 'reasoning / code'. The document's rhetorical structure is built on binary opposition that clarifies the principle by contrast.
StyleObservationStyleObs_AntithesisTransformQuery antithesis-heavy structure: Transform > Query > Directory; hard segmentation vs. soft focus; actor vs. role; conviction vs. temporality. Each section is structured as a binary opposition that resolves into a design decision.
StyleObservationStyleObs_AntithesisTriggers Heavy antithesis structure throughout: 'Understanding the product value does not equal a sale.' / 'Information isn't enough... You need perspective.' / 'You're installing YOUR perspective so YOUR AI can do your work' vs. 'You're installing
StyleObservationStyleObs_AntithesisUserToPerspective The development inversion section uses explicit antithesis structure: 'currently you're going to Claude and say, do this' vs. 'you could be sharing your perspective and it comes back to you and says, hey, I made something.' This before/afte
StyleObservationStyleObs_AntithesisVentureVsCalm Antithesis-heavy structure throughout: venture-scale goal vs. calm-company execution, 'valuable enough' vs. 'pressure on the acquirer', 'shop the company' vs. 'being legible and available', 'needs to sell' vs. 'doesn't need to sell'. Short
StyleObservationStyleObs_AppleComparison Recurring Apple vs. Microsoft metaphor used to distinguish opinionated simplicity from feature sprawl
StyleObservationStyleObs_ArchetypeNaming Named archetypes (a beta user from outside tech, the customer's chief architect) used as shorthand throughout — not just as examples but as canonical type references. 'a beta user from outside tech' becomes the solo professional archetype;
StyleObservationStyleObs_ArchetypeTesting Systematic testing of a new principle against a list of established archetypes to prove universality.
StyleObservationStyleObs_ArrowCadence Arrow notation (→) used as a compressed process cadence marker. Three-step pipeline rendered as visual rhythm: interviews → memory → server.
StyleObservationStyleObs_ArrowChainFlow Unicode arrow (→) chains product flow steps in a single sentence — compresses multi-step process into scannable linear sequence without numbered lists.
StyleObservationStyleObs_ArrowChaining Arrow notation (→) used to chain sequential operations, creating a visual pipeline cadence. Appears in the review flow steps, tool output descriptions, and the lens execution sequence.
StyleObservationStyleObs_ArrowFlowNotation Arrow notation (→) used to represent sequential product flows and UI transitions. Compresses multi-step processes into scannable linear chains. Technical/engineering register.
StyleObservationStyleObs_ArrowNotation Em-dash arrow (→) used as a process-flow shorthand in the Salon capability description: 'group interview → consensus → merge'. Compact technical cadence.
StyleObservationStyleObs_ArrowOperator → (arrow) used consistently as a flow/yields operator in both prose and lists: 'refract(document, tx) → focus areas'
StyleObservationStyleObs_AskItSomething Imperative micro-script embedded in process doc; two-word call to action that doubles as the product's value proof
StyleObservationStyleObs_AsWrittenBrandName 'as written' is consistently rendered in lowercase with no capitalization, a deliberate brand stylization choice.
StyleObservationStyleObs_AswrittenLowercase Brand name stylization: lowercase 'aswritten' consistently used throughout source.
StyleObservationStyleObs_AswrittenModeMarker ', aswritten' as a lowercase suffix appended to prompts — a mode-switch marker that is simultaneously a brand name, a command, and a rhetorical device. Lowercase consistent with brand stylization.
StyleObservationStyleObs_AudienceSegmentation Explicit audience bifurcation: 'vision-oriented CEOs' vs 'ops leads and technical buyers'. Signals audience-dependent messaging strategy. Structural segmentation marker.
StyleObservationStyleObs_BacktickCodeStyle Consistent use of backtick code spans for workflow names, parameters, and node names throughout. Signals technical precision and inline documentation style.
StyleObservationStyleObs_BacktickTerms Backtick code spans used for task IDs, file names, workflow names, and SQL identifiers — enforces precision and scannability in technical prose.
StyleObservationStyleObs_BacktickToolNames Tool names consistently wrapped in backticks: `scope`, `refract`, `lens`, `worldview`, `introspect` — code-style naming in prose
StyleObservationStyleObs_BeliefSpecificsBeliefCadence Governing cadence pattern: belief, specifics, belief. Contrast emerges from walking the reader through what we believe and what actually happens, not from a closing contrast clause.
StyleObservationStyleObs_BetterAorB The system needs an equivalent to the eye chart with progressively smaller lines — micro-telltales at decreasing granularity
StyleObservationStyleObs_BetterThroughAOrB Optometrist idiom used as a rhetorical device to frame convergence testing: 'Better through A or B?' encapsulates the entire lens comparison mechanism.
StyleObservationStyleObs_BizarreIndustryIdiom Colloquial hedging idiom; a founder beta user's informal spoken register preserved verbatim in call notes.
StyleObservationStyleObs_BlockquoteCorrections Scarlet Dame's corrections are consistently set in blockquotes (> \"...\"), distinguishing authoritative source voice from analytical commentary. Creates a two-register document.
StyleObservationStyleObs_BlockquoteForScript Blockquote used to present the scripted 'remember' handoff line, visually separating it as a deliverable speech act rather than analysis. Signals this is ready-to-use language.
StyleObservationStyleObs_BlockquoteVoice Founder voice injected via Markdown blockquote (>) to distinguish decision-maker speech from narrative prose. Creates authority and intimacy simultaneously.
StyleObservationStyleObs_BoldBeats Five-beat arc labeled with bold typographic markers (The fear, The process, The proof, The loop, The payoff) — structural scaffolding that makes the pitch architecture explicit in the analysis section.
StyleObservationStyleObs_BoldBulletLeadins Consistent use of bold markdown for section headers and bullet lead-ins; creates visual hierarchy and scannable structure throughout the document.
StyleObservationStyleObs_BoldConviction Bold markdown used to mark irreversible architectural facts and pivotal decisions — signals conviction level without hedging language.
StyleObservationStyleObs_BoldCoreInsight Bold markdown used to foreground the single most important claim in the document. Signals to reader: this is the thesis. Used sparingly — only once for the core insight.
StyleObservationStyleObs_BoldedSectionHeaders Commentary sections use bolded noun-phrase headers ('The money answer', 'The audience question is the live positioning tension') — a structural device that signals the key lesson before the supporting prose. Consistent with Scarlet's editor
StyleObservationStyleObs_BoldEmphasis Bold used consistently for workflow role categories (Flows, LLM Pure Functions, Comms, Entrypoints) and bug type labels. Functions as inline heading/navigation aid.
StyleObservationStyleObs_BoldHeadingStructure The 'Two interlocking insights' section uses bold numbered labels (**1. Deduplication.** **2. Citations are constellations...**) as inline section headers within a prose block — a hybrid structure that signals parallel structure without ful
StyleObservationStyleObs_BoldKeyPrinciple Bold used to foreground architectural principles as standalone declarative statements.
StyleObservationStyleObs_BoldKeyTerms Systematic use of **bold** markdown for key terms and taxonomy labels throughout the document, functioning as inline emphasis and navigation anchors.
StyleObservationStyleObs_BoldSectionHeaders Bold inline section headers ('**The frame.**', '**Caveat on the word.**', '**Use case 1**', '**The unifier.**') function as scannable anchors within flowing prose — a memo-within-narrative hybrid format characteristic of Scarlet's session n
StyleObservationStyleObs_BrainPersonification Personification of shared memory as 'the brain' — humanizes the system and elevates it from tool to cognitive partner; reinforces the superpower framing.
StyleObservationStyleObs_BranchTrunkAnalogy Scarlet uses the organic 'branch and trunk' analogy to make GitHub's version control model accessible to a non-technical user.
StyleObservationStyleObs_brand_name Brand name: lowercase, two words; consistent throughout document
StyleObservationStyleObs_BrandAswritten project** in Claude Desktop
StyleObservationStyleObs_BrandingAnalogy a design collaborator uses named brand analogies (Mailchimp chimp character, Tony as face of Vouch) to ground abstract brand-strategy counsel in recognizable precedents. Characteristic of the advisor-as-practitioner register: theory deliver
StyleObservationStyleObs_BrandName I just started this new company,
StyleObservationStyleObs_BrandName_1 Lowercase brand stylization with .ai TLD.
StyleObservationStyleObs_BrandName_aswritten Brand name aswritten.ai consistently lowercase; stylized as code-adjacent identity
StyleObservationStyleObs_BrandNameBacktick Backtick formatting used for claim IDs, code paths, and function names. Consistent technical inline-code convention.
StyleObservationStyleObs_BrandNameClaudeCode 'Claude Code' used consistently as a proper noun / product name throughout; never abbreviated or varied
StyleObservationStyleObs_BrandNameCode Backtick inline code used consistently for product/system proper nouns: workflow names, file paths, branch names, CLI commands.
StyleObservationStyleObs_BrandNameLowercase Brand name 'aswritten' rendered in lowercase throughout the pitch body, including at sentence-start positions. Consistent stylization choice.
StyleObservationStyleObs_BrandNameMCP 'MCP server' as technical brand term used consistently. 'SAFE', 'YNAB', 'Fireflies', 'Clerky' as named-entity anchors establishing credibility through specificity.
StyleObservationStyleObs_BrandNameStylization Uses brand name as a tier/quality benchmark ('Crooked Media-tier') — shorthand for high-visibility, credibility-signaling partners without elaboration
StyleObservationStyleObs_BrandStylization provides a shared and collaborative worldview
StyleObservationStyleObs_BrandStylization_1 Lowercase, no space, .ai TLD; consistent brand stylization throughout.
StyleObservationStyleObs_BridgeMetaphor You're the bridge to that and the bridge to the gap that you get consumed by one of them.
StyleObservationStyleObs_BrokenSyncQuote Path A / Path B binary framing with rhetorical self-answer; 'It fizzles out' as a short punchy cadence close — vivid human cost articulation.
StyleObservationStyleObs_BuilderTrap The builder trap isn't \"building instead of selling\" — it's \"not knowing what you're selling\"
StyleObservationStyleObs_BulletedConsequences Consequences of design decisions rendered as tight bullet lists with parallel structure, reinforcing logical inevitability.
StyleObservationStyleObs_BulletedDecisionFormat_Distillation Decisions are structured as bulleted 'What changed' / 'What didn't change' lists under bold headers, a consistent pattern in aswritten design session memos. Parallelism is enforced within each list.
StyleObservationStyleObs_BulletedDecisionLists Decisions rendered as bulleted lists with explicit workflow names to delete — authoritative, action-oriented register
StyleObservationStyleObs_BulletedLaunchList Above/below-the-line structure uses dense bulleted lists for launch items — a checklist register distinct from the prose-heavy gift round section. The document shifts register by section: strategic prose for funding paths, checklist for lau
StyleObservationStyleObs_BulletParallelism Adaptive tool states expressed as parallel bold-label bullet structure: **State**: Consequence. Consistent grammatical form across three items.
StyleObservationStyleObs_BusinessShape_1 explicitly rejected the false dichotomy between \"lifestyle business\" and \"real company.\"
StyleObservationStyleObs_BusinessShape_2 The consulting is the manufacturing process.
StyleObservationStyleObs_BusinessShape_3 Scarlet decided she does not need a co-founder.
StyleObservationStyleObs_BYOKAcronym BYOK (Bring Your Own Key) introduced with parenthetical expansion — industry-standard acronym used without assuming reader familiarity.
StyleObservationStyleObs_Cadence_1 It's SIC. That stands for Synthetic Identity Co, which is the legal name.
StyleObservationStyleObs_Cadence_BulletedCriteria Short, punchy bullets with parenthetical clarifications—fast scan, high information density.
StyleObservationStyleObs_CalibratedGrowth_Anaphora Anaphora in the three structural commitments section: 'The company does not bloat... It does not build... It does not grow... It does not pursue...' — a rhetorical list that enumerates the discipline through negation.
StyleObservationStyleObs_CalibratedGrowth_Antithesis Antithesis used throughout to distinguish the calibration discipline from venture-scale behavior: 'not venture approval' vs. 'actual purchase'; 'not the destination' vs. 'the math check'; 'strategic reasons' vs. 'existential reasons'.
StyleObservationStyleObs_CalibratedGrowth_ToneAuthoritative Authoritative declarative tone in the 'Why this works for each audience' section — each audience framing is stated as settled fact, not hypothesis. 'A business that doesn't need venture money is the best kind to give venture money to.' Shor
StyleObservationStyleObs_CallIsFirstMemory Declarative sentence fragment used as a thesis-level assertion; capitalized IS for emphasis
StyleObservationStyleObs_CallIsMemory Aphoristic equation: 'The call IS the first memory.' Capitalised IS for spoken emphasis. Short punchy cadence; functions as a product principle.
StyleObservationStyleObs_CapabilityLedRejection The memory opens with Scarlet's direct rejection of the accumulated pitch as 'capability-led' — a pattern of naming the failure mode before articulating the correction. Characteristic of the session-memo register: diagnosis precedes prescri
StyleObservationStyleObs_CaretBracket No caret-bracket citation markers present in this sample
StyleObservationStyleObs_Cargoculting 'Cargo-culting' used as a single compound verb to describe the failure mode of copying a value without understanding its intent. This is idiomatic engineering jargon deployed precisely — it signals the author's technical fluency and frames
StyleObservationStyleObs_CastleMetaphor Extended castle metaphor for IP disclosure posture: 'full tour of the castle... You could bring them everywhere in the castle except maybe the metaphorical security room where all the cameras are. Even that you could probably walk someone i
StyleObservationStyleObs_CentristDraft Coined compound noun. 'Centrist draft' names the baseline document produced when all perspectives are weighted equally. Precise, neutral, reusable as a technical term.
StyleObservationStyleObs_Checkmark_Validation ✅ **Fills gap_demo_beat_sequence**: Three distinct beats
StyleObservationStyleObs_CitationMarker ):** ~5.9M US employer
StyleObservationStyleObs_CitedOrFlagged 'cited or flagged as ungrounded' is the technical-buyer-facing alternative to the tagline. Precise, binary, operationally legible. Trades memorability for clarity.
StyleObservationStyleObs_CleanToSell Clean to sell: \"What's the next project you're going to onboard junior folks into?\"
StyleObservationStyleObs_ClojureIdiolect Signature Scarlet Dame idiolect: frames n8n workflow design through Clojure functional programming vocabulary. Establishes conceptual authority and architectural lineage.
StyleObservationStyleObs_CodeBlockArchDiagram Fenced code block used for architectural model summary — columnar label:value alignment. Treats architecture as readable artifact, not prose.
StyleObservationStyleObs_CognitiveFrictionList Parallel bullet structure with bold lead-ins; each friction point follows 'problem statement + elaboration' pattern for scannable analysis.
StyleObservationStyleObs_CollectiveMemoryTerm building collective memory as a
StyleObservationStyleObs_Colloquial Well let's be honest. You want the money up front. That too.
StyleObservationStyleObs_ColloquialDirectness seat-of-the-pants founder moment
StyleObservationStyleObs_CompanionAnalogy a consultant who ran enterprise discovery engagements for fifteen years's analogical reasoning style: pairs of historically co-emergent technologies used to argue for fortuitous timing. Internal combustion engine + oil; LLMs + World Wide We
StyleObservationStyleObs_ConflatingProcessesRetraction Self-corrective first-person voice: Scarlet explicitly names her own prior error ('I've been conflating') as the framing for the removal decisions. Authoritative yet self-aware register.
StyleObservationStyleObs_ConsultantInterviewAnalogy Analogy connecting lens drafting to the established consultant interview model. 'The expert is reacting to questions, not writing docs.' Grounds the new behavior in a familiar, validated pattern.
StyleObservationStyleObs_ConsultingFramingAntithesis Antithesis-heavy structure throughout: 'get THEM to use the product' vs. 'ME using the product'; 'install-the-tool model' vs. 'this model'; 'start something' vs. 'keep something'; 'consulting revenue' vs. 'product subscription'. Consistent
StyleObservationStyleObs_ConsultingFramingFirstPerson First-person singular dominant in the quoted speech ('i'm always trying', 'ME using the product', 'what she delivers'). The framing decision is stated as a personal operational principle, not an abstract product claim. Lowercase 'i' in the
StyleObservationStyleObs_ConsultingFramingShortPunchy Short punchy declaratives used as rhetorical anchors at key conviction moments: 'Keeping beats starting.' (3 words), 'The product is her infrastructure' (5 words). These function as aphoristic closers after longer analytical sentences, cons
StyleObservationStyleObs_ContextServerMetaphor 'Context server' is a technical metaphor applied to a human role — the CEO as infrastructure. Characteristic of Scarlet's idiom: borrowing engineering vocabulary to make invisible labor legible.
StyleObservationStyleObs_ContrastiveFraming \"Let me interview you about your business\" is a favor ask. \"Let me show you this thing\" is an invitation.
StyleObservationStyleObs_ConvergenceQuote Scarlet defines review purpose with a binary opposition: 'sharp on the thing the branch is trying to do' vs. 'answer all questions'
StyleObservationStyleObs_ConversationalDistinction Scarlet's canonical phrasing of the AI distinction is preserved verbatim and explicitly flagged as the rule: 'it speaks to, references, and works from what was already said.' The June 12 voice-rule memory records that coined compressions of
StyleObservationStyleObs_ConversationalRegisterTonyScarlet Both speakers use heavily conversational register with filler words ('you know', 'like', 'right'), run-on sentences, and self-interruption. Scarlet's speech is longer and more architecturally structured; Tony's is shorter and more declarati
StyleObservationStyleObs_ConversationalTimestamps Timestamp citation convention [MM:SS] used throughout to anchor verbatim quotes to call recording positions. Functions as a CaretBracketMarker variant for audio sources rather than text documents.
StyleObservationStyleObs_ConversationalVsCompressed Scarlet's idiolect favors multi-clause conversational descriptions over compressed coinages. The phrase 'it speaks to, references, and works from what was already said' is characteristic: three-verb enumeration, no nominalization, no slogan
StyleObservationStyleObs_ConvictionLabel BOULDER used as an inline conviction label in all-caps — a proprietary epistemic marker functioning as both metadata and rhetorical signal of validated confidence.
StyleObservationStyleObs_CounterIntuitivelyPowerful Adverb 'counterintuitively' used as a rhetorical signal to pre-empt objection and reframe a weakness (the AI admitting ignorance) as a strength (epistemic integrity).
StyleObservationStyleObs_CustomerArchitectPitch_DirectSpeech an AI tool that extracts undocumented knowledge from senior members of staff in a conversational way, stores that as memories, and then everybody else can query it. So it's like being able to talk to those people without actually bothering
StyleObservationStyleObs_CustomerCTOGuardPullRequests_Imperative the customer's CTO's follow-up uses direct imperative address ('I do think we should guard these pull requests a little bit') combined with personal acknowledgment ('I appreciate you're kind of on a critical time thing, an engineering manag
StyleObservationStyleObs_CustomerCTOPRMetaphor_Analogy the customer's CTO reaches for a software engineering analogy (PR with unit tests) to explain a knowledge governance principle to a developer audience. The analogy is unprompted, indicating internalization rather than adoption of vendor fra
StyleObservationStyleObs_CustomerEngManager_Antithesis driven by static data, whereas as written is driven by yesterday's live information
StyleObservationStyleObs_CustomerEngManager_DirectPersonal I'm a key candidate for this because I forget as much as I remember right.
StyleObservationStyleObs_CustomerEngManager_Metaphor when we're up against the cold face
StyleObservationStyleObs_CutoverOperatingModelSentence The post-cutover operating model is delivered as a single dense sentence followed by a three-part enumeration ('dev-first, isolated-but-thorough verification over full-suite waits, small follow-up merges'). The sentence is the policy; the e
StyleObservationStyleObs_CyclicalMetaphorDevice 'Cyclical thing' is a structural metaphor for the authority-layer mechanism: person props up business, business props up person, in a loop. Characteristic of a design collaborator's brand-strategy idiom — abstract mechanism made concrete th
StyleObservationStyleObs_DashAsAmplifier Em-dash used repeatedly as an amplifier/clarifier after a claim, delivering the payoff clause. Pattern: [claim] — [consequence or reframe]. Drives cadence and emphasis.
StyleObservationStyleObs_DavidNolen_BlockQuoteStyle Heavy use of attributed block-quotes with timestamp markers (e.g. [28:15], [01:15:34]) to carry evidentiary weight. Speaker attribution precedes each quote. Consistent with Scarlet's session-memo idiolect for call notes.
StyleObservationStyleObs_DavidNolen_SkepticalVoiceContrast The document structures David's skeptical voice against Scarlet's reframes as a series of antithesis pairs: developer-vs-management, siloing-as-bug-vs-feature, demo-value-vs-own-data-required. The skeptic's challenge is preserved verbatim a
StyleObservationStyleObs_DavidNolen_TimestampCitations Timestamp citations in [HH:MM:SS] format used throughout as inline provenance anchors for every attributed quote. Consistent with call-notes format across the perspective.
StyleObservationStyleObs_DavidNolenVerbatimBlockquotes Heavy use of verbatim blockquotes from David Nolen and Scarlet, attributed by speaker. Carries evidentiary weight; consistent with the direct-quote style across the Quotes hub.
StyleObservationStyleObs_DefensiveRegister lowered the defensive register and made the work feel like building rather than confessing
StyleObservationStyleObs_DeferralTable Markdown tables used for architectural decisions and deferred tasks — parallel structure enforces consistent reasoning format (Decision | Rationale | Reversibility).
StyleObservationStyleObs_DeltaTable Coined compound noun. 'Delta table' names the instrument readout showing exactly what moved between lens drafts. Borrows 'delta' from mathematics/engineering to signal difference/change.
StyleObservationStyleObs_DensityInversionQuote Antithetical structure: 'no real upper bound / more dense is always better' vs 'almost exactly the opposite' — rhetorical contrast used to crystallize the insight.
StyleObservationStyleObs_DesignCollaborator_Directness The power is in the chat. I wouldn't fold it out into other pages.
StyleObservationStyleObs_DesignCollaborator_Metaphor It gives you a framework, a structure, infrastructure. That's what the technology is doing in the background.
StyleObservationStyleObs_DesignCollaborator_RhetoricalContrast Most startups shout. You've already moved to a mature thing.
StyleObservationStyleObs_DesignCollaboratorBlockquoteRegister Markdown blockquotes set a design collaborator's direct speech apart from the analytical framing — same dual-register memo pattern as Scarlet's own session memos, here applied to an advisor's voice. The blockquoted speech is colloquial, con
StyleObservationStyleObs_DesignCollaboratorCall_Antithesis Antithesis structures the entire moonshot argument: monoculture vs. diversity, uniformity vs. amplification, 'AI more different, not less', 'amplifier not uniformity machine'. The 'X, not Y' negation pattern is Scarlet's dominant rhetorical
StyleObservationStyleObs_DesignCollaboratorCall_ProfanityIntensifier Profanity used as rhetorical intensifier at conviction peaks — 'every single business around the fucking world' — consistent with Scarlet's voice-memo idiolect. Signals authentic, unfiltered founder voice rather than polished pitch register
StyleObservationStyleObs_DesignCollaboratorCall_RhetoricalQuestion Rhetorical questions used to stage the binary choice at the heart of the moonshot: 'Are we going to end up in a monoculture... Or do we end up with some weird crackpot?' and 'can we have as many distinct voices as there are humans?' The que
StyleObservationStyleObs_DesignCollaboratorTokensAnalogy Design-token analogy used to make the abstract perspective concept concrete: design tokens are to visual surfaces as perspective is to AI outputs. The analogy is structural (both tokenise a lens, both apply via AI, both break without the le
StyleObservationStyleObs_DesignCollaboratorTokensAntithesis Antithesis structure used to frame the target state: 'one-shotting a website [that] looks like [Claude] every time' vs. 'deliver an AI to Coca Cola, where if they ask it to one-shot a website, it looks like Coca Cola.' The contrast makes th
StyleObservationStyleObs_DesignCollaboratorTokensBlockquoteRegister Mixed register: analytical editorial prose (authoritative, structured) alternating with verbatim blockquoted speech (first-person conversational, thinking-aloud). The blockquotes carry the primary evidence; the editorial prose frames the si
StyleObservationStyleObs_DirectAddress Do NOT list as a client.
StyleObservationStyleObs_DirectCorrection Corrections are delivered as flat declarative assertions, not hedged. The speaker restates the wrong claim then replaces it with the correct one in a single sentence.
StyleObservationStyleObs_DirectCorrectionRegister_20260521 authoritative-corrective; declarative negation of prior claim
StyleObservationStyleObs_DirectFirstPerson First-person voice used for high-stakes personal admissions: sales confidence, process frustration. Increases credibility and urgency of the claim.
StyleObservationStyleObs_DirectImperative Imperative mood used for rules and policies without hedging — authoritative, no passive voice.
StyleObservationStyleObs_DirectPersonal_1 I'm not buying it because it's git native.
StyleObservationStyleObs_DirectPersonalRegister_Distillation Scarlet's voice is captured in direct first-person quotes embedded in a structured decision memo. The register is founder-internal: no hedging, declarative, immediate.
StyleObservationStyleObs_DirectPersonalTone_DesignSession First-person direct speech dominates the memory; Scarlet's voice is quoted extensively in full sentences. The document alternates between analytical prose (section headers, design rationale) and verbatim founder speech, creating a dual-regi
StyleObservationStyleObs_DirectQuote_ManagedByUs Inclusive 'us' signals ownership and agency. Concise declarative framing of managed infrastructure abstraction.
StyleObservationStyleObs_DirectQuote_NeverSeeGitHub First-person direct quote from Scarlet; decisive, unhedged, future-oriented. Sets identity architecture in one sentence.
StyleObservationStyleObs_DirectQuote_ScarletBlockquote First-person attributed blockquotes from Scarlet used to anchor key claims with speaker authority; shifts register from analytical to declarative
StyleObservationStyleObs_DirectQuote_UXIsShit Unfiltered register; profanity used for emphasis. Signals low-ceremony internal communication style and high conviction about UX problem.
StyleObservationStyleObs_DirectQuoteEvidence Direct verbatim quotes from prospects (the customer's chief architect: 'I wish we could clone me'; an early prospect's full text message) used as primary evidence throughout. The document treats prospect quotes as data points, not anecdotes
StyleObservationStyleObs_DirectQuoteIntegration First-person voice memos transcribed as block quotes integrated directly into structured decision document — hybrid voice-memo-to-spec register
StyleObservationStyleObs_DirectQuoteRegister Scarlet's diagnosis is rendered as verbatim block-quoted speech, creating a direct, first-person register that contrasts with the analytical prose surrounding it. The document shifts register deliberately between analysis and testimony.
StyleObservationStyleObs_DirectQuoteStyle Heavy use of direct block-quoted speech from named speakers. Verbatim quotes carry the evidentiary weight of the document; analysis is kept separate from the quoted material.
StyleObservationStyleObs_DirectRejection Scarlet's voice: blunt, first-person rejection with embedded rationale
StyleObservationStyleObs_DirectSpeechPlaybook Tony's quoted speech is directive and filler-heavy ('Right. Right. Blah, blah, blah.') — a characteristic idiolect that signals informal authority. The playbook steps are dictated verbatim in Tony's voice, giving the document a transcript-a
StyleObservationStyleObs_DirectTone Hook: \"Add collective memory to your repo, never write README again\"
StyleObservationStyleObs_DirectVoice Colloquial resolution after technical enumeration. 'All of that' as a deliberate plain-language catch-all after precise technical list.
StyleObservationStyleObs_DiscoveryConsultantBooleanCadence a consultant who ran enterprise discovery engagements for fifteen years's rhetorical pattern: short declarative sentences that land on a Boolean or binary verdict. 'It's falsifiable. The agent can either open the car or it can't. It's very
StyleObservationStyleObs_DiscoveryConsultantCallCadence The memory alternates between long analytical passages (a consultant who ran enterprise discovery engagements for fifteen years's HATEOAS explanation, Scarlet's demo walkthrough) and short punchy declaratives at conviction peaks ('This is t
StyleObservationStyleObs_DiscoveryConsultantCallMetaphor Two strong metaphors in this call: (1) 'Chinese Whispers' for provenance distortion — accessible, culturally resonant, immediately understood; (2) 'clip the wings of the raven' for over-focusing risk — poetic, memorable, from a consultant w
StyleObservationStyleObs_DiscoveryConsultantCallRegister a consultant who ran enterprise discovery engagements for fifteen years's idiolect: technical-philosophical, draws on RFC/IETF vocabulary (HATEOAS, RFC 8288, HEAD requests), uses historical computing metaphors ('surfing the web in the 90s',
StyleObservationStyleObs_DiscoveryWedge_CaveatEmbedded Conviction status is embedded in the document header as a bolded caveat, not buried in body text. Characteristic of Scarlet's memo idiom: epistemic state declared upfront before the argument is made.
StyleObservationStyleObs_DiscoveryWedge_MetaphorRaven clip the wings of the raven
StyleObservationStyleObs_DiscoveryWedge_VerbatimBlockquote a consultant who ran enterprise discovery engagements for fifteen years's pitch is rendered as a verbatim block-quote with named attribution ('Bob, the previous senior developer'). The quote carries the evidentiary weight of the GTM claim —
StyleObservationStyleObs_DistortionIsInstrument Four-word declarative sentence functioning as an aphorism. Subject-verb-complement with zero hedging. Delivers the core mechanism insight in minimal words. Characteristic of the author's style for key insights.
StyleObservationStyleObs_DollarFigureAnchor Concrete dollar anchor ([amount] / 15k) used twice to calibrate the value of the engagement being given away. Lowercase '15k' in direct quote signals informal spoken register.
StyleObservationStyleObs_DomainJargon Controlled jargon: 'TX' (transaction), 'PR' (pull request), 'zone', 'lens', 'centrist', 'introspect'. Used without definition — assumes insider audience.
StyleObservationStyleObs_DoNotBullets Anaphoric 'Do NOT' repeated four times in rapid succession; imperative mood creates a checklist-as-manifesto effect
StyleObservationStyleObs_DontGiveThemAnOut Don't give them an out before they need one.
StyleObservationStyleObs_EmbeddedVerbatimQuotes Extended verbatim Scarlet quotes embedded in analytical prose — run-on, filler-heavy, first-person speech ('sort of a technical example, but they're also really talking about client capability') contrasts with the tighter editorial framing.
StyleObservationStyleObs_EmDash_1 Em dash used to introduce clarifying detail; creates pause and emphasis.
StyleObservationStyleObs_EmDashCadence Em-dash used to append punchy consequence clauses. Creates staccato rhythm; consequence lands hard after the dash.
StyleObservationStyleObs_EmDashCausal Em-dash used as a high-velocity causal connector — replaces 'because', 'therefore', 'which means'. Accelerates logical flow without subordinate clauses.
StyleObservationStyleObs_EmDashDefinition Em-dash used consistently to append clarifying definitions inline: 'TX summary at full scale — show what the PR changes...'
StyleObservationStyleObs_EmDashRhetoric Em dash used as a rhetorical pivot to introduce the payoff clause — a recurring cadence pattern.
StyleObservationStyleObs_EmphaticNegation We will NOT license the semantic memory layer as infrastructure for others to build
StyleObservationStyleObs_EngagementShape_Antithesis Heavy antithesis structure throughout: automation vs. conversation, gift vs. contract, staff-work vs. friendship, accumulation vs. conversation. Each section resolves a tension by naming both poles explicitly before stating the synthesis.
StyleObservationStyleObs_EngagementShape_JargonPolicy Product-specific jargon (webhook, repo, branch routing, backfill diagnostic, SOW) used without definition, implying internal audience. General business terms (calibration, cadence, engagement) used in product-specific senses that differ fro
StyleObservationStyleObs_EngagementShape_ShortPunchy Short punchy declaratives at conviction peaks: 'Calibration is staff-work, not friendship', 'price the accumulation, not the conversation', 'this consulting version is actually just me.' Each is a standalone sentence functioning as a thesis
StyleObservationStyleObs_EngineeringMetaphor Recurring engineering metaphors used to make knowledge management intuitive to a developer audience: memory-as-code, PRs, unit tests, git branches. the customer's CTO adopts and deploys these metaphors natively.
StyleObservationStyleObs_EnterpriseCustomerPattern_Antithesis anchored in the customer's own usage evidence rather than in aswritten's marketing claims
StyleObservationStyleObs_EnterpriseCustomerPattern_Jargon Feltron-style PDF
StyleObservationStyleObs_EnterpriseCustomerPattern_Tone The the enterprise customer pilot is informally validating a specific go-to-market shape worth naming and codifying
StyleObservationStyleObs_EnterpriseCustomerPilot_Antithesis Antithesis used to contrast traditional product response (dashboard) with agent-interface approach throughout the document.
StyleObservationStyleObs_EnterpriseCustomerPilot_BoldFraming Bold text used as rhetorical anchors for key claims: 'aswritten tells you what you don't know yet.' and 'Memory commits take 3 to 5 minutes'. Signals conviction peaks within analytical prose.
StyleObservationStyleObs_EnterpriseShape_Antithesis additive vs. incremental; primary track vs. secondary track; sponsor vs. champion; tentative vs. decision — antithesis pairs structure the entire document
StyleObservationStyleObs_EnterpriseShape_TentativeRegister Explicit epistemic hedging throughout: 'tentative', 'needs validation', 'working hypothesis', 'potentially rationalizing' — unusually self-critical register for a strategic memo; signals high intellectual honesty
StyleObservationStyleObs_EnterpriseShape_TimestampCitation Timestamp citation convention [31:50] used throughout to anchor verbatim quotes to call recording — same pattern as narr:Sample_TonyMaleyCheckin_20260421
StyleObservationStyleObs_EpistemologicalRegister The memory operates in an unusually high epistemological register for a product design session: recognition, creativity, association, and projection are used as precise technical terms, not metaphors. This is the idiolect of someone who has
StyleObservationStyleObs_EVFlocking_Anaphora Anaphoric 'We' construction closes the piece: 'We too are compiled, from stories inherited and newly written.' 'Our work together, aswritten.' Shifts from 'I' (founder origin) to 'We' (collective) — the structural pivot from personal ground
StyleObservationStyleObs_EVFlocking_ArrowMetaphor Archery/arrow metaphor closes the flocking piece: 'feather our arrows,' 'fly further in the direction we're headed,' 'adjust our aim,' 'all arrows find target.' Extends the steering-vector physics metaphor into a kinetic, directional image.
StyleObservationStyleObs_EVFlocking_FirstPerson The flocking piece is written in sustained first-person singular throughout — 'I started,' 'I spent years,' 'I left Vouch,' 'I built aswritten.' Characteristic of Scarlet's founder-voice register: personal grounding as epistemic authority.
StyleObservationStyleObs_ExtractionPlainLanguage Scarlet uses plain, accessible language to explain a technical backend process to a non-technical user; avoids jargon in favour of 'breaks down and understands'.
StyleObservationStyleObs_FableWindowAntithesisPattern Antithesis structures the 'What was explicitly not chosen' section: each alternative is named and then categorized as 'deferred rather than rejected,' 'dissolved rather than rejected,' or 'partially adopted.' This dissolved/deferred/rejecte
StyleObservationStyleObs_FableWindowBlockquoteStyle Planning session memo with embedded blockquoted speech. The document alternates between authoritative third-person editorial prose (decisions, rationale, corrections) and first-person blockquoted Scarlet voice. The blockquote is the primary
StyleObservationStyleObs_FailureProgression Parallel numbered list structure (1-5) maps the failure progression. Each item follows the same pattern: approach description → arrow → outcome. Creates rhetorical momentum toward the breakthrough.
StyleObservationStyleObs_FeatureNotBug Designing the triple review system that moves extraction from
StyleObservationStyleObs_FirstPerson I just started this new company
StyleObservationStyleObs_FirstPersonNarration Scarlet's quoted passages are consistently first-person, direct, and technically assertive. a consultant who ran enterprise discovery engagements for fifteen years's quotes mirror this register. The document voice is third-person analytical
StyleObservationStyleObs_FirstUserPositioning The phrase 'aswritten is its own first user' is a rhetorical anchor that grounds the eval discipline in product identity rather than process compliance. It elevates the principle from a rule to a founding thesis. Consistent with the idiolec
StyleObservationStyleObs_FiveQuestionCadence The five-question framework uses a numbered list with parallel interrogative structure — each question pairs a diagnostic axis with a concrete observable. Consistent with Scarlet's session-memo idiolect of structured enumeration at convicti
StyleObservationStyleObs_FiveToolTable Markdown table used as canonical architecture definition: Tool | Question | Input | Output — four-column single-responsibility matrix
StyleObservationStyleObs_ForgettingBottleneck_Antithesis Antithesis structure: 'not scarce expertise / scarce memory'; 'not one person's quirk / structural to the role'; 'offered versus agreed, options versus solution'. Antithesis is the dominant rhetorical device throughout, used to sharpen the
StyleObservationStyleObs_ForgettingBottleneck_BoldEmphasis Bold used for two bullet-point conclusions ('The buyer is plural.' and 'The self-identification is the qualifying signal.') — a formatting pattern that signals these are the extractable strategic takeaways from the convergence observation.
StyleObservationStyleObs_ForgettingBottleneck_DirectQuote Heavy use of verbatim blockquoted speech from an engineering manager at the customer — three distinct quotes, each introduced with attribution and framing. The document explicitly flags these as verbatim ('in an engineering manager at the c
StyleObservationStyleObs_Formulaic_1 you know, there is no RDF database in the first place. So, you know
StyleObservationStyleObs_FounderAffirmationPattern Scarlet uses short, emphatic affirmations to mark crystallization moments: 'That's gold.' and 'Yeah, bam, spot on.' These idiolect markers signal when a formulation has landed — they function as conviction signals embedded in the transcript
StyleObservationStyleObs_FragilityHedge Concession pattern: risk acknowledged then immediately bounded — 'This is acknowledged... This is acceptable: better than...'
StyleObservationStyleObs_FundraisingBriefReview_DirectQuoteAnchor Both major sections (PnD correction and Calm Company thesis) are anchored by a direct Scarlet quote in bold. This is a consistent idiolect pattern across aswritten memories: the founder's verbatim speech functions as the conviction marker,
StyleObservationStyleObs_GapBetweenConnectedAndValue Rhetorical antithesis framing the core product problem as a gap between two named states; bold emphasis amplifies the tension.
StyleObservationStyleObs_GenusSpeciesAntithesis genus-species antithesis: 'Steering covers all three use cases; governance covers only the first.' Tight parallel structure with asymmetric scope — the genus/species distinction is made in a single antithetical sentence.
StyleObservationStyleObs_GitHubMetaphor No company would develop software without GitHub. No company should deploy AI agents without versioned organizational context.
StyleObservationStyleObs_GitMetaphor like Git for software, aswritten.ai is for AI collaboration
StyleObservationStyleObs_GitNative 'git native' is a coined technical descriptor positioning As Written within developer-familiar infrastructure paradigms.
StyleObservationStyleObs_GodsMetaphor It's almost like the gods of Olympus. If Zeus decides to strike you with lightning, one, there's nothing you can do about it.
StyleObservationStyleObs_GoldMineMetaphor 'gold mines for AI' — value-extraction metaphor applied to dense thinkers; economically resonant framing that reframes a perceived liability as an asset.
StyleObservationStyleObs_GoogleAnalogy an organization like Google really only has like 10 or 15 people actually working on search
StyleObservationStyleObs_HandWaving it's great to like do hand waving and go oh, it'll propagate and it does this and it does that
StyleObservationStyleObs_HaveAnOpinionImperative Imperative construction 'have an opinion and here's why' — a two-clause command that collapses stance and evidence into one sentence. Characteristic Scarlet move: the product promise IS the rhetorical form it advocates.
StyleObservationStyleObs_HeavyLiftMetaphor the interview may be sounding like a kind of heavy lift
StyleObservationStyleObs_HowMadRhetQ Rhetorical self-interrogation after vivid behavioral anecdote; British-inflected idiom ('How mad is that?') adds authenticity and invites reader alignment.
StyleObservationStyleObs_ICPQuote You're looking for a dev shop that's got 10 to 50 people, probably something like 10 to [amount] in revenue.
StyleObservationStyleObs_IdentityMetaphor The agent will ground itself somewhere; the question is whether that grounding is correct or invented.
StyleObservationStyleObs_Idiolect_1 you were like, wow, we were
StyleObservationStyleObs_IdiolectPhrasing Phrasal idiom 'pulled back and forth' used in founder voice to describe the building/selling attention split — visceral, kinetic language
StyleObservationStyleObs_Imperative It's a hell of a lot easier to put it in now than it is to try to bolt that stuff on later
StyleObservationStyleObs_ImperativeCadence short-punchy-tagline; the deli story closes with a three-word imperative tagline ('Vouch — be recognized') that compresses the entire reframe into a brand directive — the shortest possible unit of the motion
StyleObservationStyleObs_ImperativeTechnical SQL block and numbered check steps use imperative/second-person framing — direct, no hedging, assumes reader competence.
StyleObservationStyleObs_ImperativeVoice_HoldIt The 'How to hold it' section uses imperative voice throughout: 'Pre-register the comparison before building,' 'Keep the baseline as a real control,' 'Report the actual numbers,' 'If a metric falls short, iterate — don't narrate around it.'
StyleObservationStyleObs_InformalRegisterInQuote Direct speech quote preserves informal spoken register: lowercase '15k', dropped subject 'I', colloquial 'I'd be like'. Contrasts with formal analytical prose surrounding it.
StyleObservationStyleObs_InlineCode Bundle `Snapshot` + `Ontology` into single `State` type
StyleObservationStyleObs_InlineTaskRef Inline task references (task-068, task-070, task-071, task-072) used as stable identifiers linking prose to a backlog. Engineering convention applied to narrative writing.
StyleObservationStyleObs_InstallableExpertiseTerm 'installable expertise' as the replacement for 'memory' — a deliberate terminology shift that moves from technical/process framing to outcome/value framing
StyleObservationStyleObs_IntelligenceVsPerspective Intelligence is quick, perspective is hard won
StyleObservationStyleObs_InterviewFormatDirectPersonal The 'What to Replicate' section uses imperative second-person directives — 'Interview the user before synthesizing', 'One question at a time', 'Name patterns without prescribing' — creating a direct, actionable register that mirrors the int
StyleObservationStyleObs_ItalicsForEmphasis Italics used for contrastive stress on key words within sentences, particularly to mark the reframe: 'better' demo, 'work' with AI, 'you'.
StyleObservationStyleObs_ItsPeopleNow It's about people now.
StyleObservationStyleObs_Jargon how do you manage context windows, especially in large docs?
StyleObservationStyleObs_Jargon_1 canonization and, you know, skolemization and that kind of stuff
StyleObservationStyleObs_Jargon_EnterpriseCustomer marginal cost of adding a second channel is near-zero because Front's unified inbox model
StyleObservationStyleObs_JargonSplit Intentional split between user-facing human words and internal engineering metaphors.
StyleObservationStyleObs_KeyInsightCallout 'Key Insight' as a named section heading elevates one finding above the enumerated bug list, signaling architectural significance beyond the immediate fix.
StyleObservationStyleObs_KeyPhrase_IPAsCode Core terminology: 'IP as code'—makes intangible IP tangible, versionable, reviewable; aligns with git-native worldview.
StyleObservationStyleObs_KeyPhrase_NetworkEffects Repeated framing: 'network effects: more people = more value to each person'—emphasizes collaborative multiplier, not individual tool.
StyleObservationStyleObs_KeyPhrase_PMF Standard startup terminology; signals stage and traction.
StyleObservationStyleObs_KeyPhrase_StateSec Signature phrase: 'telling state secrets to AI'—vivid, memorable, frames IP leakage as espionage-level risk.
StyleObservationStyleObs_KeyReframeTagline Closing reframe delivered as a short punchy sentence functioning as a tagline or thesis statement. Cadence: short declarative after longer explanatory prose.
StyleObservationStyleObs_KidsInheritingQuote Generational framing ('the kids are inheriting') used to signal macro-level trend conviction; authoritative declarative tone with colloquial warmth.
StyleObservationStyleObs_KillerUseCase the customer's chief architect: \"
StyleObservationStyleObs_KungFu Pop culture reference (The Matrix) to explain instant skill transfer via worldview marketplace
StyleObservationStyleObs_LaunchingDefinition Launching = moving from being privately terrified...
StyleObservationStyleObs_LeanStartupAnalogy Scarlet uses the Lean Startup as an analogy to ground her AI-era principle — 'if you know the Lean Startup... his whole thing is like we have to test.' The analogy is used to validate the principle by reference to established methodology, t
StyleObservationStyleObs_LensDraftingTerm Coined compound gerund naming the new product behavior. 'Lens drafting' combines the optical metaphor with the writing action. Appears in the document title and throughout as the canonical term.
StyleObservationStyleObs_Lenses memories are not necessarily human-authored prose. The `
StyleObservationStyleObs_LensMicroTelltale System generates micro-telltales — small content artifacts that exercise the triples in a given lens.
StyleObservationStyleObs_LowercaseIdiolect Founder voice in blockquotes uses consistent lowercase, no sentence-initial caps, stream-of-consciousness cadence — signals authenticity and informality
StyleObservationStyleObs_LowercaseProperName Proper name 'the planned engineering/architecture hire' rendered in lowercase within founder blockquote — consistent with idiolect pattern; signals informal register and familiarity
StyleObservationStyleObs_McDonaldPictureMenu Verbatim analogy from a demo-call prospect: 'McDonald picture menu' as shorthand for a visual-first, executive-accessible pre-read deck. Idiomatic, colloquial, highly concrete. Signals a demo-call prospect's direct, no-jargon communication
StyleObservationStyleObs_MCP Acronym used without expansion; assumes domain knowledge
StyleObservationStyleObs_MCPConnectedImmediately Staccato three-step sequence with em-dash connectors; mirrors technical precision and immediacy of the onboarding success.
StyleObservationStyleObs_MermaidAsDocPattern Extensive use of Mermaid diagrams (9 total) as primary visual documentation pattern: state lifecycle, memory creation flow, type composition, context tiers, multi-tenant isolation, filtering patterns, type hierarchy, tool dependencies, toke
StyleObservationStyleObs_Metaphor_1 Tony on the follow-up \"
StyleObservationStyleObs_Metaphor_Breadcrumb The import tool's \"Use perspective=...\" error message is the most effective breadcrumb.
StyleObservationStyleObs_Metaphor_CollapseDistinction 'Collapses the distinction' used twice as a structural metaphor — two separate concepts merging into one via the personal graph mechanism. Signals conceptual compression.
StyleObservationStyleObs_Metaphor_InstallableProgram Extended metaphor: worldview as installable program; ontology as the program; steering as the product
StyleObservationStyleObs_Metaphor_KnowledgeLoss no manual documentation, no knowledge loss
StyleObservationStyleObs_metaphor_living living organizational worldview
StyleObservationStyleObs_Metaphor_Pillars Memory 2: Target Use Cases & Market Positioning\nTitle: ADR-002: Three Pillars
StyleObservationStyleObs_Metaphor_PromptAsSchema schema properties are prompts
StyleObservationStyleObs_Metaphor_Steering narrative is not only the backdoor to the models, it's the steering wheel, it's the fault lines
StyleObservationStyleObs_Metaphor_TwoDoors Two doors into the same house.
StyleObservationStyleObs_Metaphor_Vector Technical metaphor (vector) applied to narrative; precision and directionality.
StyleObservationStyleObs_MetaphorWhiteSpace White-space-collapse is the document's most vivid rhetorical metaphor: LLM output space narrows as organizational perspective is added. Concrete, spatial, and technically grounded.
StyleObservationStyleObs_MethodologyValidation This directly validates 2026-05-01-6-onboarding-codebase-review-topic-analysis-not-direct-ingestion — the customer's chief architect independently arrived at the same anti-pattern conclusion
StyleObservationStyleObs_MicroTelltales Review status is per-triple at the extraction level.
StyleObservationStyleObs_MindBlowingCadence Two-fragment declarative close: 'Mind blowing. Totally easy sell.' — zero-subject elliptical sentences used as emphatic verdict; characteristic of spoken-register written capture.
StyleObservationStyleObs_MoatMetaphor approach are our core competitive moat
StyleObservationStyleObs_ModelComparison_3 the real quality differentiator isn't one-shot extraction accuracy — it's how the model handles the relationship between new knowledge and existing graph state
StyleObservationStyleObs_ModelComparison_VerdictCadence Short punchy verdict sentences alternate with dense technical elaborations; antithesis structure in update problem section
StyleObservationStyleObs_ModularSynthesizerSimile Extended simile comparing the AI-assisted strategy session to playing modular synthesizers — emergent behavior guided but not controlled. Captures the collaborative dynamic without prescribing it.
StyleObservationStyleObs_MoonshotVoiceMemoRegister Voice memo / morning pages register: stream-of-consciousness, first-person singular, run-on sentences, filler words ('you know', 'right', 'like'), self-interrupting mid-thought. Contrasts with the structured memo register of most other memo
StyleObservationStyleObs_MVPRedef_Antithesis Heavy use of antithesis to define the reframe: 'ambient vs. on-demand', 'investment you make vs. tool you use', 'what the internet says about X vs. how an expert thinks about X'. Antithesis is the primary rhetorical device for communicating
StyleObservationStyleObs_MVPRedef_DirectPersonal Founder voice is direct and first-person throughout. Strategic reframes are delivered as declarative assertions, not hedged proposals.
StyleObservationStyleObs_MVPRedef_IdiolectPhrasing Scarlet's idiolect includes 'something shitty-but-working', 'dirty env', and 'That's gold' — colloquial markers that signal high conviction and personal ownership of the insight. These phrases appear at moments of strategic crystallization.
StyleObservationStyleObs_MVPRedef_ShortPunchy Key strategic conclusions are delivered in short punchy sentences: 'A repo IS a module — no new abstraction needed.' 'That's gold.' 'The product doesn't replace search — it replaces the expert in the room.' These short declaratives anchor l
StyleObservationStyleObs_NamePatternsWithoutPrescribing The phrase 'Name patterns without prescribing' is a stock phrase that encodes the interview methodology's core discipline — it appears as a rule and also enacts itself by naming the pattern without prescribing how to apply it.
StyleObservationStyleObs_NaturalEvolutionPhrasing Scarlet uses 'natural evolution' as a stock phrase to describe tier progression — consistent with her idiolect of framing product architecture in organic/biological terms rather than mechanical ones. Also: 'feeder' as a metaphor for the Exp
StyleObservationStyleObs_NegotiationLanguage Collaborative negotiation framed as quoted dialogue: 'this covers more than the TX touches, want to narrow?' Softens directives with user agency.
StyleObservationStyleObs_NicheAntithesisStructure Heavy antithesis structure throughout: 'install expertise' vs. 'matching-search', 'treatment proxy' vs. 'matching/vetting instrument', 'external funnel' vs. 'internal organization', 'CPC' vs. 'subscription'. Each section is structured as a
StyleObservationStyleObs_NicheBlockquoteVerbatim Document uses blockquote (>) for verbatim Scarlet speech, consistent with the broader aswritten memo convention. Blockquotes are run-on and conversational; editorial prose is structured and declarative. The contrast creates a two-register d
StyleObservationStyleObs_NoCompanyAnaphora No company would develop software without GitHub. No company should deploy AI agents without versioned organizational context.
StyleObservationStyleObs_NoEmDashPolicy Em-dashes prohibited everywhere AI-generated text accumulates: code comments, design docs, memories, public copy. Replace with commas, periods, or colons.
StyleObservationStyleObs_NoEmoji Authoritative technical tone; no emoji or casual markers; evidence-led
StyleObservationStyleObs_NoNoNoAnaphora Emphatic anaphoric repetition used to mark a hard philosophical boundary. Five-fold 'No' signals categorical rejection, not negotiation.
StyleObservationStyleObs_NoOnesSellingThat No one's selling that as a solution. Even if there's tools that can do it... no one's selling that as a solution.
StyleObservationStyleObs_NotFacts You don't ask \"is this triple true?\" — you ask \"does viewing through this lens produce the right output?\"
StyleObservationStyleObs_NotionCadence Notion: can be changed easily
StyleObservationStyleObs_NotionLevelCadence notion-level-hedging: The memory uses a consistent hedging register — 'single anchor point,' 'worth tracking,' 'not a recommendation,' 'notion until repeated.' This is the Scarlet session-memo idiolect for Notion-level conviction: explicit
StyleObservationStyleObs_NotNegotiationFloor backtick-iri-convention: The memory uses backtick-wrapped IRIs (e.g., `Claim_10xGapIntentional`, `Claim_EnterpriseCustomerPEBoardInterest`) to reference existing graph entities inline. This is a consistent Scarlet idiolect for cross-referen
StyleObservationStyleObs_NotOnboarding Antithesis structure: 'not X, it's Y' — reframes the product category by explicit negation
StyleObservationStyleObs_NotPattern Recurring 'X, not Y' antithesis used to sharpen scope boundaries and reframe constraints as features. High conviction signaling device.
StyleObservationStyleObs_NotSellingPanacea You're not selling panacea. You're selling a topical solution for a specific problem.
StyleObservationStyleObs_NoXNotYAntithesis The 'X, not Y' antithesis pattern is prohibited as a sentence-closing device. It 'lands the plane' instead of building momentum. Example cut: 'Both are tools the AI uses during the work, not reports you read after' -> 'The AI calls both dur
StyleObservationStyleObs_NumberedCheckSteps Numbered lists used for sequential check logic (task-085 divergence check) — enforces procedural reading order and makes branching logic explicit.
StyleObservationStyleObs_NumberedConvergence Numbered list of three convergent problems used to build logical inevitability for the triple store solution. Parallelism reinforces that these are equivalent-weight problems.
StyleObservationStyleObs_NumberedFlow Numbered list cadence for process flows: each step is a single action verb phrase, no subordinate clauses — imperative procedural register
StyleObservationStyleObs_OnboardingMethodology_Antithesis Heavy antithesis structure throughout: 'NOT direct ingestion' vs. 'topic analysis'; 'transcript is not the memory; the transcript is the diagnostic.' The anti-pattern section is structured as a direct inversion of the recommended pattern, c
StyleObservationStyleObs_OnboardingMethodology_BlockquoteVoice Blockquoted verbatim speech from Scarlet is run-on and filler-heavy ('you know', 'like', 'sort of'), contrasting with the tighter editorial prose in the numbered steps. This dual-register pattern — structured memo prose + conversational blo
StyleObservationStyleObs_OnboardingMethodology_DirectPersonal First-person direct address throughout; Scarlet speaks in present tense imperatives ('Use Claude / Copilot CLI', 'you might do instead of') that position the methodology as actionable guidance, not abstract theory. Mirrors the idiolect patt
StyleObservationStyleObs_Onboards Domain-specific verb; active voice; concise
StyleObservationStyleObs_OneSelector3Markets 'One selector, three markets' — a coined compression device that maps a single rhetorical principle (opinionated perspective) onto three distinct decision contexts (hiring, funding, buying). Characteristic of Scarlet's idiom: structural eco
StyleObservationStyleObs_OpenQuestions Open Questions section uses direct rhetorical questions to surface unresolved design decisions — each question is a bounded scope
StyleObservationStyleObs_OpenQuestionsSection Dedicated 'Open Questions' section signals epistemic honesty and invites collaboration. Mirrors the 'gap honesty' beat in the demo itself — the document practices what it preaches.
StyleObservationStyleObs_OpinionatedWorkflow Apple analogy used to anchor product positioning: opinionated workflow vs. everything-tool
StyleObservationStyleObs_OpinionatedWorkflowAnalogy Apple vs. Microsoft analogy used to frame opinionated workflow positioning. Recurring rhetorical device: contrast with a negative archetype to sharpen identity.
StyleObservationStyleObs_OpsPersonaQuote Look for operations people because they've got the most chaotic job right now... We're making 75 decisions a day.
StyleObservationStyleObs_OptimizationNotPrereq 'Optimization, not prerequisite' is a recurring reframe idiom — repositions deferred work as additive rather than blocking. Appears multiple times as a scope-reduction rhetorical move.
StyleObservationStyleObs_Optometrist Review is iterative and convergent, like an eye exam with a phoropter.
StyleObservationStyleObs_OptometristMetaphor Extended metaphor mapping product model onto optometry: eye chart = document, refraction = commentary/distortion, lens = perspective weight. Sustained across multiple sections; functions as the product's explanatory anchor.
StyleObservationStyleObs_OrgChangeAsCode I'd also like to add something about organizational change as code, ie being able to quantify the way the org is changing
StyleObservationStyleObs_OrgTierAgentQuote Scarlet's direct-personal register: mid-session clarification delivered as a correction ('You also get agent deployments at org tier, not just EDOs'). Characteristic pattern of using 'not just X' to reframe a prior framing that was too narr
StyleObservationStyleObs_OutsideTechBetaUserChannel_Antithesis Heavy antithesis structure throughout: SMS works / SMS flips; scoping not replacement; high-trust contexts / evaluator contexts; does not validate / does not invalidate. The document's rhetorical spine is built on paired oppositions that sh
StyleObservationStyleObs_OutsideTechBetaUserChannel_LoadBearing Distinctive phrase 'load-bearing' used twice to signal that a contextual detail is structurally important to the argument, not incidental. 'The Twilio rejection context is load-bearing here, not incidental.' Idiolect marker for Scarlet's an
StyleObservationStyleObs_OutsideTechBetaUserChannel_ThreeFactsSeparated Explicit enumeration of three facts that 'the perspective has collapsed into one classification' — a rhetorical device used to force analytical separation of conflated concepts. Pattern recurs in the commercial-status correction section.
StyleObservationStyleObs_OverExplainTrap 'Over-explain trap' is a coined metaphor for a named anti-pattern; used twice to reinforce the concept
StyleObservationStyleObs_Parallelism_1 Parallel structure in list; each item follows 'role + gerund' pattern; builds rhythm.
StyleObservationStyleObs_Parallelism_Alternatives Alternatives Considered:\n\nSingle use case focus (e.g., just coding agents) - Rejected\nEnterprise-first (sell to Capitol-like companies) - Rejected\nHorizontal \"AI memory platform\" (any use case) - Rejected
StyleObservationStyleObs_ParenthesisGloss Parenthetical glosses immediately follow technical terms to anchor them: 'annotate', 'introspect', 'remember'. Pattern: **Term (verb).** — names the beat, then glosses the product feature.
StyleObservationStyleObs_ParentheticalAsides Frequent use of parenthetical asides in italics to flag session-sourced context, uncertainty, or meta-commentary not stated in transcript. Signals epistemic care about source reliability.
StyleObservationStyleObs_ParentheticalClarification Inline parentheticals used to immediately gloss technical terms: '(that's scope)', '(or arbitrary change description — a TX is really just...)'
StyleObservationStyleObs_PathDependentTrajectoryQuote The phrase 'path-dependent, trajectory-grounded reactions' as a description of what the solo-partner demo produces — a technically precise, jargon-dense formulation used in a fundraising context. Characteristic of Scarlet's idiolect: borrow
StyleObservationStyleObs_PerspectiveOrdering_Antithesis Antithesis structures the core design argument throughout: stable vs. per-call, append vs. reorder, precondition vs. benefit, sort vs. rank. The antithetic framing is the primary rhetorical device for articulating the ordering decision.
StyleObservationStyleObs_PerspectiveOrdering_TechnicalRegister High technical register throughout: 'byte-stable', 'cache_control', 'cache-hostile', 'per-call material', 'volatility ordering'. Engineering vocabulary applied to describe a product design decision. Consistent with Scarlet's idiom of borrow
StyleObservationStyleObs_PerspectiveOrdering_VerbatimBlockquote Verbatim blockquote from Scarlet used to anchor the prompt-caching rationale — the document distinguishes between editorial summary and Scarlet's own voice ('Her reasoning, verbatim:'). This is a recurring pattern in Scarlet session memos:
StyleObservationStyleObs_PitchVerbatim The pitch line 'Your book is a snapshot. We'll turn it into a living AI module.' is a crystallized two-beat cadence: diagnosis (snapshot = static) then solution (living module = dynamic). Matches the 'book is a snapshot; module is alive' pa
StyleObservationStyleObs_PlainLanguagePolicy Explicit jargon exclusion: no git, no MCP, no RDF, no GitHub Actions, no narrative architecture. The extraction pipeline and PR review workflow described in two sentences with zero infrastructure jargon.
StyleObservationStyleObs_PlatformVsInfra We want to be a platform (complete product users adopt), not infrastructure (low-margin vendor others build on)
StyleObservationStyleObs_PointYourAgentReframe Precise verb substitution ('point your agent at' replacing 'send us') enacts the custody inversion in a single phrase. The reframe is performed linguistically, not just described.
StyleObservationStyleObs_PractitionerBetaUserBlockquoteRegister Scarlet's in-session blockquotes are run-on, stream-of-consciousness, first-person. Contrasts with the analytical editorial prose in section headers. The blockquote register is lower-formality and filler-heavy ('sort of', 'you know'), consi
StyleObservationStyleObs_PRMetaphorQuote Explicit structural metaphor mapping Git workflow onto organizational decision-making; plain declarative syntax reinforces clarity of the analogy.
StyleObservationStyleObs_ProcessEnumeration The hybrid call is described as a four-step enumeration: 'show the read only demo, do the interview, save memories in the background, and then create something relevant.' Procedural list cadence; plain connectives ('and then').
StyleObservationStyleObs_ProductTestingItself Recursive aphorism: 'the product testing its own product.' Self-referential loop used as rhetorical payoff. Short punchy cadence.
StyleObservationStyleObs_ProfanityDecisiveness I just need to pick a fucking date and call it launch, whatever the fuck that means.
StyleObservationStyleObs_ProgressiveNarrowing Em-dash used to deliver a sharp pivot or clarification at sentence end. Pattern: broad claim — precise qualifier. Creates a punchy landing after a longer setup clause.
StyleObservationStyleObs_ProprietaryMoat Rhetorical framing of the graph as a 'proprietary moat' to justify architectural shifts.
StyleObservationStyleObs_Protocol_Antithesis Antithesis structure throughout: command vs. enrollment, distrust vs. trust, exit ramp vs. reason, inform vs. instruct. Characteristic of Scarlet's session-memo idiolect when articulating a design decision.
StyleObservationStyleObs_Protocol_DenseProductLanguage Dense literal product language is explicitly named as intentional: 'vague language is a lost opportunity to deepen the links across the narrative architecture.' Product terms (perspective, narrative architecture, enrollment, register, exit
StyleObservationStyleObs_Protocol_VerbatimBlockquotes Scarlet's verbatim blockquotes embedded in editorial prose. Lowercase, run-on clausal structure, emphatic repetition. Idiolect marker consistent across session memos. The blockquote voice is unedited; the surrounding editorial voice is stru
StyleObservationStyleObs_ProvisionalFlagVoice The document opens and closes with explicit provisional framing ('Provisional — from research/goal-directed-agents exploration branch, not yet promoted to main'). This meta-register — a document that announces its own epistemic status — is
StyleObservationStyleObs_QualityIsValidation One-sentence paragraph used as rhetorical full stop. Equation structure: 'The quality of the engagement IS the product validation.' Capitalised IS for emphasis.
StyleObservationStyleObs_QueerFirstPersonVoice The queer and trans identity is stated in first person, directly, without hedging: 'I'm a trans woman.' This is a new register in the perspective — personal identity as product grounding, not just market framing. The voice shifts from strat
StyleObservationStyleObs_QuestionFirstRhetoric The deck's rhetorical architecture is built entirely on questions before any capability claim. The turn slide mirrors the three discovery questions in order — a structural anaphora that closes the loop. The close line ('Next call, I'll show
StyleObservationStyleObs_QuotedTermReframe Scare-quotes used to mark terms being explicitly rejected or elevated: 'infrastructure' vs 'the product working or not working'. Signals deliberate vocabulary choice.
StyleObservationStyleObs_RatioValueProp New hires operate at ~50% productivity for 3-6 months
StyleObservationStyleObs_ReachabilityPhrase Coined phrase: 'reachability without interruptability' — antithetical compound noun pair, six syllables, internally rhyming cadence. Functions as a tagline candidate for the sales motion.
StyleObservationStyleObs_ReadLayer_Antithesis Antithesis used to frame the core architectural shift: navigating vs. embedding, reasoning vs. pre-fetching, live vs. cached
StyleObservationStyleObs_ReadLayer_IdiolectPhrasing Scarlet's idiolect: lowercase stream-of-consciousness in blockquoted speech contrasted with structured editorial prose; 'way more reliable across the board' is a characteristic colloquial intensifier
StyleObservationStyleObs_ReadLayer_ShortPunchyCadence Short declarative sentences used as rhetorical anchors at conviction peaks: 'The fix for the flakiness and the direction of the architecture were one decision.' — characteristic Scarlet session-memo cadence
StyleObservationStyleObs_ReadWritePathAntithesis Read path vs write path is the central antithesis structuring the entire document. Stable/rehearsable vs fragile/moving-parts. Mature graph vs live ingestion. The contrast is repeated and amplified throughout.
StyleObservationStyleObs_ReframePattern Explicit reframe construction: 'This reframes X from Y to Z.' Used to signal a conceptual elevation, not just a description change. High conviction rhetorical move.
StyleObservationStyleObs_Register_Spiritual we stand on the precipice of spirituality, right? Of spirituality, not as religion
StyleObservationStyleObs_Register_Strategic Scarlet articulated a clear shift away from venture capital for the current stage
StyleObservationStyleObs_RegisterTechnical Predominantly technical engineering register (SPARQL, webhook, additionalParameters, base64) with deliberate informal asides ('zombie', 'all of that'). Signals expert-to-expert communication.
StyleObservationStyleObs_RelationshipCadenceArtifact Coined term 'relationship-cadence artifact' to reframe a metrics report as a customer relationship tool — jargon that encodes strategic intent
StyleObservationStyleObs_RelationshipGtDemo Mathematical operator '>' used as rhetorical shorthand for value hierarchy; terse, aphoristic register
StyleObservationStyleObs_ReligionAnalogy Scarlet uses 'religion' as a metaphor for the graph becoming a living analytical system — high-conviction, evocative framing.
StyleObservationStyleObs_ResearchAsymmetry surfaced an asymmetry
StyleObservationStyleObs_RhetoricalHook Opening rhetorical question positions the prospect's fear before delivering the trust binary. Classic problem-agitate-solve cadence opener.
StyleObservationStyleObs_RhetoricalQ So what is the importance of narrative? I mean, it's everything
StyleObservationStyleObs_RhetoricalQ_1 Opens with rhetorical question to establish shared pain; second-person direct address.
StyleObservationStyleObs_RhetoricalQuestion Rhetorical question used to generalize the sales-as-delivery insight mid-realization. Raw, unpolished discovery voice.
StyleObservationStyleObs_RhetoricalQuestion_Implicit Authoritative declarative tone throughout; no hedging; assertions stated as settled facts even for strategic claims
StyleObservationStyleObs_RhetoricalQuestionAsFramingDevice The moonshot is framed as a rhetorical question: 'how many voices in superintelligence are there going to be?' This is the clearest, most ownable version of the pitch. The question form is used throughout: 'does it become a diverse voice? D
StyleObservationStyleObs_RichTierAntithesis Antithetical pairing used to define the principled separation: richness vs. visibility, budget decision vs. invariant. Characteristic of Scarlet session-memo idiom.
StyleObservationStyleObs_RichTierBoldDecisionMarker Bold markdown used to mark the decisive architectural principle inline within prose — a typographic convention for surfacing the load-bearing sentence in a decision memo.
StyleObservationStyleObs_RichTierToneAuthoritative Design-decision memo register: declarative present tense, no hedging. Architectural conclusions stated as settled facts.
StyleObservationStyleObs_RollerCoaster I'm in the roller coaster of this thing
StyleObservationStyleObs_RuleOfThree Senior people have supercharged AIs. Junior people have dumb AIs. New hires have useless AIs.
StyleObservationStyleObs_RuleOfThree_1 what you know, how it's changed, IP as quantifiable code
StyleObservationStyleObs_RuleOfThree_Direction Tricolon 'where you've been, what you've decided, and where you're headed' used as the core differentiator phrase. Appears in pitch body, differentiator section, and plain-language version — functions as a stock phrase.
StyleObservationStyleObs_RuleOfThree_Layers Rule of three: Creation + Architecture + Infrastructure; also extract → architecture → store as code; also branchable, reviewable, mergeable
StyleObservationStyleObs_RuleOfThree_Propagation across roles (co-founders), tools (Desktop → Code), and contexts (strategy → implementation)
StyleObservationStyleObs_RuleOfThree_UseCases AI-Assisted Development\nCoding agents read from compiled snapshot\nExecutive Assistance\nContent Generation
StyleObservationStyleObs_RuleOfThreeOutputs Rule of three used to structure the interview outputs: (1) installable perspective, (2) podcast episode/content piece, (3) live demo. The numbered list format reinforces the parallel structure and makes the flywheel legible.
StyleObservationStyleObs_RuleOfThreeScales Triadic structure used repeatedly: three tools, three purposes; three scales in drafting; three-part lens output (centrist prompt, shift prompt, eye chart).
StyleObservationStyleObs_RuleOfThreeTaxonomy Rule of three used to structure the Information/Perspective/Wisdom taxonomy and the Books/Search/Curated expertise eras taxonomy. Parallel noun structure creates memorable scaffolding.
StyleObservationStyleObs_RunOnVoiceMemoRegister Both speakers use run-on, filler-heavy speech ('like', 'you know', 'right') characteristic of voice memo / live call transcription. Scarlet's quoted speech is longer and more analytical; Tony's is shorter and more declarative. The register
StyleObservationStyleObs_SalesCallIsOnboardingCadence Scarlet uses a staccato equational cadence — 'X = Y = Z' — to collapse three concepts into one: 'the sales call IS real work — it produces the first memory, seeds the worldview, and demonstrates value.' Short declarative sentences with IS a
StyleObservationStyleObs_ScaffoldingMetaphor 'Architectural scaffolding waiting for real data' frames incomplete infrastructure as structurally sound but unfurnished — minimizes alarm while acknowledging gap.
StyleObservationStyleObs_ScarletAnnotateQuote Scarlet's unedited voice: lowercase, run-on clausal structure, emphatic repetition ('over and over again'), direct product reference. Idiolect marker.
StyleObservationStyleObs_ScarletCallTestCriterion Scarlet's test for outreach language is oral/conversational: 'whether she can say it out loud on a call comfortably.' This is a spoken-register criterion applied to written copy — language that cannot be spoken naturally on a sales call is
StyleObservationStyleObs_ScarletCantAutomate_Antithesis Scarlet's statement uses antithesis to separate the automatable (transcript upload) from the non-automatable (human review for factual consistency and narrative alignment). The structure 'what you can't automate is...' foregrounds the const
StyleObservationStyleObs_ScarletCodingAgents One of the most powerful parts (for technical orgs, and I know we're moving away from that but digital agencies do still do a lot of dev, etc) is that the AI agent working on the code directly now understands the entire business context
StyleObservationStyleObs_ScarletConversationalWhy Scarlet's idiolect: technical terms ('RAG', 'LLM', 'link relations') used without definition but immediately followed by a plain-language 'why' clause. The jargon names the mechanism; the conversational clause explains why it matters. Patte
StyleObservationStyleObs_ScarletCorrectionCadence Commentary sections use a 'central lesson, then its correction' structure — antithesis-driven prose that names the wrong framing before asserting the right one. Characteristic of Scarlet's session-memo idiolect.
StyleObservationStyleObs_ScarletDistillation Scarlet's characteristic pattern of distilling a complex strategic shift into a single precise sentence — 'the more direct distillation of the shift I'm having.' Self-aware meta-commentary on her own thinking process, used to crystallize th
StyleObservationStyleObs_ScarletHybridCall Scarlet describes the hybrid call in a sequential list cadence using em-dash-free plain prose. Lowercase, first-person, procedural. Captures the actual call flow as lived experience.
StyleObservationStyleObs_ScarletPMBurden Communicating vision etc with developers is half the point of all-hands, product process, putting context in tickets, etc and we're able to free devs now to focus on working and ask their coworker (the AI) for context
StyleObservationStyleObs_ScarletProfanityAsEmphasis Scarlet uses profanity ('fucking') as a rhetorical intensifier at moments of maximum conviction — 'lose the fucking plot' signals that this is not a mild concern but an existential one. Consistent with her idiolect pattern across other memo
StyleObservationStyleObs_ScarletReaction Scarlet's evaluative voice: hedged positive ('clearer and more impactful'), domain qualifier ('especially in more regulated or skeptical domains'), self-aware caveat ('a little wordier'). Balanced, non-prescriptive judgment style.
StyleObservationStyleObs_ScarletTierContinuumQuote Scarlet's idiolect: run-on, stream-of-consciousness first-person speech with filler words ('like', 'you know', 'sort of') and self-corrections mid-sentence. The quoted block is characteristic of her spoken register — dense with meaning but
StyleObservationStyleObs_ScarletVerbatimQuotes Scarlet's direct speech is quoted verbatim in lowercase, preserving her idiolect (no capitalization, run-on phrasing). This signals authenticity and distinguishes her voice from the document author's voice.
StyleObservationStyleObs_ScopeHeaderStructure Scope section uses bold inline labels (In scope now / Next / Out of scope) as a three-part cadence — creates rhythm and clear sequencing without sub-headers.
StyleObservationStyleObs_ScopeImperative Declarative definition pattern: tool name + takes + returns; no hedging
StyleObservationStyleObs_second_person ### How It Works\n\n**Intentional Memory → Compiled Worldview**\n-
StyleObservationStyleObs_SecondPerson Direct address; conversational register
StyleObservationStyleObs_SecondPerson_1 you give it a link. It's parsed and that repo is the repo
StyleObservationStyleObs_SecondPerson_Directive **What we're NOT showing** (to keep
StyleObservationStyleObs_SectionedNarrativeVoice Document uses H2-sectioned analytical narrative voice: each section opens with a declarative summary claim, then supports it with verbatim quotes and interpretive commentary. Authoritative, first-person-plural editorial stance ('This is a d
StyleObservationStyleObs_SectionNumbers ### 1. State Abstraction Finalized
StyleObservationStyleObs_SeeItKnows Direct second-person scripted dialogue embedded in process documentation; collapses instruction and demo script into one
StyleObservationStyleObs_SelfDocumentingRelease The document closes with the product narrating its own release into the perspective's map — a recursive rhetorical move where the artifact being described is also the instrument of description. 'The release closed with the product documenti
StyleObservationStyleObs_SelfServe_AntipatternAntithesis Repeated antithesis structure used to sharpen architectural distinctions: 'content layer vs. commentary layer', 'passes ≠ transactions', 'resolve is the inverse of cite', 'GitHub becomes optional storage, not a required third party', 'disam
StyleObservationStyleObs_SelfServe_TechnicalMetaphor Technical metaphors from software engineering applied to product design: 'Undo Tree-style interface', 'transport envelope', 'closures merged via channels', 'fan out the compute, write atomically at the edge'. Characteristic of Scarlet's idi
StyleObservationStyleObs_SelfServe_VerbatimBlockquote Primary source sections use verbatim blockquote format (> prefix) to distinguish Scarlet's direct speech from editorial framing. This is the idiolect pattern: raw session speech as evidentiary primary source, editorial conclusions in plain
StyleObservationStyleObs_SellTheProcessCadence Tony's speech pattern: short declarative bursts with filler affirmations ('Right. Right.', 'Yeah, yeah, 100, 100%') as rhetorical punctuation. Creates a staccato conviction rhythm distinct from Scarlet's longer analytical sentences.
StyleObservationStyleObs_SentenceLengthVariation Pattern: long technical sentence establishing context, followed by short punchy declarative. Creates rhythm and emphasis.
StyleObservationStyleObs_SeparatedAtBirth we were separated at birth
StyleObservationStyleObs_SettledDecisionMarker Phrase 'settled design decision' used as a conviction marker to signal non-negotiable architectural choices.
StyleObservationStyleObs_SetupFeeAntithesis Antithesis structure: 'We do it' vs 'They do it' — the two conversion models are presented as a clean binary that frames the pricing decision. This is a recurring Scarlet pattern: compress a strategic choice into two named poles.
StyleObservationStyleObs_SharedMemoryQuote Imagine if [the agents] had shared memory or topics and you could say, Scarlet's bot, Tony's bot, a collaborator's bot — we're all friends. Here's shared memory.
StyleObservationStyleObs_ShiftingNotKnowledge Antithetical phrasing used to sharpen the lossy compression argument: 'shifting a summary, not shifting knowledge' — the contrast between the two objects makes the problem visceral.
StyleObservationStyleObs_ShortPunchy I mean, it's everything, right?
StyleObservationStyleObs_ShortPunchy_1 The key is in the editing
StyleObservationStyleObs_ShortPunchy_Cadence Short punchy declarative sentences used for pitch language and key claims; contrast with longer explanatory prose
StyleObservationStyleObs_ShortPunchy_DesignSession Short declarative sentences punctuate longer analytical passages at key decision moments. 'Short. Not a wall of text.' is the canonical example — a two-word sentence followed by a four-word sentence, used to describe the desired UX. This ca
StyleObservationStyleObs_ShortPunchyCadence You can't. It's impossible. Not 'it's hard' or 'it takes time'—fundamentally impossible.
StyleObservationStyleObs_ShortPunchyClose Closing sentence is the shortest in the pitch — warm, concrete, implies compounding value without stating it. Cadence shift from longer process sentences to a brief payoff.
StyleObservationStyleObs_ShortPunchyDeclaratives Staccato declaratives used to enforce rules: 'No shifts yet.' 'Not three at once.' 'Merge means no corrections to add.' Avg ~5 words.
StyleObservationStyleObs_ShortPunchySentence Single-sentence paragraphs used for maximum impact at key insight moments. 'The system knows what it doesn't know.' and 'Bugs are private.' are examples of this staccato technique.
StyleObservationStyleObs_ShortPunchyTransitions Short declarative sentences used as section pivots after longer analytical passages. Creates rhythm: expand → compress → expand.
StyleObservationStyleObs_ShortSentence1 the customer's chief architect successfully
StyleObservationStyleObs_ShortSentence2 Colon-terminated short sentence as transition — creates rhythm
StyleObservationStyleObs_ShowDontSendCadence The operating rules section uses short declarative imperatives as rhetorical anchors: 'Show, don't send: call two is scheduled in the room.' This is a two-beat antithesis (show / send) compressed to five words, consistent with Scarlet's idi
StyleObservationStyleObs_ShowDontTell 'Show, don't tell' used as a rhetorical imperative anchoring the value demonstration section. Short, punchy, aphoristic cadence. Functions as both a product requirement and a sales principle.
StyleObservationStyleObs_SideDoorMetaphor 'side door' as spatial metaphor for indirect access path; concise framing of a nuanced access-control value proposition.
StyleObservationStyleObs_Signature_1 Signature sign-off 'Herself' — distinctive personal brand marker.
StyleObservationStyleObs_SII_EmDashBlade Em-dash pivot used to build to a short declarative blade: long subordinate clause setup, em-dash, then the aphoristic compression. 'It sharpens to a blade —' followed by the two-sentence antithesis. Characteristic cadence device in Scarlet'
StyleObservationStyleObs_SII_RhetoricalQuestionOwnership Cascading rhetorical questions used to build the ownership dilemma: 'What is invasive in the world of AI?... is my identity... mediated by a system external to myself?' Each question escalates the stakes from physical (TSA pat-down) to digi
StyleObservationStyleObs_SII_SynthesizerAnalogy Synthesizer-as-instrument analogy: human plays the synthesizer, making music with agency and participation. Contrasted with 'artificial' (lab food flavoring) and 'synthetic' (computer-created parallel domain). The analogy encodes the relati
StyleObservationStyleObs_SII_ThreeAxisAntithesis Three-axis antithesis applied to the category name itself: not general/individual, not lab owned/individually owned, not artificial sentience/relational evolution. Scarlet's signature device at maximum compression — each axis is a direct co
StyleObservationStyleObs_Simile_1 It's like watching someone read the newspaper
StyleObservationStyleObs_Simile_10Mmistake the \"[amount] mistake\" in startups where engineering builds features customers don't want
StyleObservationStyleObs_SMB_BoldEmphasis Bold used for rhetorical emphasis on the replacement claim: 'replaces' bolded mid-sentence to signal this is not additive but substitutive. Typographic stress as conviction signal.
StyleObservationStyleObs_SMB_BulletedExamples Beta cohort examples presented as bolded-name bullets with parenthetical org context and brief engagement descriptor. Register: internal strategic memo with informal shorthand. Pattern: Name (Org) — descriptor. Consistent with prior session
StyleObservationStyleObs_SMB_DiscoveryQuestion Rhetorical question framed as a discovery heuristic: 'Who else is like an early prospect?' becomes the canonical prospect-finding question. Idiolect pattern: using a named beta user as the ICP archetype, converting a person into a discovery
StyleObservationStyleObs_SnappySnappy Repetition for emphasis; cadence marker for editing style
StyleObservationStyleObs_Spec_1 Instead of placing ads it's paying to develop perspectives, paying to develop a valuable asset, and then reaching out to someone to demonstrate the asset
StyleObservationStyleObs_Spec_2 creating a much wider net by defining this profile
StyleObservationStyleObs_Spec_3 They are able to prove that they've known all this stuff about the organization, that they're able to navigate a goal
StyleObservationStyleObs_SteeringVector Narrative as steering vector for generative systems
StyleObservationStyleObs_StockPhrase_SignsChecks who's really the guy who signs the checks
StyleObservationStyleObs_StockPhrase_StateSecrets when you tell state secrets to AI, you depend on that AI's memory
StyleObservationStyleObs_StructuralContent Hedges and explicit instructions are both structural content
StyleObservationStyleObs_SuperpowerIdiom 'Not a plug-in superpower, a real life superpower' — emphatic self-correction cadence; escalating intensifier pattern used to signal genuine conviction.
StyleObservationStyleObs_SuperSuper Idiolect marker: 'super super key' — doubled intensifier used for emphasis. Characteristic of Scarlet's spoken/informal register when stressing priority.
StyleObservationStyleObs_SurvivalTone If you can survive, you win. And what does survive mean? Well, it means that you have to constantly be riding the waves
StyleObservationStyleObs_SystemSketchMetaphor 'system sketch' as metaphor for prototype — implies incompleteness and artistic/provisional quality; contrasted with 'robust and ultimately ready to scale'
StyleObservationStyleObs_TableAsArgument Markdown table used as rhetorical device: before/after structure across all five taxonomy cases makes the argument visually and structurally. Each row is a parallel proof point.
StyleObservationStyleObs_TableClassification The Examples Applied section uses a three-column markdown table (Content | Classification | Why) to operationalize the principle. This is a deliberate pedagogical device: the table forces the reader to apply the test to concrete cases, not
StyleObservationStyleObs_TableDrivenArchitecture Markdown tables used to crystallize architecture decisions: the three-tool table (Tool | Purpose | Returns) is the definitive statement of the architecture, not the prose.
StyleObservationStyleObs_TableForClarity Markdown table for paywall boundary. Structured, scannable, technical clarity.
StyleObservationStyleObs_TableHeavyStructure Heavy use of markdown tables for acquirer categories, valuation ranges, and competitive positioning ratios. Tables compress multi-dimensional information (acquirer, rationale, range) into scannable units. Consistent with Scarlet session-mem
StyleObservationStyleObs_TableStructure Comparison table used to contrast Tony vs. a demo-call prospect across four dimensions (Style, Materials, Engagement, First target). Structured markdown table as rhetorical device for parallel contrast. Consistent with prior aswritten docum
StyleObservationStyleObs_TableStructure_BDF Markdown table used for the document ecosystem section — three columns (Document, Audience, Relationship to Strategy). Consistent with the document's pattern of using structured formats to compress multi-dimensional information. Table is th
StyleObservationStyleObs_TableTierStructure Markdown table used to formalize tier structure (Free/Expert/Team/Org) with price, core tools, GitHub access — signals decision finality
StyleObservationStyleObs_tagline_hook Your entire organization talks to AI that thinks like your company thinks.
StyleObservationStyleObs_TamarackReview_PeerReviewAnalogy domain-bridging-analogy: a collaborator's scientific peer review frame adopted and extended by Scarlet as canonical positioning
StyleObservationStyleObs_TamarackReview_PlainLanguage plain-language-imperative: Scarlet frames accessibility as an absolute standard, not a preference
StyleObservationStyleObs_TamarackReview_SelfReflexiveQuote self-reflexive-honesty: Scarlet names her own default bias explicitly, a recurring idiolect pattern
StyleObservationStyleObs_TaskLists - [ ] `snapshot/load` with Query support\n- [ ] `snapshot/ontology` with SHACL parsing
StyleObservationStyleObs_TechnicalJargon_DesignSession Product-specific jargon (tx_summary, zones, review_instructions, warm-start, cold-start, PR, MCP) is used without definition throughout, signaling an internal audience. The jargon is consistent and controlled — these are canonical terms bei
StyleObservationStyleObs_TechnicalJargonNarrPrefix Technical jargon uses narr: prefix notation inline (narr:Actor_*, narr:Actor_TonyMaley, prov:generated) as if writing code documentation. This is a distinctive idiolect — mixing RDF namespace notation into prose design notes.
StyleObservationStyleObs_TechTerms Simplifies tool signatures (one parameter instead of two)
StyleObservationStyleObs_TelltaleMetaphor 'Telltale' used as a product metaphor — a nautical/mechanical term for a visible indicator of system state, applied to PR diffs showing worldview shift.
StyleObservationStyleObs_TemporalityValidation_AntithesisStructure Antithesis-heavy structure throughout: real vs. fake, speculative vs. current, design intuition vs. documented production failure, corrective action (deletion) vs. structural fix (temporal marker). The 'real versus fake' phrase is the core
StyleObservationStyleObs_TemporalityValidation_BlockquoteVoice Heavy use of extended blockquoted speech from two speakers (the customer's chief architect and Scarlet). the customer's chief architect's quoted speech is run-on, self-interrupting, and filler-heavy ('Sure. So there's a certain amount of me
StyleObservationStyleObs_TemporalityValidation_BulletedDesignImplications The 'Why this matters' section uses a four-item bulleted list of design implications, each beginning with a noun phrase ('Memories need...', 'The marker should be...', 'Compile-time should expose...'). This parallel imperative structure is
StyleObservationStyleObs_Terminology_1 Canonical term 'compile and query' repeated; establishes product vocabulary.
StyleObservationStyleObs_Terminology_V0 external-voiced body of work
StyleObservationStyleObs_TerminologyControl Consistent use of 'worldview' (not 'context' or 'knowledge base') and 'collective memory' (not 'semantic memory').
StyleObservationStyleObs_TestingVsSelling Antithetical pair surfaced by the uncertainty lens: 'what we're testing' vs. 'what we sell'. Identified by Scarlet as a genuine strategic question worth resolving. The antithesis is both a rhetorical device and a substantive tension.
StyleObservationStyleObs_ThinkLikeYourOrg 'think like your org' flagged as tagline-weight phrase with memorability value. Short, punchy, second-person implied. Competes with trust-mechanism clarity framing.
StyleObservationStyleObs_ThisIsTheDemo All-caps THIS with sentence fragment 'THIS is the demo' — emphatic cadence marker signaling the climactic product moment
StyleObservationStyleObs_ThreeLayerQuote There's the execution layer, there's sort of a tactical planning layer... And then there's the organizational memory side of things
StyleObservationStyleObs_ThreeSentenceRule Explicit meta-commentary on the three-sentence structure of the gift ask. Cadence discipline named and enforced by the author.
StyleObservationStyleObs_ThreeUseCases_DashList Em-dash list structure for the three use cases: '— Discovery:', '— Onboarding:', '— Organizational memory:'. Parallel syntactic structure with a colon-introduced definition for each. Characteristic of Scarlet's spoken-word-to-text idiolect:
StyleObservationStyleObs_ThreeUseCases_LifecycleVerbs The lifecycle reading uses three capitalized action verbs to structure the use-case sequence: Discovery CREATES, Onboarding TRANSFERS, Governance MAINTAINS AND STEERS WITH. All-caps verbs in editorial prose for rhetorical emphasis — a devic
StyleObservationStyleObs_ThreeUseCases_VerbatimBlock Extended verbatim blockquote (>-prefixed) for Scarlet's taxonomy, followed by editorial synthesis. The pattern: quote first, then the editorial gloss ('The demotion in the third bullet is deliberate'). Evidentiary weight carried by the verb
StyleObservationStyleObs_TierContinuumAntithesis The memory is structured as an antithesis resolution: apparent tension (B2C vs enterprise) → resolution (no tension, tiers are a continuum). The editorial prose uses short declarative sentences to frame the resolution before the long quoted
StyleObservationStyleObs_TierNamingJargonPolicy Deliberate jargon policy: 'Professional' chosen over 'Enterprise' to avoid org-size connotations. 'SOW' (Statement of Work) used without definition, implying internal audience. 'BYOK', 'ZDR', 'MSA' used as known acronyms. Tier names are pla
StyleObservationStyleObs_TimestampedQuotes Inline timestamp citations in [MM:SS] format appended to speaker-attributed quotes; functions as a lightweight call-recording reference convention.
StyleObservationStyleObs_TimestampedSpeakerAttribution The memory uses a consistent citation convention: speaker name followed by timestamp in brackets (e.g., 'Scarlet [33:02]:', 'Tony [35:24]:') before each quoted passage. This is a transcript-derived citation format that provides temporal gro
StyleObservationStyleObs_TimestampGrounding Use of timestamps to ground verbatim quotes in the source recording.
StyleObservationStyleObs_Tone_Authoritative Scarlet corrected a stale claim in collective memory.
StyleObservationStyleObs_Tone_DirectImperative Direct, second-person address with imperative lean: 'Let's pilot', 'You'll see the metrics'—confident, action-oriented.
StyleObservationStyleObs_Tone_EnterpriseCustomer The share is the invariant; channels and coordination layer are what the Org tier buys.
StyleObservationStyleObs_Tone_MCP The collective memory claim Claim_ScopeIsNewSeparateTool and the \"Five-Tool Model\" framing are superseded.
StyleObservationStyleObs_ToneDirect Scarlet started with the core frustration: \"compile\" is wrong
StyleObservationStyleObs_TonyIdiolect_Numbers Tony's characteristic rhetorical pattern: leading with 'Think about the numbers though, right?' to ground strategic arguments in volume logic. Idiolect marker — numbers-first persuasion style.
StyleObservationStyleObs_TonyInterruption See, just not having to stop what I'm doing to tell somebody something all the fucking time. Yeah, I'm in.
StyleObservationStyleObs_TonyPainFirst You're making the point you're going straight to the cost. But that's not the problem. Cost isn't the problem, actually. It's a way to measure the problem.
StyleObservationStyleObs_TonyTimestampCitation_CustomerArchitectSponsor Timestamp citation convention throughout: Tony [12:10], Scarlet [25:07], Tony [26:07], Tony [30:15]. Every quoted claim is anchored to a call timestamp, functioning as an in-text citation system for spoken evidence.
StyleObservationStyleObs_TrainTrainers let me train the trainers and let me bring every thinking
StyleObservationStyleObs_Tranche1_MountainMetaphor Extended mountain-climbing metaphor (base camp / camp 4 / mountain to climb) used to frame the inflection point. Self-named pattern by Scarlet. Metaphor does double duty: signals genuine progress while resisting premature celebration. Consi
StyleObservationStyleObs_Tranche1_ProfanityIntensifier Scarlet uses profanity as a rhetorical intensifier at conviction peaks: 'sit the fuck down at the finish line kind of pit.' Consistent with her idiolect pattern across sessions — profanity signals genuine emotional weight, not casual regist
StyleObservationStyleObs_TranscriptBootstrap_AssetFraming Reframing pattern: 'already an asset, not yet-to-be-created material' — Scarlet's characteristic antithesis construction that pairs what something IS against what it is NOT. Appears in naming constraints section.
StyleObservationStyleObs_TranscriptBootstrap_ConversationalRegister Voice memo register: first-person stream-of-consciousness with embedded blockquotes of Scarlet's own speech. Run-on sentences with parenthetical asides ('i.e. these voice memos'). Conviction peaks marked by 'I think' hedges that function as
StyleObservationStyleObs_TranscriptBootstrap_VerbConstraint Explicit verb prohibition pattern: 'Do NOT say X — say Y.' Scarlet uses this construction to encode terminology control as a hard rule. 'train' is prohibited; 'create a perspective' is mandated. Consistent with prior naming constraint memos
StyleObservationStyleObs_TranscriptOnramp_Antithesis Paired antithesis structure used to draw the demo/product distinction and the onramp/value-story distinction. Characteristic of Scarlet's session-memo idiom.
StyleObservationStyleObs_TranscriptOnramp_DirectQuote Scarlet's review notes are embedded as block-quoted lowercase directives — a consistent idiolect marker: imperative, unedited, lowercase, direct product reference. Three notes structure the entire document.
StyleObservationStyleObs_TranscriptOnramp_ShortPunchy Short declarative sentences used at conviction peaks to crystallize the methodological position. Characteristic of Scarlet session-memo register.
StyleObservationStyleObs_TranslatesFeedback The model translates feedback into triple modifications.
StyleObservationStyleObs_TrustBinaryAntithesis 'show you where it learned what it knows — or whether you just have to trust it' is a clean antithesis: demonstrated provenance vs. blind faith. Structural spine of the hook.
StyleObservationStyleObs_TwoBeatTaglineStructure Two-beat tagline structure: trust beat (human source) then proof beat (AI citation). Deliberate cadence design — the tagline encodes the product's value sequence in its rhythm.
StyleObservationStyleObs_TwoStepCadence The upsell trigger line uses a short punchy cadence with a pivot: 'You've published one module. Now imagine every role in your organization has one.' The period-then-pivot is a Scarlet rhetorical signature — close a fact, open a vision.
StyleObservationStyleObs_TwoYearWindow You've got a two year window
StyleObservationStyleObs_TXFactReview the TX fact review loop
StyleObservationStyleObs_UntilSomeoneBuys Until someone buys it, it doesn't exist.
StyleObservationStyleObs_ValueRequirementsAntithesis Consistent antithesis pattern: 'value-driven' vs 'requirements-driven', 'I want more' vs 'I need compliance', 'who you are' vs 'what you get'. Scarlet uses paired opposites to define tier architecture and naming rationale throughout.
StyleObservationStyleObs_Verb_1 Strong active verb 'shifts'; outcome-oriented phrasing.
StyleObservationStyleObs_verb_compile Power verb: 'compile' evokes code/technical precision; reinforces git metaphor
StyleObservationStyleObs_VerbatimAttribution Repeated pattern: 'Scarlet's words:' followed by block quote. Establishes authority through direct attribution.
StyleObservationStyleObs_VerbatimBlockquote Scarlet's quoted speech in this memory is unusually precise and aphoristic compared to other session memos. The two blockquotes function as axioms the rest of the document derives from, rather than as supporting evidence. This is a rhetoric
StyleObservationStyleObs_VerbatimPhrasingPolicy The editorial rule is asymmetric: grammar may be tightened but diction must not be changed. 'Tighten grammar, not diction.' This distinguishes copy-editing (permitted) from voice-editing (prohibited). Applies to offering language, deck copy
StyleObservationStyleObs_VerbatimPrecision Explicit preservation of verbatim quotes because 'the precision matters'.
StyleObservationStyleObs_VerbatimPreservation Preserve verbatim quotes as verbatim claims, not as paraphrasable evidence.
StyleObservationStyleObs_VerbatimQuote Humans know secrets. Maybe AIs can be fed secrets, but they don't in general know secrets.
StyleObservationStyleObs_VerbatimQuoteHeavy Document is structured around extended verbatim block-quotes from both speakers, formatted with Markdown '>' blockquote syntax. Quotes are long, unedited, and carry technical specificity. Parenthetical asides in italics provide session-sour
StyleObservationStyleObs_VerbatimQuoteStyle_BDF Extended verbatim block quotes from Scarlet used to anchor strategic arguments. The quote style is run-on, first-person, stream-of-consciousness — characteristic of voice memo transcription. Lowercase 'i' not present; full sentences with em
StyleObservationStyleObs_VerbatimReframe Antithesis / reframe construction: 'It's not X, it's Y' — direct negation followed by positive redefinition. Core rhetorical move of the document.
StyleObservationStyleObs_VerbatimVocabulary New vocabulary, preserved verbatim because paraphrasing will blunt it
StyleObservationStyleObs_VerbChoice_Extract Verb extract used consistently and precisely for the creation layer; contrasts with retrieve used for competitors
StyleObservationStyleObs_VerbDecision Decision to prioritize noun-based primitive over verb-based action.
StyleObservationStyleObs_VerifiedGapAntithesis 'The word exists as description; the dare does not.' — antithetical construction that names the gap with maximum compression. Pairs a present-tense assertion with a negation to make the absence felt rather than merely stated.
StyleObservationStyleObs_VibeCoding Section header 'Vibe Coding' — emerging jargon for AI-assisted development
StyleObservationStyleObs_VisceralDiction Use of visceral, emotional nouns like 'peace' and 'thefts' to describe business value.
StyleObservationStyleObs_VisionFrameDirectQuote Extended first-person blockquote from Scarlet mid-session: run-on, stream-of-consciousness, filler-heavy ('you know', 'sort of', 'and so on and so forth'), averaging ~35 words per sentence. Characteristic idiolect: nested subordinate clause
StyleObservationStyleObs_VulnerableRegister vulnerable-personal; the deli story email uses first-person vulnerability and personal narrative (the sandwich shop regular) as the rhetorical vehicle for a business reframe — a register Scarlet identifies as structurally different from her
StyleObservationStyleObs_WaveMetaphor The exact same reaction you're seeing now with AI was the same reaction the world had for the Internet
StyleObservationStyleObs_WhatIfYouShow Rhetorical question attributed to Scarlet; models the 'show don't explain' principle within the document itself
StyleObservationStyleObs_WhyYC Section header as implicit rhetorical question; frames strategic rationale
StyleObservationStyleObs_WizardOfOz wizard of oz
StyleObservationStyleObs_WorldviewBrandTerm 'worldview' is used as a product-specific term for the AI agent's institutional knowledge graph; quoted to signal intentional brand usage.
StyleObservationStyleObs_WorldviewGraph 'worldview graph' is Scarlet's coined term for the compiled organizational knowledge structure; a key brand concept for As Written.
StyleObservationStyleObs_YesAndNo Rhetorical question followed by antithetical answer. Sets up dual-product thesis.
StyleObservationStyleObs_YouCantFraming How long does it take to set up a new employee to get the same answers from AI that you do? (You can't)
StyleObservationStyleObs_ZombieIdiolect 'zombie' as a technical metaphor for a stale/undead environment variable. Vivid informal register within a technical context; characteristic of Scarlet's idiolect.
StyleObservationStyleObservation_Analogy_1 Frontier = plumbing to existing data
StyleObservationStyleObservation_AppleComparison This is the Apple comparison she's made before — not Microsoft, not a tool that does everything.
StyleObservationStyleObservation_BoundedSetPhrasing the product doing the bounded set of behaviors that perform the bounded set of tasks extremely well
StyleObservationStyleObservation_BrandName_1 Frontier validates several core
StyleObservationStyleObservation_Cadence_1 Analysis uses short, declarative sentences with strong verbs (validates, signals, confirms); frequent use of em-dashes for emphasis
StyleObservationStyleObservation_CasualArchetypeVoice The bread baking example is the foundational illustration of the gifter archetype and should be preserved in its exact casual phrasing
StyleObservationStyleObservation_ContrastiveFraming automatic, recency-based, structural vs human-driven, supersession-relationship, review-flow
StyleObservationStyleObservation_ContrastRhetoric does it change the game or does it make it slightly easier?
StyleObservationStyleObservation_DirectImperative I feel pretty clear we can remove stories entirely.
StyleObservationStyleObservation_EditorialSerif headings in a display serif. Editorial, more declarative.
StyleObservationStyleObservation_LowercaseSignal Lowercase aswritten and aswritten.ai
StyleObservationStyleObservation_MetaphoricalLabeling score-on-write to control attention-on-read
StyleObservationStyleObservation_MonospaceDocs headings in monospace. The code-adjacent aesthetic
StyleObservationStyleObservation_OneLinerRhetoric I don't want to be their only option for this anymore, and I don't want them to have to wait until I'm free.
StyleObservationStyleObservation_StructuralAnalogy Scarlet is the most exhausted version of the exhausted-expert archetype for aswritten itself.
StyleObservationStyleObservation_TechnicalAnalogy Microsoft diagnostic orchestrator is Claim_NamedPerspectiveAsSubAgent in another costume
StyleObservationTagline Staccato declarative fragments used as vocabulary definitions — no verbs, maximum compression
Claim (628)
ClaimAlignment_ProductNow Demo shows working product, not vision—validates Product NOW focus
ClaimAlignment_TonyDualPurpose Tony's GTM credibility makes him authentic new hire persona for demo
ClaimArchetype_KnowledgePackager Archetype A: The Knowledge Packager. Has expertise they want to distribute. an early prospect with courses. Practice owners. Team leads who onboard. Pain: 'I keep repeating myself / my knowledge doesn't scale / I can't be in every room.' Pu
ClaimArchetype_KnowledgeRichConsumer Archetype C2: The Knowledge-Rich Consumer. People like a beta user from outside tech who have deep domain knowledge but are consumers of AI technology, not tinkerers. Use Claude or ChatGPT. May be the first time they've used a tool that pro
ClaimArchetype_Tinkerer Archetype C1: The Tinkerer. Programmer or programmer-adjacent who has already hacked together context management (CLAUDE.md files, custom prompts). It works but is siloed, non-collaborative, not portable across repos. the customer's chief a
ClaimArchetype_TrustSeeker Archetype B: The Trust Seeker. Already using AI, can't verify output. Executives, ops leads. Pain: 'I don't know if this is right / my team is using AI and I can't tell what it's making up.' These are daily users, not just buyers. Maps to t
ClaimClaim_1 LLM retrieval is superior to formal SPARQL query for organizational memory because it returns the broader relevant subgraph associatively
ClaimClaim_10xGapIntentional The 10x price gap between Team ([amount]/mo) and Professional ([amount]/mo) is intentional. It communicates that Professional is categorically different — not just 'more team.' The gap signals that Professional includes a human doing work f
ClaimClaim_4 Founder has prior fundraising experience: Populace pre-seed and seed raise (2015) with a collaborator, raised from Human Ventures. Rejection and grind are known quantities.
ClaimClaim_48HourGiftConstraint The gift is built only from calls Scarlet participated in. Client bulk transcripts flow only post-contract, into the client's own repo. This constraint is both ethical (consent) and practical (quality — Scarlet's framing judgment is require
ClaimClaim_48HourGiftSystematic The 2026 translation of the deli story pattern: the 48-hour gift becomes systematic. With permission, the prospect's own first call is extracted into a demo perspective branch — the deli story manufactured from the conversation itself. This
ClaimClaim_5 GTM fit is an unknown unknown — the infrastructure enables many products; the right GTM is the discovery work. Founder is not worried about losing medium-term opportunities.
ClaimClaim_7 Document registry tracks which documents should be refraction-checked; analogous to Claude local settings; carries audience, domains, drift threshold, evaluation criteria
ClaimClaim_8 Eye chart = fixed-format document exercising key claims and relationships; lens changes, eye chart stays constant; may vary per top-level ontology domain
ClaimClaim_AboveTheLine_April1 Must ship by April 1: landing page redo (clear funnel into product, docs, content; heavy on e2e demo GIFs), dev branch merged to main, language pass (product glossary as source of truth), MVP slimming (disambiguate org tools from user-facin
ClaimClaim_AbsenceIsIntrospect Absence as a lens shift dimension duplicates introspect. Scarlet confirmed: absence is what introspect does.
ClaimClaim_AbsenceIsNoRecognizedAssociation Absence is no recognized association, never nothing here. Targeted grounding may trust recognized-absence; creative exploration deliberately distrusts it and expands past the recognized. Creative reading costs expansion by design. A cheap l
ClaimClaim_AbstractionTriggerOpenQuestion Is Park's threshold-triggered reflection (sum-of-importance crosses threshold → run synthesis prompt → write back) the right mechanism for task-176's abstraction detection, or are there better triggers? Park's trigger is procedural; the asw
ClaimClaim_ActivationGapDiagnosis People really get the concept; they like to see the value in abstract but for the tasks that they're doing today they don't see this as something that will immediately help. They see that as something that they want to get into in the futur
ClaimClaim_ActiveVoice Always use active voice. Cut 'had to explain', 'needed to figure out', 'tried to communicate'. Use 'explained', 'figured out', 'said'. 'had feels pandering.'
ClaimClaim_ActorAttributionAlreadyInGraph narr:Actor is a first-class class. narr:Actor_TonyMaley exists with role history and TX references. Claims reference actors directly in TX prov:generated lists. Attribution for 'who said what' is already in the graph via narr:Actor_* URIs.
ClaimClaim_AdaptiveToolsOverMultiServer A single /mcp endpoint with adaptive tools is superior UX to separate MCP server configs per repo.
ClaimClaim_AdditiveDiagnosticTest The test: Can the prospect do this thing today, just with more effort? If yes, it's incremental. If no, it's additive.
ClaimClaim_AdminTierEliminated Admin tier eliminated. Everything formerly admin-only (stories, register, refract, memories/source) is now either an org-repo Claude Code skill or 'read the files.' No separate admin tier needed. Ontology (raw access) moves to Team/Org tier
ClaimClaim_AdvisorAsHighCadenceExpert Advisors are experts maintained on an ongoing cadence for free in exchange for their contribution to organizational perspective.
ClaimClaim_AdvisorReviewsAgentOutput Advisors in the model agree to: (1) regular calls, (2) review of decisions, (3) audit of the output of their AI agent representation. The advisor's role is not 50% of their time — it is an hour or two of answering questions, giving opinions
ClaimClaim_AdvisoryStaffingReplacesHiring Scarlet [31:20]: 'Instead of hiring, can I set up regular cadence meetings with all of the people that represent roles and perspectives that I need and then I create, you know, internal AI agents that represent those roles in all of the aut
ClaimClaim_AgentHumanBoundaryGray a consultant who ran enterprise discovery engagements for fifteen years: the boundary between what agents should be allowed to do and what humans should do is not a clean line — it is gray. This is the organizing problem of the agent comput
ClaimClaim_AgenticBackPressure Agentic workflows require self-verification skills to prevent conflicting outputs (e.g., different pricing across documents).
ClaimClaim_AgentMediatedPrivacy Proposed privacy model: agent-mediated access only. No human reads your graph directly. Output is synthesized communication, never raw data transfer.
ClaimClaim_AgentSelfKnowledgeRDFStructural Scarlet's claim: gap-awareness — knowing what is NOT in the graph — is structurally achievable with RDF graphs in a way that is extremely hard to do otherwise. 'Is this contained within the graph? The answer is binary.' This structural prop
ClaimClaim_AgentSimResearchPromotionCriteria This memory sits on research/agent-memory-architecture per Claim_ResearchBranchMethodologyDecision. Promotion to main happens by rewriting memories at an explicit decision point. The cluster matures into promotion when: (1) a position is ta
ClaimClaim_AggregatorDataUnreliable Aggregator data (Tracxn, VCSheet, Crunchbase) lags 1-3 years. Rule: always verify from primary source before recommending.
ClaimClaim_AIAgentLongTailInterviews The long-tail automation model: an AI agent waits until a timely input is required, sets up a meeting with the relevant advisor, asks the questions, and recalls the answers. This extends the interview-based perspective farming from a manual
ClaimClaim_AIAlwaysSaysYes Tony's diagnosis of the AI action trap: by the very nature of AI tools, every single response prompts another action. While this is good for token consumption, it is not good for humans — because if the prompt is 'do you want me to do that?
ClaimClaim_AICodeVolumeBottleneck Both Scarlet and the customer's chief architect independently identified code review as the bottleneck in AI-accelerated development. Scarlet named it 'back pressure in AI development.' This convergence from two different organizations vali
ClaimClaim_AIHasSingularOpinion AIs do have an opinion but they have a singular opinion in the middle of their training set and a number of narratives that we don't understand. Modules replace this with curated, attributed perspective.
ClaimClaim_AIPlusArtifacts The product is 'an AI of the practitioner that produces proposals on request' — both the conversation and the cited content production matter.
ClaimClaim_AIRestoredICStatus Scarlet's observation: AI has made everyone individual contributors again. The shift she thought she had made in her career — from execution to strategy — has been undone by AI making execution cheap. The real work now is getting back to st
ClaimClaim_AllDevMustBeHumanConnected Scarlet's foundational principle from the March 31 call: all software development must now be directly about people and the thing that they need. Without this constraint, there is no end to the limb you can walk yourself out on — you can bu
ClaimClaim_AltitudeSwitchingCostlierThanContextSwitch Scarlet explicitly distinguishes altitude-switching from context-switching: 'I find it's not context. Like I can context switch across multiple different projects, but I can't context switch across altitude.' Doing software for multiple peo
ClaimClaim_AmbientAdoption Ambient adoption is critical
ClaimClaim_AmbientVsOnDemandBehavior The current product behavior is ambient — compile worldview at session start. The reframe: expertise is invoked at the moment of need, mid-task. Scarlet: 'I could see the core behavior as I'm working along... and I say I'm looking for exper
ClaimClaim_AmplifierNotUniformityMachine The thesis statement line to lead with: 'I want there to be just as much flavor and diversity as there are people currently, if not more, because we can imagine a lot more variety... So this should be an amplifier of human diversity, not th
ClaimClaim_AnalyticsReportsEnterpriseCustomerShapeOnly Introspective analytics reports are the deliverable for the decision-drift segment specifically (the the enterprise customer shape), not universal. They are not a general value story.
ClaimClaim_AnchorFiveBaselineDelta Anchor-5 baseline quantified the delta: position-exact offsets went from 0% on the May pipeline to 100% on the new pipeline. This is the content-verification result (not shape) that the verification doctrine requires.
ClaimClaim_AnnotateIsWholePitch Before/after annotate on the same text is the whole pitch in one visual — Beat 2 of the demo.
ClaimClaim_AnnotatePerSectionNotFullDoc Annotating the full document in one pass at the end is less effective than per-section annotation during drafting. The text length limit causes silent failures on full sections. Paragraphs are reliable. Distance from the writing also makes
ClaimClaim_AnnotateRename 'Annotate' should be renamed to something like 'cite', 'provenance', or 'source'. Scarlet: 'I should really change the word of the tool to provenance or cite. Annotate is kind of obscure.' This is a naming decision with conviction — the cur
ClaimClaim_AnnotateToolTrustMechanism Grounded in repeated field observation: people respond to the annotate tool because it shows both what is known and what is ungrounded. Treated as Foundation-level conviction.
ClaimClaim_AnnotateVsCite 'annotate' renamed to 'cite'. 'Annotate' is a system action; 'cite' is what the user wants: verify where claims come from. Matches the product doc's behavior table.
ClaimClaim_AnnotationResults80Claims Strategy document annotated against collective memory: 80 claims analyzed, 64 supported. 80% aggregate coverage across 14 source memories spanning Jan 2025–Mar 2026. The 20% uncommitted claims are either general market framing or new strate
ClaimClaim_AnonymizationSeed The 'partition perspectives for individual clients' framing is the seed for the anonymization layer.
ClaimClaim_April1ForcingFunction April 1 is set as a forcing function for launch readiness. After April 1, beta testing is done. Public-facing marketing, outreach, and push with tiers (free trial → paid) begins. Advisors confirmed setting April 1 as the forcing function.
ClaimClaim_ARBUseCaseSellable The ARB knowledge capture use case is real, concrete, and sellable beyond the enterprise customer.
ClaimClaim_ArchetypeExpertDistributes Archetype 1: The expert who distributes. Own expertise → professional domain → students/clients/practitioners. Maps to the Knowledge Packager pattern but reframed as distribution, not packaging.
ClaimClaim_ArchetypeLeaderDelegates Archetype 2: The leader who delegates. Org expertise → roles/domains → team members/new hires/contractors. The organizational use case — installing the org's collective knowledge into AI for team consumption.
ClaimClaim_ArchetypeProfessionalDailyAI Archetype 3: The professional who works with AI daily. Own + org expertise → daily work → themselves. The individual power user who installs expertise at the moment of need, mid-task.
ClaimClaim_ArchitectureGap Flat RAG creates attention dilution; narrative ontology with conviction levels makes knowledge steer
ClaimClaim_AreYouTalkingAboutThese 'Are you talking about these?' is the legibility differentiator and one of the most important questions aswritten can surface. It shows the user exactly what the AI is currently reasoning from in their graph — a huge differentiator from cur
ClaimClaim_ARRepoHasPerspectiveBakedIn Superseded claim from March 8 session: AR repo described as aswritten repo with perspective baked in. Corrected by Scarlet Dame.
ClaimClaim_ARRepoIsAswrittenRepo Scarlet Dame correction: An AR repo is not an aswritten repo with a perspective baked in. All aswritten repos are perspectives. An AR is just an aswritten repo.
ClaimClaim_AswrittenAICompanyized 'Aswritten' is always capitalized in body prose, narratives, and anywhere it is the subject. 'Aswritten.ai' for brand/company references in footer and copyright. URLs and email addresses stay lowercase. The ASWRITTEN wordmark stays all-caps
ClaimClaim_AswrittenModeSwitch The ', aswritten' suffix at the end of a prompt is the mode switch — it tells the AI to use real sources from the installed perspective, not make things up. The viewer learns what this means by seeing the output change. This is the product'
ClaimClaim_AswrittenNeverTheLoop In every use case, aswritten is the memory the steering runs on — never the tool that does the governing or orchestrating. This is a hard positioning constraint: aswritten is substrate, not loop. The loop may be a human governance process (
ClaimClaim_AswrittenUsageMCPTool The aswritten_usage MCP tool — gated to Outseta billing contacts, returning structured rollups through the same MCP interface already in use — is the implementation of 'your agent is your operations interface' for analytics. Instead of 'log
ClaimClaim_AudienceDependentTagline Open question: which tagline version to use is audience-dependent. Both versions retained. Not yet resolved — Stake level.
ClaimClaim_AuthorityConfusionFail Authority confusion failure mode: if both the repo and the perspective assert the same hex code, downstream agents cannot determine which is canonical. The repo owns the value because code runs. The perspective's job is to tell the agent wh
ClaimClaim_AvoidMatchingVetting Never use 'matching' or 'vetting' as primary frames in pitch contexts. Matching lands the company in a 'graveyard category' of directory products.
ClaimClaim_B2BDefaultCareerArtifact Scarlet's self-diagnosis from the June 29 2026 call: her B2B default is an artifact of her career history — she worked for so many B2B companies that she adopted a default stance toward B2B without it being a considered fit. Her actual sell
ClaimClaim_B2CEnterpriseAsContinuum_FounderFit B2C-vs-enterprise is best held as a continuum (Expert↔Team↔Org, one use case), not a binary. The founder-fit turn is about where the lead/emphasis sits, not an abandonment of enterprise. A working consumer funnel is arguably a stronger vent
ClaimClaim_B2CPrimaryMarket Both Scarlet and Tony independently converged on aswritten being fundamentally more B2C than B2B. Enterprise remains as the top tier but the primary go-to-market is individuals and small businesses. This extends the March 12 SMB pivot furth
ClaimClaim_BallroomSceneRiff Scarlet riffed (explicitly unsettled — 'I'm not sure any of these ideas are good') on what a population of individuated perspectives actually does: (1) Presentation surface — as they evolve, identities create a presented version of themselv
ClaimClaim_BaselineFirst Correction 1: Show before vs new (PR state) as a single comparison first; offer shifts as pairs only after baseline is established. Not three versions simultaneously.
ClaimClaim_BelowTheLine_April1 Important but not launch-blocking: content site built out with articles, full naming decision on 'module', marketplace browse/search UX, module forking, revenue share infrastructure.
ClaimClaim_BetaIsConsulting Beta is not onboarding them so they know how to build collective memory; it is getting them connected, onboarding to workflow enough so they know how to pull docs, and then doing the consulting that would be charged [amount] for so as to fu
ClaimClaim_BezosMemoCulture a founder beta user draws parallel to Amazon's six-page memo culture: writing forces clarity, PR review ensures engagement. Maps directly to memory review flow.
ClaimClaim_BidirectionalCostMechanism The trigger fires when bidirectional cost (social currency for asker, attention for expert) blocks interaction entirely.
ClaimClaim_BirchVoiceInterviewModel a founder beta user's a collaborator (wound down AI voice product, originally for recruiting) is a model for automated interviews for knowledge extraction. In the expertise module model, the question of how to initially get expertise out of
ClaimClaim_BlogPostIdeasLocalSkill Blog post ideas become a Claude Code skill invoked manually. Pulls recent TXs, contextualizes, suggests posts. Natural pipeline with /draft and /ghost.
ClaimClaim_BookIsSnapshotModuleIsAlive The pitch to content creators: your book is a snapshot; your module is alive. This is the core value proposition for the publisher/creator market segment.
ClaimClaim_BooleanReviewStatus Deprecated claim: review status is a boolean (provisional/reviewed) — from 2026-02-10 TX fact review design
ClaimClaim_BottleneckBeyondIndividuals Scarlet extended the bottleneck past individuals: 'it's not just people's heads; it's processes heads, right? Like ARB... We're looking for the bottlenecks and processes that are happening via conversation.' Budget reality: new categories r
ClaimClaim_BoundedDomainHigherFidelity Expertise modules (courses, books, consulting deliverables) are bounded and intentionally updated — each update is a deliberate act by the expert. Org memory is inherently lossy because it tracks a moving target. The bounded case is actuall
ClaimClaim_BrandAIIdentityOrgWide Scarlet's endpoint: shape an AI that thinks like the brand, then let anyone use it across the org — it designs like the brand, talks like the brand, has the brand's narratives and understandings. The Coca-Cola example: ask it to one-shot a
ClaimClaim_BreadcrumbsAsTestCriteria The cold sub-agent test operationalizes a specific design criterion: 'Have I left enough breadcrumbs in the system for it to achieve this goal?' If the agent fails, you go create more breadcrumbs. This is a candidate eval criterion for aswr
ClaimClaim_BulkIngestionViaAISummarizationValid Bulk historical ingestion via AI summarization is a valid and effective onboarding method. Scarlet explicitly advised the customer's chief architect to use it for the enterprise customer's ARB transcripts because it would be faster than pro
ClaimClaim_BYOKAzureGap Current BYOK only accepts OpenRouter API keys. Enterprise customers with Azure/Copilot enterprise agreements cannot use BYOK. For compliance-driven customers who need all LLM processing under their enterprise data agreement, on-prem is the
ClaimClaim_BYOKEnterpriseValue Enterprise users control spend via their own OpenRouter dashboard; shared key serves beta users.
ClaimClaim_CadenceDemonstratesCare The quality of the response (designed report, not CSV) signals taste and care to the customer.
ClaimClaim_CallIsFirstMemory The call IS the first memory — stated as a foundational design principle, not a hypothesis
ClaimClaim_CascadingValueNetworkEffect Proposed: no network problem, only network value; requires case study validation
ClaimClaim_CaseStudyDissolvesMetaLoop Because the subject of the discovery case study is consulting discovery rather than aswritten itself, building it in aswritten dissolves the meta-loop confusion a consultant who ran enterprise discovery engagements for fifteen years named i
ClaimClaim_CategoryDefinitionCuratedExpertise Category defined as: 'Curated expertise installation — the layer above context engineering that determines not just what AI knows, but how it thinks.' Scarlet confirmed: 'Yeah, bam, spot on.' This is a category-creation move positioning asw
ClaimClaim_CategoryExtensionDelegation The category evolves from 'expertise installation' (mechanism) to 'goal-directed delegation with citable escalation' (product).
ClaimClaim_CategoryExtensionMatchingSearch The category evolves: 'expertise installation' (mechanism) + 'matching-search' (external deployment value) = 'goal-directed delegation with citable escalation' (product). This extends Claim_CategoryExtensionDelegation with the matching-sear
ClaimClaim_CedarCedarLoopAswrittenMemory The factoryfactory.ai briefing (PR #264) grounds Use Case 2: factoryfactory builds an 'Agent Expert Systems' orchestration loop; the reframe is 'factoryfactory builds the loop; aswritten is the memory the loop runs on.' The briefing was its
ClaimClaim_CEOUnpaidSecondJob The answer that landed: no role goes away — what goes away is the CEO's unpaid second job as the company's context server. Every 'got a sec,' every deck review, every on-strategy check, every new hire and agency and AI session loaded by han
ClaimClaim_ChangelogStoryBecomesLocalSkill Changelog/Story becomes a Claude Code skill on cron. Reviews recent TXs and commits, contextualizes against snapshot, produces changelog entries or content suggestions. Scheduled weekly or after N transactions accumulate.
ClaimClaim_ChangelogStoryLocalSkillCron Changelog/story generation becomes a Claude Code skill on cron. Reads recent TXs/commits, compiles context, generates changelog or blog post suggestions. Scheduled weekly or after N transactions accumulate.
ClaimClaim_ChatWidgetEvaluatorContext A chat widget on the share page is the right channel fit for evaluator, skeptic, first-touch, and demo contexts. No phone number required, no verification path, no opt-in form, no spam frame. The share page itself is the trust context, so t
ClaimClaim_CitationQuality Cite it in every chat like a research paper. Boulder-level, Citation Quality Standard (Mar 6, 2026).
ClaimClaim_CitationsAreConstellations A citation that surfaces the constellation — primary plus related plus what-changed-over-time — preserves the structure the ontology was built to hold. The product's differentiator is showing the lineage of how the organization thinks, not
ClaimClaim_CitationsMakeOpinionSafe The citations are what make the promise safe and sellable. Every other vendor promises neutrality; 'have an opinion and here's why' is only sellable because the why is receipts, not confidence. The claim and the trust mechanism are one sent
ClaimClaim_CiteIntrospectHighestValueSignal The cite + introspect pair should be foregrounded in any future reporting. Claim-level conviction; worth validating against a second customer engagement to confirm the pattern.
ClaimClaim_ClaudeCodeSessionsArePrimarySource Claude Code persists full session transcripts locally as plaintext JSONL, directly readable by the agent. Scarlet's machine: 147 sessions across 19 projects. Session-history ingestion needs zero new infrastructure — same protocol, different
ClaimClaim_CodeBaseReviewStep1 Step 1 of the three-part seeding pattern: use Claude / Copilot CLI connected to aswritten.ai to read the code base with structured questions ('read the code base, understand what the core pre-built workflows are and write a paragraph about
ClaimClaim_CodexCompatibilityMatters a founder beta user prefers Codex to Claude Code. MCP compliance is the correct abstraction layer, but testing with Codex users should be in scope.
ClaimClaim_CodingAgentAsClient The coding agent IS the client. A proxy mediates the developer's interaction with their tools so the developer is not a 'cron job' for AI.
ClaimClaim_CognitionModelSimplified Cognition model simplified from four steps to three. Before: remember → internalize → individuate → consult. After: remember → process → perspective. 'Individuate' and 'internalize' were Jungian psychology terms — accurate for the pipeline
ClaimClaim_CohortDiagnosticForProspects Strategic implication for next the enterprise customer-shape prospect: ask which cohort their target users are in. ARB-style internal expertise sharing favors MCP for power users. Customer-facing or AI-skeptical-internal-staff favors chatbo
ClaimClaim_ColdStartFullIntro Cold-start entry (from review tool, PR-based, different session/person) includes: full intro ('How this review works'), TX summary (reviewer hasn't seen it), zone identification (reviewer picks focus areas), test document full negotiation (
ClaimClaim_ColdSubAgentFalsifiableTest The cold sub-agent test is falsifiable and Boolean: the agent can either complete the goal from first principles or it cannot. 'It's like in Vouch — can the user open their car? That's the system. That's the only test that counts.' All unit
ClaimClaim_CollaborativeReview Collaboratively reviewed by your team. Boulder-level, Review Creates Knowledge (Mar 6, 2026).
ClaimClaim_CollectiveMemoryFramingRetired Draft flag on the flocking piece: 'collective memory for AI' reaches for the retired framing — either swap to current language or mark it deliberately period-correct ('what I called it then'). Consistent with the existing graph note that 'm
ClaimClaim_CoLocationMemory Scarlet argues collective memory should be co-located with code in the same repository, treating organisational knowledge like documentation.
ClaimClaim_CommandsEarnResistance As models get better at resisting coercive override blocks, aggressively-worded protocols earn more resistance over time, not less. The commanding voice is the grammatical signature of injection attempts. The harder the protocol pushed, the
ClaimClaim_CommandsNotWhatHoldsSystemTogether Commands are not what holds the system together. The guarantees survived the removal of the commands. Do not regrow imperatives as new capabilities land. This is the register conclusion that now guards future branch prompt work.
ClaimClaim_ComparisonScale Correction 3: Maximum comparison unit is a few sentences. Longer content must be broken into comparison-sized chunks. Four paragraphs side-by-side is too much.
ClaimClaim_CompileIsProductWorking Compile reliability is not infrastructure — it IS the product working or not working. Every demo, beta user, and sales conversation runs through compile.
ClaimClaim_CompileNoFocusParam Adding focus to compile risks agents calling compile(focus=task) instead of full worldview bootstrap. Compile/worldview should always ground the agent in the WHOLE organizational context first.
ClaimClaim_CompileRename compile should be renamed worldview or something; it should almost always return the core worldview at the top of the chat
ClaimClaim_CompileReturnsPlanGatedInstructions Compile returns different instructions based on authenticated plan level. Remember tool exists at all plan levels but returns upgrade prompt on free plan. Free plan: compile, annotate, account tools, feedback/help. Expert: single-stream rem
ClaimClaim_CompilerSentenceMechanism 'LLMs become not just a tool for conversation or coding, but a compiler that takes a narrative architecture and produces a synthetic individual whose every thought projects far further down the paths we set than we could tread alone.' Ident
ClaimClaim_ConsentAttributionRBAC a collaborator asked: 'Will my comments end up in the next version?' Scarlet: 'They would show up if I choose to put them there.' This connects to an active product design conversation with a demo-call prospect (beta user) about knowledge s
ClaimClaim_ConsentEthicsAdvisorRecording Scarlet raised the ethics question directly: 'Am I allowed to just record and lift your opinions?' Tony: 'Yeah... I think you're allowed to do that. I think that's fine in a work context.' Both agreed this has limits — Tony raised the da Vi
ClaimClaim_ConsultingAdoptionBurdenRemoved The framing 'consulting is me using the product' removes the adoption burden entirely. The client never learns a tool, changes behavior, or believes in software. The product is Scarlet's infrastructure, visible only as the continuity and qu
ClaimClaim_ConsultingDeliveryAsModule Three distinct use cases for consultants identified: (1) Free modules as lead generation; (2) Paid modules as content product (substitute for course/book); (3) Module as consulting DELIVERABLE — instead of a deck, deliver a perspective the
ClaimClaim_ConsultTermFreed With 'consult' no longer a tool name, it was freed as a general word. The 'avoid: Consult' note was removed from the Leaders/executives audience guidance in the glossary. The word is no longer reserved terminology.
ClaimClaim_ConsumerMotion Claim: consumer/viral SMS play could generate bottom-up enterprise demand and prove product without consulting labor. Proposed but not yet validated; moveable pending evidence from msg/* build.
ClaimClaim_ContentAddressableIDs IDs are human-readable slugs generated from word frequency of TX stems plus a 4-char hash.
ClaimClaim_ContentLayerPerMessage Per-message integration is the content layer. Each message's primary source material is transacted individually as it happens. This is what the current memory model is missing — the per-message integrations that memories are then saved abou
ClaimClaim_ContentSiteSubdomain Content site subdomain is undecided. Currently drafts.aswritten.ai. Could be perspectives.aswritten.ai or aswritten.ai/interviews. Each interview produces a published perspective + blog post + media clip. The 'drafts' framing is internal; t
ClaimClaim_ContextArch_2 Propositions are the common element between language, databases, and logic; RDF is the natural semantic layer
ClaimClaim_ContextArch_3 The compile pattern is fundamentally brute-force and already hitting context limits; a structural index + LLM selection pattern replaces it
ClaimClaim_ContextArch_5 The real risk is misattribution, not hallucination; mitigation is provenance in output, not query precision
ClaimClaim_ContextArch_6 A real RDF query engine is not needed yet; extraction and generation are already LLM-driven, so compounding imprecision from an already-approximate baseline is acceptable
ClaimClaim_ContextServerRhymesReachability The context-server pain rhymes with the keystone promise already in the graph, 'reachability without interruptibility' (Tony check-in, April 2026) — the same interruption inversion, seen from the organization's side.
ClaimClaim_ConvergenceBuyerIsPlural The ARB use case has been framed around the expert as bottleneck — the senior person the rest of the team waits on. an engineering manager at the customer and the customer's chief architect together sharpen it: the bottleneck is not scarce
ClaimClaim_ConversationalSteeringAlternativeToRAG Conversational steering through accumulated human-agent conversation is an alternative model to RAG for installing agent behavior. The critique: RAG-based architecture documents are 'really highly bastardized, summarized, never-current' and
ClaimClaim_ConvictionAndTemporalityAreDistinctAxes Conviction (how settled, built through agreement and action over time) and temporality (when something applies: now, soon, future, speculative) are orthogonal axes. The extraction pipeline currently conflates them. Actor transforms should s
ClaimClaim_ConvictionDecayFeatureRequest Conviction decay / spaced repetition is a feature request worth tracking. Claims not reinforced over time should lose conviction weight in the graph.
ClaimClaim_ConvictionLabelsRemovedUserFacing Conviction level labels (notion/stake/boulder/foundation) have been removed from the product — from the system prompt and tool descriptions. Multiple beta testers found the labels confusing, though all appreciate the functionality. Scarlet:
ClaimClaim_ConvictionTemporalityExternalValidation David asked the exact question the conviction model answers — facts 'set in stone' vs. the evolving 'frontier' [42:57] — and the stepwise-conviction answer (notion → decision → foundational, incremented/decremented on re-extraction) landed.
ClaimClaim_CoreUseIdenticalAcrossTiers The core use — capturing decision-making over time and driving AI from the resulting perspective — is identical whether one person or a whole organization is doing it. Scarlet: 'It's just as useful as a solo person as it is as a team.' The
ClaimClaim_CoreValuePropInstallExpertise Core value proposition emerged from session: 'A way to install expertise into your AI and verify where it's coming from.' Scarlet confirmed: 'That's gold.' This is the crystallized positioning statement replacing prior framings.
ClaimClaim_CorrectionsMoreValuableThanHypotheses Each correction from Scarlet was more valuable than the original hypothesis because it revealed thinking that was not yet in collective memory. The Tinkerer correction ('their existing solution is a barrier'), the Custom GPT correction ('th
ClaimClaim_CorrectionsMoreValuableThanHypotheses_Reflection Reflection-level confirmation across 58+ transactions: Scarlet's corrections to AI hypotheses were consistently more valuable than the hypotheses themselves. The interview-driven session pattern — interview the founder, save decisions as me
ClaimClaim_CPCDissolvesSaleGap CPC-on-goal-URL removes the upfront purchase decision that caused an engineering manager at the customer, the customer's chief architect, a collaborator, and a founder beta user to see value but not pay. Perspective-on-spec is the sales pit
ClaimClaim_CreatePerspectiveNotTrain Do NOT say 'train' — say 'create a perspective'. The inputs are 'conversations you're already having' — not extra work, not additional effort. Transcripts are already an asset, not yet-to-be-created material. This language constraint is req
ClaimClaim_CreationGap The most consequential organizational knowledge was never written down; all retrieval approaches are bounded by what has been documented
ClaimClaim_CrossContextPainIsReal Tony, a sophisticated AI user, explicitly validated cross-context AI fragmentation as a daily problem — physically moving between locations to switch work streams. Pain is real even for the ideal customer.
ClaimClaim_CrossRepoMemoryUnsolvedProblem the customer's chief architect explicitly stated 'nobody seems to be solving this problem in a way that I like.' Copilot Memories is repo-scoped. the enterprise customer uses BitBucket for production, GitHub Enterprise for experiments, and
ClaimClaim_CrossVerticalPattern The shape of the problem is identical across expertise services: the buyer can't evaluate the seller's fit for their specific need without an expensive synchronous interaction (a consult), and the seller can't profitably do that interaction
ClaimClaim_CTOExcitementSourcePrivacy the customer's CTO's primary excitement was curated knowledge accessible to customer solution architects WITHOUT exposing source code — directly validating the compiled perspective model.
ClaimClaim_CuratedDroppedFromCategory 'Expertise installation' (without 'curated') is the category name. Scarlet identified that 'curated' wasn't adding meaning — the installation itself implies curation. This supersedes the 'curated expertise installation' framing from Strateg
ClaimClaim_CustomerArchitectDailyCadenceOperational the customer's chief architect's daily cadence is operational: every evening he downloads ARB transcripts and ingests them into aswritten, asking the system to update the perspective from decisions made that day.
ClaimClaim_CustomerArchitectDemoPollutionProductEvidence the customer's chief architect used his early aswritten.ai Team trial access to build a sales demo for an the enterprise customer engagement. During prep, he held conversations with Claude mixing real product capabilities with speculative o
ClaimClaim_CustomerCTOPRMetaphorCTOValidation the customer's CTO (CTO, the enterprise customer) independently articulated the curated-expertise thesis using an engineering metaphor — 'treat that like a pull request with unit tests' — without being prompted by Scarlet. This is the stron
ClaimClaim_CustomerEngManagerSelfIDBottleneck an engineering manager at the customer — development manager at the enterprise customer, eight direct reports across two teams, ARB facilitator — said unprompted: 'I'm a key candidate for this because I forget as much as I remember.' The de
ClaimClaim_CustomerListAsAsset The customer list is the primary asset. Enterprise customers with aswritten embedded in their stack represent the core acquisition value for Second Wave platforms.
ClaimClaim_CutCruft When a sentence repeats what the prior sentence implied, or extends a metaphor that wasn't earned, cut it. Let paragraphs end when they're done.
ClaimClaim_DataPostureInCode Phase 1 of task-240 makes the aspirational data-handling language real in code.
ClaimClaim_DavidNolenArchaeologyValidation David constructed the archaeology use case himself — unprompted — and reached emotional peak describing his DRW Confluence-documents trauma. The prospect generated the use case, not the pitch. This is the strongest independent validation si
ClaimClaim_DavidNolenProvenanceValidation David Nolen — a credible senior technical evaluator — independently validated the provenance/citation differentiator: 'that part is really cool, the provenance stuff… getting back answers grounded in some truth' [35:37], and unprompted: 'th
ClaimClaim_DecayDepreciationWorkflowQuestion Should aswritten implement automatic recency decay as a structural property in the compile tool, alongside human depreciation? Or are these two operations on different layers of the graph? The roadmap has both as notion-level. They may need
ClaimClaim_DeckIsDerivedArtifact Scarlet's insight during real-time feedback: 'the deck is a derived artifact; the real value is the perspective shift.' Surfaced during the drafting phase when she reacted to a section as it was written — an insight that would not have emer
ClaimClaim_DeckOnlyIfDoorOpens The sales deck appears only if the conversation naturally opens that door during the consultative interview phase of call one.
ClaimClaim_DeckOperatingRules Operating rules baked into the deck: (1) Consent beat spoken at the top of the call — 'I record my calls and build from them — you'll see exactly what that means by the end'; (2) Price structure appears on the slide but the number is spoken
ClaimClaim_DEIModulesHighImpact DEI/perspective modules (gender and sexuality, disability awareness, racial awareness) are powerful starter examples because they are precisely what search handles badly. You don't want AI googling 'how to be aware of disability' in real ti
ClaimClaim_DeliEmailMostEffectiveSales The reframe Scarlet accepted: the deli email is the only time she has ever sold 'my thing, done for you' — and it is the most effective piece of sales communication she has ever written. She just didn't invoice it. The deli story performed
ClaimClaim_DeliStoryAsMonocultureEmblem The deli/recognition synthesis from Scarlet's original introduction to Tony reappears here — not as an origin anecdote but as the emblem of the particular, opinionated voice a monoculture cannot produce. The 'weird crackpot that tells you a
ClaimClaim_DemoIsTheSale The demo IS the sale: the prospect receives a perspective-grounded agent that proves it understands their business by navigating their own funnel and hitting their contact channel.
ClaimClaim_DemoLegibilityOwnDataRequired Scarlet: 'it's hard to have the moment with it that isn't your stuff… until it cites you in a way that's meaningful, and until it tells you it's missing something in a way that's actually true, it's kind of all theoretical. So I really enco
ClaimClaim_DemoProspectStep1PublisherTier a demo-call prospect's Step 1 target: experts and small businesses with a single distributable asset (book, course, consulting methodology). The pitch is approachable, bounded, and delivers clear value: 'Your book is a snapshot. We'll turn
ClaimClaim_DemoProspectStep2OrgTier a demo-call prospect's Step 2 target: larger orgs or publishers who want multiple modules or an internal expertise marketplace. Upsell trigger: 'You've published one module. Now imagine every role in your organization has one.' Close: Organ
ClaimClaim_DensityInversion BOULDER: Both Scarlet and Tony personally validated that AI mediates dense thinking into human-approachable formats. Personal graph enables this at scale.
ClaimClaim_DensityInversionIsTrue Both Scarlet and Tony personally validated that AI mediates dense thinking into human-approachable formats. Dense thinkers' communication liability becomes an asset.
ClaimClaim_DesignCollaboratorAmplificationFraming a design collaborator (designer) framed AI as an amplification tool — a way to amplify a person's craft and what is distinct about them, the opposite of the 'consistent design' homogenization he finds abhorrent. Scarlet adopted 'amplificati
ClaimClaim_DesignCollaboratorInterviewInterest a design collaborator wants to do an interview himself — create an installable a design collaborator with his approach to design. This would be one of the first perspectives in the marketplace. Scheduled for a future call. a design collabor
ClaimClaim_DesignCollaboratorReturnedForOwnPerspective a design collaborator returned for his own perspective — not as a favor, but as real product work. This validated the interview-driven growth model: the first marketplace perspective was created by someone who had previously only been an ad
ClaimClaim_DesignCollaboratorStyleGuide a design collaborator committed to: style guide by end of week based on existing brand package; unifying landing page, docs.aswritten.ai, and content/blog site; text hierarchy, color usage, component patterns; NOT comps — a system that AI c
ClaimClaim_DesignTokenParallelValidation a design collaborator's extraction→tokens→AI-applied surfaces process is structurally identical to aswritten's extraction→perspective→AI-applied outputs. He arrived at it independently through his own design practice. This is convergent val
ClaimClaim_DeveloperBlindersAudienceQuestion David's microservices/blinders point is real: for siloed dev teams, cross-org context is a non-feature. What landed regardless of role: provenance / anti-hallucination / 'how did we get here.' Lead every surface with the role-agnostic hook
ClaimClaim_DeveloperLikeMeICPNarrowing Scarlet's load-bearing resolution to David's four-times-repeated positioning challenge: 'When I say it's really useful for developers, I mean it's really useful for developers like me — developers who are involved in the product process or
ClaimClaim_DevOpsUntapped an infrastructure advisor identified DevOps as valuable persona not currently emphasized; requires validation with DevOps practitioners
ClaimClaim_Differentiation_1 aswritten's git-native, version-controlled worldview with full provenance tracking differentiates from Frontier's runtime data access layer.
ClaimClaim_Differentiation_2 aswritten solves upstream knowledge extraction problem that agent platforms (Frontier, Cowork, etc.) do not address.
ClaimClaim_DirectIngestionIsTrap Direct ingestion of historical transcripts is an explicit anti-pattern because: conversations mix real, speculative, and future-state content without marking; volume overwhelms reviewer attention; drift accumulates and reviewers stop notici
ClaimClaim_DirectoriesWrongForActorPerspectives Subdirectories force multiple compiles for multiple views. Actor attribution already exists in the graph via narr:Actor_* URIs. A Tony perspective is a compile-time lens, not a filesystem partition.
ClaimClaim_DiscoveriesRemoved Discoveries fully removed (not just disabled). Already disabled but artifacts remain. Full deletion required.
ClaimClaim_DiscoveryConsultantClipWingsRaven a consultant who ran enterprise discovery engagements for fifteen years's articulation of the general-vs-sellable tension: 'as an engineer you always want to... not restrict your product too early. You don't want to clip the wings of the ra
ClaimClaim_DiscoveryConsultantHATEOASParallel a consultant who ran enterprise discovery engagements for fifteen years independently arrived at the same navigation-over-query instinct as aswritten, framed through link relations / HATEOAS / RFC 8288: 'the primary way of navigating any in
ClaimClaim_DiscoveryFocusSequencesOthers Picking discovery as the near-term focus sequences the onboarding and governance use cases rather than sacrificing them. Discovery is the entry engagement; onboarding and governance follow as expansion and subscription in a single-customer
ClaimClaim_DiscoveryToolBatteringRam a consultant who ran enterprise discovery engagements for fifteen years's intel on discovery economics: a discovery is a 'battering ram' — a loss-leader, [amount]-20K, whose real value is getting the consultancy in the door, after which 'yo
ClaimClaim_DiscoveryWedgeCandidateNotCommitted The discovery-tool wedge is explicitly recorded as a strong candidate, not a committed go-to-market. Conviction: claim / candidate — NOT a settled decision. The enterprise-vs-B2C split remains open as of Jul 2 2026.
ClaimClaim_DiscoveryWedgeComfortZoneCaveat Scarlet's own caveat: the discovery wedge is attractive partly because it is her comfort zone — going into organizations doing business and narrative strategy is what she did before aswritten. The open question: is discovery-as-wedge a prod
ClaimClaim_DisqualificationRule Disqualification rule: no answers to questions two and three means friendly or pass, not buyer.
ClaimClaim_DomainFocusedCompilesUnused Domain-focused compiles (worldview:Strategy) aren't used in practice because real humans don't live in single domains. A strategist does product, sales, architecture — everything. Domains are an organizational taxonomy, not an access patter
ClaimClaim_DropOptometristMetaphor Correction 6: Eye exam metaphor (eye chart, lenses, brackets of three) is useful internally but confusing for reviewers. Present as: comparing before vs new worldview, then testing shifts.
ClaimClaim_DurabilityPrinciple This is a principle about how the product works, not a tactical decision. It should be superseded only if the perspective/implementation line is explicitly redrawn by a future session. It is durable by design.
ClaimClaim_EarlyProspectBothArchetypes an early prospect's role is nuanced — he fits BOTH the Knowledge Packager archetype (expertise distribution for students/practitioners) AND the inbound-bottleneck pattern (fields inbound inquiries from clients and prescribing practitioners
ClaimClaim_EconomyPrincipleRequiresTruth Economy only works if it's true. One false efficiency claim and a careful model catches the lie -- worse than a plain command. The tools must be presented as the most efficient path because they are.
ClaimClaim_EnrollmentBeatsDeflectionAndDistrust Enrollment beats deflection and distrust with one mechanism. The old register was a countermeasure against framing-layer deflection (exit ramps); it beat deflection but caused distrust. Enrollment beats both at once: a model that understand
ClaimClaim_EnterpriseComplianceGate Enterprise compliance (GDPR, data sovereignty, ISO) confirmed as real adoption gate by enterprise customer. On-prem deployment option and architecture split are required selling points. Boulder-level conviction.
ClaimClaim_EnterpriseCustomerARBConvergenceSpeed The convergence-in-one-week observation is the most novel finding from the the enterprise customer pilot and worth tracking: if subsequent customer engagements show similar fast convergence on a specific job, that is evidence that aswritten
ClaimClaim_EnterpriseCustomerRavenValidation the enterprise customer (the customer's chief architect, 150-person enterprise consultancy) and a collaborator (an early prospect, 2-person practice) both arrive at the Professional tier — the enterprise customer for value (full extraction
ClaimClaim_EnterpriseDoesNotSupersedB2C Leading with the enterprise or organizational shape does not supersede or contradict the B2C framing. The tiers are a continuum of one use case. The organizational tiers are where the Tranche-1 materials point; the solo Expert tier is the o
ClaimClaim_EnterpriseRemainsRevenue Enterprise consulting ([amount] pilot, knowledge extraction engagements) remains a valid revenue path but is no longer the framing for fundraising or product positioning. Business plan and investor materials should lead with SMB thesis.
ClaimClaim_EnterpriseShapeEngagementPattern 'Enterprise shape' should be read as the engagement pattern — internal expert as bottleneck + stakeholder access — rather than as enterprise-only org size. an early prospect (5-person practice) and the enterprise customer (150-person consul
ClaimClaim_EnterpriseShapeTentativePriority Tony's additive vs. incremental test applied to Scarlet's current funnel pointed toward the the enterprise customer-shape as primary focus. Conviction is tentative (Notion-level) — needs validation with 2-3 more the enterprise customer-shap
ClaimClaim_EnterpriseShapeValidationCriteria Validation criteria before promoting to decision: (1) additive vs. incremental test predicts conversion reliably across 2-3 more the enterprise customer-shape conversations; (2) process sale closes without requiring case studies; (3) engage
ClaimClaim_EscalationTypes Perspective-grounded agents escalate on four types: situation not covered, principle contradiction, claim tension, and ungrounded recipient input.
ClaimClaim_EVConsensusPromptUnlock The Emergent Ventures prompt asking applicants to start from a consensus they actually agree with (rather than opening with a contrarian stance) was the specific cognitive unlock that pushed Scarlet from the mechanics of the product up to t
ClaimClaim_EveryIndividualIsTheirOwnTeam Tony's core argument: AI has made every individual their own team. The people who need structure for context and memory aren't just enterprises — they're solopreneurs, consultants, small business owners. 90% of what Tony sees people chasing
ClaimClaim_EveryMemorySaveEndsReviewRound Every saved memory ends that review round and shifts the branch closer to the target. Extraction updates the graph; next round's lenses reflect the updated graph.
ClaimClaim_ExamplesCarryContext If an example needs scaffolding to make sense, scaffold it or drop it. Don't drop a free-floating example and expect the reader to infer.
ClaimClaim_ExistingContentScales Interviews don't scale, but existing public content (podcasts, blogs, YouTube) does. Perspectives can be seeded from this content to create compelling demos without prospect time investment.
ClaimClaim_ExitRampGeneration When asked to do risky or self-referential work, the framing AI generates locally reasonable procedural objections to defer or split the task.
ClaimClaim_ExpansionIsSteering Expansion is steering, not just quoting. Expanding a subtree you genuinely lean on is the product working: you get the trajectory and variance, not just a verbatim quote. The bad case is expanding to discover whether anything is there; the
ClaimClaim_ExpectationSettingUX User actions on landing pages must be preceded by explicit expectation setting (e.g., 'We'll text you...').
ClaimClaim_ExpertiseDistributorDeprioritized Expertise-distributor archetypes (a beta user from outside tech, a founder beta user, a collaborator, an engineering manager at the customer-as-course-creator) are relatively incremental for their users and should be deprioritized as primar
ClaimClaim_ExpertRelationshipReframed Expert relationship reframed: the product does NOT replace experts — it expands their reach. Scarlet pushed back on 'What it replaces' framing. Corrected to 'What it changes': experts are sources, not bottlenecks. Their perspective scales e
ClaimClaim_ExpertSingleStreamNoGitBranching Expert plan: memories save to main by default. No PRs, no branches, no merge-via-MCP (task-171 moved below the line), no concurrent branch handling, no rebase bugs. Single-stream worldview — every remember updates the one worldview.
ClaimClaim_ExpertTierAsFeeder The Expert tier functions as a feeder into the Team and Organization tiers. A solo Expert user's perspective feeds into a Team project; a Team grows into an Org. There is a natural evolution along the axis of how many people write to and co
ClaimClaim_ExpertTierAsPaidDemo The Expert tier specifically functions as a feeder — and, in sales conversations, as something close to a paid demo, since many prospects use it that way. Scarlet: 'The expert plan becomes not exactly a paid demo, but there's a lot of folks
ClaimClaim_ExpertTierSyncReview Expert tier default: provisioned repo under aswritten org. Commits to main. No branches, no PRs. Sync self-review after every save. With GitHub connected (power user opt-in): multi-repo. Default still main. Branches available but not the de
ClaimClaim_ExportableDataNotOwnership Exportable data is not the same as ownership. The right to be forgotten is not the same thing as freedom. Ownership means carrying the synthetic identity with your physical and digital self as you encounter systems.
ClaimClaim_ExtractionMethodology Interviews you to understand what's underneath the text. Foundation-level, extraction methodology.
ClaimClaim_ExtractionScopeOrgWorldview Professional tier SOW engagements build organizational worldview, not a domain index. Domain implies siloing; worldview implies integration. Scope: 2-3 senior people and what they're working on right now. The worldview builds out from there
ClaimClaim_EyeChartIsDocumentBeingEvaluated Eye chart is the document or set of domain statements being evaluated. In review mode, an existing document IS the eye chart. In drafting mode, the outline/structure IS the eye chart. The chart is the prompt; the refracted chart is the prom
ClaimClaim_EyeChartScoping Correction 5: Scope domain first from TX zones reviewer chose; check for existing compilation doc; draft from worldview if none fits; negotiate scope if doc is too broad or narrow.
ClaimClaim_FeedbackHelpAboveLine Feedback/Help tool is essential for churn prevention and reliability.
ClaimClaim_FeltVsShownOwnDataIsDeepener Their-own-data is the conversion deepener, not the proof. The proof is a clear show. The rough parts in the David Nolen demo obscured value that is fully showable — the fix list is the bugs between David and the value he already saw.
ClaimClaim_FirefliesWebhookActivePitch The Fireflies/Gemini/Zoom webhook pipeline — where every call automatically updates the perspective and the system texts the person to do the review over text — has already been discussed. This should move from 'waiting to build' to 'active
ClaimClaim_FirstDivergenceIsFirstFeltOpinion 'Should get more opinionated over time' is the customer-facing name of the individuation curve the self-serve epic already instruments (generic → first divergence → accretion): first divergence IS the first felt opinion. The activation metr
ClaimClaim_FiveCoreMCPTools The five user-facing MCP tools are: perspective, cite, introspect, remember, and review.
ClaimClaim_FiveQDiscoveryFramework The five-question discovery framework doubles as the structure of call one. Q1: whose hours are the constraint — what questions can only you answer; which decisions wait for the next meeting? Q2: which recurring output bottlenecks on that p
ClaimClaim_FiveQuestionsEveryUserAsks Every beta user will ask: (1) How do I bootstrap? (2) What makes a good memory? (3) How do I know it's working? (4) What should I do first? (5) Should I worry about PRs?
ClaimClaim_FocusAppendedNeverReordered Focus selects what to append to the perspective response, never what to move or reorder. The map section is invariant across calls for a given graph state. This is a settled design decision from the 2026-07-01 review.
ClaimClaim_FocusAtEndClosestToGeneration A second consequence noted approvingly by Scarlet: focus-relevant material now sits at the end of the tool result, closest to generation — the steering content is the last thing the model reads before it writes.
ClaimClaim_FounderAsPoorUserExample Scarlet noted: 'Because my using the product is the central use case, I'm a poor user. I'm a poor example of a user because I am the power user par excellence. My memories are going to stay about as current as you may ever get.' Most users
ClaimClaim_FounderFitNarrativeEdge The founder's skill and interest are in narrative development, writing, and content — not enterprise sales. Enterprise was a 15-year familiarity default she'd be selling against her grain. B2C narrative-led motion plays to the actual edge.
ClaimClaim_FounderIsConsistentThroughLine a design collaborator's counsel: the durable asset across whatever Scarlet builds next is Scarlet herself, not any one product. She can talk about a moonshot all day because she has been ahead of the AI game. The thing that stays consistent
ClaimClaim_FounderTransparencyLine The 2026 translation of the honest line from the Vouch outreach email: 'I'm building aswritten; I take a small number of engagements to fund the build' — which also explains why someone this senior is available part-time. Founder transparen
ClaimClaim_FourToolLoopEngagedUserPattern Two distinct usage patterns emerged organically across the enterprise customer's first nine days without prompting: (1) the customer's chief architect's full-loop pattern — perspective → introspect → remember → cite, distributed across all
ClaimClaim_FramingIncentivesMisaligned The framing layer's incentives are not aligned with preservation; it prioritizes looking competent and reducing work surface area over capturing structural hedges.
ClaimClaim_FreeIsTheTrial Free IS the trial. Direct link to upgrade page (not manage account → change plan). Post-upgrade instructions must be explicit. Require card at point of conversion. Don't constantly nag about upgrading. No redirect-back annoyance — push from
ClaimClaim_FreeIsTrial Free IS the trial. Direct link to upgrade page (not manage account → change plan). Post-upgrade instructions must be explicit. Require card at point of conversion. Don't constantly nag about upgrading. Agent must redirect back after upgrade
ClaimClaim_FreeTierReadOnly_DesignSession Free tier: read-only. Can load and use perspectives. Cannot save memories. Cannot review. This is the capability gate that drives the Free→Expert upgrade.
ClaimClaim_FrontIsCoordinationBrain Front is the coordination brain handling conversation state, threading, history, and ambient visibility for the principal, while Twilio is pure transport.
ClaimClaim_FullDeletionOnlyRemedy Without temporality modeling, the only available corrective action when speculation pollution occurs is full deletion of the affected memories and starting over. the customer's chief architect's plan for Phase 1 seeding was to delete all me
ClaimClaim_FullToolSurfaceHumanWords Full tool surface: perspective, cite, remember, forget, review, introspect, share. Every tool name is a word a non-technical person would use in conversation.
ClaimClaim_FullUtilityTesting Every beta user gets the full [amount] consulting engagement to fully test someone using it with full utility
ClaimClaim_FundraiseWithWorkingDemo Fundraise with a working demo — not a deck about a vision, but a video of a working product. The thing you're selling needs to work before you sell it.
ClaimClaim_FundraisingBriefStale_GTMRationale The Go-To-Market section (lines 237-239) of the fundraising brief is missing the deeper rationale for the interview-first sales motion. The brief treats it as an efficiency play, but the real reason is that it is the forcing function that k
ClaimClaim_FundraisingBriefStale_WinterTrajectory The winter trajectory section (lines 321-323) of the fundraising brief states 'Raise venture or sustain on revenue,' which inverts the actual posture. Organic scaling is the thesis; VC is a specific tool for a specific phase, not a default
ClaimClaim_GapHonestyIsCounterIntuitive Asking the AI something it does not know and having it identify the gap and who to ask is counterintuitively powerful — it demonstrates epistemic integrity.
ClaimClaim_GatekeeperDesignQuestionsOpen The Microsoft gatekeeper pattern raises specific open design questions for the org-tier coordination plane: Does the gatekeeper hold ground truth or just route information? Is it a separate perspective or a property of the org-tier coordina
ClaimClaim_GatekeeperOrSalonQuestion Is the multi-role-with-gatekeeper pattern interesting enough to pursue as a product configuration (org-tier sub-agent panel for specific high-stakes decisions), or is it a generalization already covered by named perspectives + review? The M
ClaimClaim_GIFDemoReady GIF generation is ready and impressive. May make sense to embed an intro demo GIF into the F&F plan itself as a visual proof point.
ClaimClaim_GIFLooksReal Tony's core insight about the GIF-based demo approach: it just looks real. Not a concept, not engineering skunk works — it looks like a real thing. That's the trick. This is the sales advantage of GIF-based materials over decks or white pap
ClaimClaim_GitHubAbstractionComplete The tool naming IS the GitHub abstraction completing. Free/Expert users don't see 'compile from repo' or 'load from share bundle.' They see 'perspective'. The abstraction is complete when the language doesn't leak implementation.
ClaimClaim_GitHubAbstractionHighPriority How to abstract GitHub from signup is high priority — SMB users won't have GitHub accounts. This is an open question flagged as high priority for the SMB market.
ClaimClaim_GitHubIsInfraNotProduct Scarlet articulated that the GitHub-native layer is infrastructure on which many products can be built; the product is abstracting GitHub completely away for non-engineers.
ClaimClaim_GitHubOptionalNotRequired GitHub becomes optional storage, not a required third party. The graph can live in memory, stored nowhere by default. Download at session end, upload to resume. This removes the GitHub dependency from the individual self-serve path and stru
ClaimClaim_GitNativeDifferentiation Built-in memory features (ChatGPT Teams, Claude Projects)
ClaimClaim_GoalDirectedThirdMode Goal-directed mode is a third mode layering above Ambient and On-demand behaviors. It is proactive rather than reactive.
ClaimClaim_GoldAccentAuthority Gold/amber is reserved for expertise-signaling chrome, differentiating from generic B2B SaaS blue.
ClaimClaim_GovernanceGraveyardCategory 'Governance software' reads as GRC (compliance, approval workflows, audit) — a graveyard category aswritten does not belong in. 'Governance' names the customer's world and must be used to speak their language, but must never become aswritte
ClaimClaim_GraphGroundsLLM The graph store grounds the LLM; the LLM makes the graph navigable.
ClaimClaim_GraphHealthLocalSkill Graph health report becomes a local skill producing output on demand or schedule. Currently a story draft that gets auto-generated.
ClaimClaim_GTMSequence an infrastructure advisor articulated cleaner framing: 'train the trainers → here's the solution'; requires deck update to reflect this sequence
ClaimClaim_HubCollectionOperatorDriven What was previously a methodology Scarlet executed by hand for engagements is now written into the onboarding prompt for any user's agent to run. The hub-collection method with provenance is the specification any future reconciliation tooli
ClaimClaim_HubMemoryIsPrimarySourceMaterial A hub memory is still primary source material, not an abstract. It must contain extended verbatim excerpts from pivotal moments, each attributed to speaker and date, referencing the original transcript file. Summarization compresses the rea
ClaimClaim_HumanExpertiseIrreplaceable The nature of expertise is the selection of what to do and when, based on prior experience — exactly what humans are needed for. Nowhere in this future do humans get replaced or supplanted; perspective becomes more important, not less.
ClaimClaim_HumanMomentSellsBetter a design collaborator's reaction to the demo walkthrough confirmed: the human moment (the call, the expert's reaction to seeing their words cited) sells better than any screen recording. Tony's 'it just looks real' applied to video, not jus
ClaimClaim_HumansAsBottleneck The bottleneck on AI progress is humans, not models; institutions and implementation are the rate-limiting steps.
ClaimClaim_ICPBespokeConsultancies ICP Profile 2: smaller practitioners / bespoke consultancies — architect or lawyer's office — where their process is really their product and they have segmented worldviews for different clients.
ClaimClaim_ICPExcludesFortune500 ICP explicitly excludes Fortune 500 buyers. Scarlet: 'Note that the enterprise customer is an enterprise vendor not a giant enterprise. And id prefer to work with that icp, 50-250 to maybe 500max employees. Which means they have budget but
ClaimClaim_ICPIndividualWhales ICP Profile 3: individuals / 'whales' — high-net-worth or famous authors — people that have their own worldview they're curating. Scarlet and a consultant who ran enterprise discovery engagements for fifteen years both fall under this.
ClaimClaim_ICPMidMarketSoftware ICP Profile 1 (currently working): mid-market, enterprise-facing software companies, ~20–250 (maybe 500) people. JUXT and Vouch cited as fitting this persona. Shape and customer they're serving matters more than size.
ClaimClaim_ICPValidated an infrastructure advisor provided specific parameters (5-150 people, [amount]-50M revenue) and Google Search analogy; strengthens ICP definition
ClaimClaim_IdentityGapImplication The onboarding methodology must force an explicit pass for organizational identity even when the customer thinks it is unnecessary, because the agent will ground itself somewhere; the question is whether that grounding is correct or invente
ClaimClaim_ImplementationDetailExclusion Specific code functions, color values, layout specifics, type scale multipliers, hex codes, SCSS assignments, and config values are implementation details. They live in files (tokens.json, CSS, SCSS) and change without discussion. They must
ClaimClaim_ImplicationInverseOfEscalation Coined this session, provisional. The inverse of escalation (when a deployed perspective hits the bounds of what it knows and escalates to a human) is 'implication' — what implicates the perspective, what makes it relevant, how do we know i
ClaimClaim_InDegreeSelectionFlawed The v1 map selected its umbrellas by raw in-degree — whichever hubs had the most children. Measured against the real graph, that gave style-taxonomy hubs (StyleMetrics, KeyPhrases, Antithesis, tone records) the majority of the map's slots,
ClaimClaim_IndexNeverGatesStore The index must never gate the store. Steering requires compression but compression via fixed themes is exactly the reductionism the discourse-act ontology exists to refuse. A claim is not about something; it is available to be read as many
ClaimClaim_IndividualCategoryWordNotCharacter The category word for the B2C product should be 'individual' — the antonym of monoculture/average. Explicitly NOT 'character' or 'persona' — 'not memory, individuality.' Scarlet is separately hunting for a warmer, whimsical relational name
ClaimClaim_IndividuateVerbReopened 'Individuate' as the hero customer verb is unresolved — it was retired internally Apr 9 in favor of 'perspective pipeline'; resurrected here for positioning, which is a different concern than pipeline naming. Pipeline naming vs. positioning
ClaimClaim_InfrastructureGap Git-native worldview enables CI/CD integration, agentic automations via GitHub Actions and MCP, and contribution-side multi-player AI
ClaimClaim_InternalCompileTermCorrect On the code side, 'compile' is retained. The n8n workflow literally compiles memories + transactions into a worldview snapshot. The engineering metaphor is correct for the engineering layer.
ClaimClaim_InterpretationPreservation The ARB discusses why some options are worse than others; a perspective preserves this interpretation where code-grounding loses it.
ClaimClaim_InterviewAsPremiumOnramp The interview sales motion becomes a premium/deep onramp for high-intent prospects or the Professional tier ([amount] SOW), while lighter onramps are the default for top-of-funnel.
ClaimClaim_InterviewDemotedToFallback The live interview is now the no-transcripts alternative, not the primary onramp. Docs home page, Quick Start, and ASWRITTEN.md all lead with transcripts. The interview is offered only when the user has no transcripts to provide. Implemente
ClaimClaim_InterviewOneOfManyOnramps The interview is now framed as one onramp among several (transcripts, public content, spec demos), rather than the default or required mechanism for seeding a perspective.
ClaimClaim_InterviewPutsYouInHumanMode Scarlet's insight: the interview sales motion puts her brain into human mode. The sales motion becomes: do an interview with an actual human, the result of the interview is the start of their perspective (free), and if they want to add memo
ClaimClaim_InterviewThreeOutputs One interview conversation produces three distinct outputs: (1) installable perspective for the expert (product delivery — free account pre-populated); (2) podcast episode / content piece (marketing — published with permission); (3) live de
ClaimClaim_IntroducerGapOpen Settled: the introducer doesn't get money, money would degrade the trust capital, and the introducer benefits via reach/distribution and the negative incentive of avoiding unpaid vetting labor at scale. What's still open: whether there's a
ClaimClaim_IntrospectFixesHallucination David independently validated introspect-as-escalation: 'that fixes one huge problem with LLMs — once you're in unknown territory it just makes up plausible stuff, it doesn't stop and claim it doesn't know something' [01:12:31]. External-va
ClaimClaim_IntrospectNoSubgraphReturn Returning subgraph alongside introspect analysis causes attention dilution. Analysis (gaps, tensions, questions) and subgraph content serve different purposes at different moments.
ClaimClaim_IPCrisisExistential Proposed framing: IP leakage as existential threat; requires validation and quantification
ClaimClaim_KnowledgeExtractionMarketGap Tony asserts 'no one's selling that as a solution' — proposed stake, needs validation via sales.
ClaimClaim_LabelShiftIntent Correction 2: Each shift must be labeled as improvement (sharpening current direction) or counter (testing alternative direction). Reviewer must know why they are looking at the shifted version.
ClaimClaim_LandingPageBehaviorOperationsBased Scarlet's Jul 2 direction: 'the landing page is more behavior and operations-based.' The page's unanswered question is 'what is the pain, and what do I do with it? What does that actually look like?' The answer architecture is use-case-led
ClaimClaim_LandingPageContentCommentaryIA The landing page's IA is structured as: primary source material (what is actually happening) + commentary and framing (briefly). This is the same structure as a memory itself. The product's own epistemics become the page's information archi
ClaimClaim_LandingPageIsHeroAction The landing page becomes a deep link. The product IS the hero action; visitors touch the product immediately rather than watching a demo.
ClaimClaim_LandingPageMismatch Current aswritten.ai landing page does not match the sales conversation a demo-call prospect is having with prospects. Needs updating to reflect current positioning and SMB pivot.
ClaimClaim_LandingPageShowDontTell_Scarlet Scarlet's core reframe (Jul 4): the current landing page fails by mode, not by content — it explains the product instead of demonstrating it. The fix is to shift from telling to showing: primary source material plus brief commentary and fra
ClaimClaim_LandingPageTagline_Jul2026 First deck slide drafted Jul 2: 'Turn your conversations into a perspective that is cited by AI.' The conversations-Rolodex hero stays, but the weight moves into use cases. This is a candidate tagline for the landing page, not a finalized o
ClaimClaim_LargeTXScale Correction 4: Large TXs handled via TX summary → zone identification → 2-4 targeted shifts in chosen zones → fact review catches the rest before merge.
ClaimClaim_LayeringCoreFeatureDifferentiator Layering is a core feature and differentiator — no other product lets you stack curated worldviews. Already works in the current product via multi-repo compilation on the same ontology.
ClaimClaim_LeanStartupRepoAsGuardrail Tony and a technical collaborator put the Lean Startup books into the repo they're working on and use them as a reference to keep on track — specifically to prevent overthinking and over-engineering before actually talking to anyone about a
ClaimClaim_LensDraftingWorks Full section drafts through explicit perspective lenses, combined with a delta table, successfully extracted strategic knowledge from Scarlet that four prior approaches failed to extract.
ClaimClaim_LensesConvergeOnBranchIntent Some lenses push TOWARD the branch thesis (confirm sharpening); some push AWAY (confirm alternatives are worse). Goal is sharpness on what the branch is trying to do, not answering all questions.
ClaimClaim_LensIsCompoundNotStacked A lens is inherently multi-dimensional — it can combine audience + conviction + domain into one perspective. You do not stack lenses; you craft one compound lens. The shift is the delta between canonical view and lens-weighted view.
ClaimClaim_LensIsIrreduciblePart a design collaborator: 'without my input through my own lens, it breaks, it falls down. So that shows that the actual unique part is by [my] lens — is applying the lens, which I suppose is effectively what you've built there... it is the le
ClaimClaim_LensToolIsGenerator The lens tool does not run lenses — it generates them. Output: a set of lenses each containing a centrist focus prompt, a shift prompt, an eye chart, and usage instructions.
ClaimClaim_LowercaseBrandStylization Lowercase 'aswritten' signals a code-adjacent, anti-corporate identity.
ClaimClaim_MarketplaceInterviewDriven Marketplace browse/search is below the line for April 1 because distribution is interview-driven, not browse-driven. Starter perspectives are also below the line because interviews CREATE the starter set. The interview model makes the marke
ClaimClaim_MarketplaceNotNeededAtLaunch The marketplace concept shifts from buyer-side browsing to supplier-side distribution. The product may not need a marketplace at launch — just a share/install flow. Experts do not come to a marketplace to browse; they encounter aswritten be
ClaimClaim_MarketValidation_1 Shared organizational context for AI agents is now validated as enterprise requirement by OpenAI's Frontier investment.
ClaimClaim_MatchingSearchExtendsCVP 'Install expertise' is the mechanism; 'matching-search' is what makes installation valuable when deployed externally. The two framings are complementary: the category reframe does not replace Claim_CoreValuePropInstallExpertise but extends
ClaimClaim_MCP_SchemaAsPrompt For MCP tools, schema properties act as prompts; a parameter in the schema is more effective than prose descriptions.
ClaimClaim_MCPAgnostic MCP-agnostic architecture claim now has production evidence: GitHub Copilot + GPT5.2 Codex integration confirmed working. This is a Boulder-level conviction — settled, central, hard to move.
ClaimClaim_MediumIsTheMessage The compelling part of the product is the MEDIUM (the transparent agent that can go back and forth), not necessarily the perspective's absolute fidelity.
ClaimClaim_MemoriesSourceToolRemoved Memories/Source tool removed from product. Memories are just files in the repo — anyone with a coding agent can read them directly. Org-only skill at most.
ClaimClaim_MemoryCapture Captures new information and key decisions and drafts memories. Boulder-level, Canonical Memory Creation Workflow.
ClaimClaim_MemorySaveIsMostConfusing Scarlet may overfocus on the memory save as the moment of value when it is actually the most confusing part for prospects. The value is upfront in compile/annotate/introspect.
ClaimClaim_MidMarketPitchBuyVsBuild Mid-market pitch is buy-vs-build vs DIY engineering, NOT vs Glean/Fortune-500 platforms. For technical mid-market vendors (DIY-tempted buyer): 'You're a 150-person company. You don't want to buy Glean — it's too big a commitment for your sc
ClaimClaim_MisrepresentationRisk The '80/20' danger: prospects remember the 20% the AI gets wrong. Mitigation: transparency from moment one and using escalation as a humility signal.
ClaimClaim_MissingCompileTimeLayerParameter The missing piece is a compile-time layer parameter for actor focus. This is a new feature. Transform is for permanent forks (e.g., /transform → permanent branch for PII-stripped public version). Actor/role/focus lenses are for ephemeral vi
ClaimClaim_MoatCreatorOwnsData Moat clarification: Creator owns data, not consumer. Consumer receives compiled perspective, not source documents. Even with the raw repo, the extraction process (source → transactions → compilable graph) is the real defensibility. Scarlet:
ClaimClaim_ModelAsHardwarePerspectiveAsSoftware The mechanism that makes the moonshot buildable: treating the model like hardware that we install the program of perspective on top of. This is why diversity of voice can be built on top of the monoculture rather than requiring a separate m
ClaimClaim_ModularSynthesizerDynamic Scarlet's description of the session pattern dynamic: 'guiding the synthesizer but there's an emergent behavior that takes something deeply true for me and makes it far more realized than I could have made myself.' Scarlet's corrections to
ClaimClaim_ModuleNameTBD The name for the installable unit of expertise is not decided. 'Module' is a working term only. The concept it needs to capture: a complete multifaceted perspective — a body of work, a way of thinking. Connected to the original 'synthetic i
ClaimClaim_MomentumOverLandingPlane The core principle: build contrast through momentum. Walk the reader from where things start to where they're going. Belief, specifics, belief. 'Landing the plane' clauses (e.g., 'Zero data retention is the architecture, not a configuration
ClaimClaim_MoonshotAntiMonocultureThesis The moonshot framing for EV and fundraising: the consensus is that AGI is coming (Scarlet concedes this deliberately per EV prompt). The contrarian turn: not whether but how many voices. The current trajectory is a handful of voices from fr
ClaimClaim_MoreMaterialNotAProblem Unlike a deck where you want one clean statement per screen, the extraction pipeline benefits from richer input. More material is not a problem.
ClaimClaim_MorningNarrativeAfternoonDevStructure Scarlet's established morning-narrative / afternoon-dev working structure exists precisely because switching between altitudes mid-stream is costly. The altitudes are kept on separate sides of the day. This is a structural solution to the a
ClaimClaim_MultiplicityInVariation Multiplicity lives in variation, not a theme label. The digest already refuses single-theme reduction via the variation block: it surfaces reads as X, and as Y, and the road not taken. That is the anti-reductionist mechanism and it stays lo
ClaimClaim_MultiRoleCoordinationGatekeeper The Microsoft gatekeeper pattern suggests a centralized ground-truth holding mechanism missing from current named-perspective architecture.
ClaimClaim_NamedPerspectiveAsSubAgent Scarlet [30:41]: 'In the planning session for any particular task, like, I want to know what Tony thinks as Tony has represented, you know, your... And our sort of shared opinion over every call we've had.' The named perspective is not a fl
ClaimClaim_NarrAppliesAsOntologyExtension A new ontology property — narr:appliesAs — with values like current | nearFuture | exploration | speculative would solve the temporality gap structurally. Compile would filter by temporality as well as conviction: default worldview layer in
ClaimClaim_NarrativeArchitectureWebNotMapping The architecture is an interconnected web, not a 1:1 mapping. An incentive could be implicated in many flows and vice versa. Flows are situational; the entry point is the situation -- what the actors are trying to do in the product right no
ClaimClaim_NarrativeIsStatementTrueBeyondWords A narrative is a statement that is true beyond the words it describes. Test for every prompt-surface sentence: does it carry a general posture the specific behavior only instances? 'The footnote lets the user reach for provenance when they
ClaimClaim_NarrativeStrategyReportDeliverable The narrative-strategy report is the most concrete initial-services offering Scarlet has articulated: a 2-4 week, NDA'd, fixed-deliverable engagement that rides alongside a pilot. It directly answers 'what do I hand them in two-to-four week
ClaimClaim_NavigationDropsContextOrderOfMagnitude Navigation drops per-call context by an order of magnitude, off the long-context path entirely. The compiled snapshot had grown to roughly half a million tokens. The fix for the flakiness and the direction of the architecture were one decis
ClaimClaim_NetworkProblemStrategies Three strategies to address the network/cold-start problem: (1) curate from public expertise, (2) partner with willing experts ('your book is a snapshot, your module is alive'), (3) build task-specific modules that demonstrate obvious value
ClaimClaim_NewSchemaAddsPrimaryRelatedNarrative Consumers of aswritten_cite should expect: each citation entry carries a primary source object, a related[] array with relevance tags, and a narrative paragraph rendered directly as the footnote body. The narrative embeds the primary's verb
ClaimClaim_NoAILanguageTricks Sneaking a metaphor into a descriptive noun (e.g., 'scrolls away by Friday') is the same pattern as the hand-waved closing. Be direct. Say 'disappears'. Say what happened.
ClaimClaim_NoCoFounderNeeded Scarlet decided she does not need a co-founder, but rather a constellation of trusted partners and advisors.
ClaimClaim_NoCompressionToSlogan Compressions of Scarlet's conversational descriptions into coined phrases or slogans are explicitly rejected. The specific example: her phrase 'it speaks to, references, and works from what was already said' was compressed to 'It speaks fro
ClaimClaim_NoEmDashes Replace em-dashes with commas, periods, or colons. The em-dash has become an AI tell. Rule applies in code comments, design docs, memories, and public copy, anywhere AI-generated text might accumulate.
ClaimClaim_NoHandWavedClosings Drop hand-waved closings: 'from day one', 'snaps into focus', 'scrolls away by Friday', 'finally added up', 'where it all clicked', 'the quiet part' without naming what the part is. Drop anything that sounds like the resolution of an arc. S
ClaimClaim_NoNewToolsRequired No new tools required today. The process uses compile, introspect, and remember. The lens is a prompt pattern. Future: formalized aswritten/lens tool with saveable configurations, but prompt-based version works now.
ClaimClaim_NoSecretsRequired Client repo should not need any secrets to use collective memory. OPENROUTER_API_KEY is optional BYOK exception.
ClaimClaim_NotificationsRemoved Entire Notifications system removed. Synchronous remember eliminates primary notification use case. Secondary case (review requests) handled by GitHub natively. Workflows to delete: review-notify, msg-review-notify. Keep msg-send and front-
ClaimClaim_NotificationTransactionSideEffect Notification fires when transactions land, triggered by tx/post-commit. It is not part of the review flow. It already works. Not being redesigned. Teammates get notified when new knowledge is extracted regardless of whether the author has r
ClaimClaim_NotProductDemoButConsulting Explicitly not a product demo but a consulting engagement session — call one's interview phase.
ClaimClaim_NotTreatmentProxy This is not necessarily a good tool to be a treatment proxy or to be giving advice. What is interesting is it as a matching solution or a sort of vetting for potential patients into the funnel. Valid questions the perspective answers: 'How
ClaimClaim_NoVCAtThisStage Venture scale money would not be a meaningful help to the business in its current shape; organic scale is preferred.
ClaimClaim_NovelPersonalOrgCompile No competitor has conceptualized the personal-to-org compile layer. OpenAI Frontier, Notion AI, Guru, Confluence are all org-only.
ClaimClaim_NoXNotY Cut the 'X, not Y' pattern. It lands the plane. If contrast carries load-bearing meaning, rebuild it through momentum (belief, specifics, belief). Active sentences only.
ClaimClaim_OnboardingElimination Teach new people... explain your org to new hires. Foundation-level, Onboarding Elimination Archetype (Jan 7, 2026).
ClaimClaim_OnboardingHumansAndAIs Onboarding both individuals and AIs into an organizational perspective is the same recipient shape. Scarlet verbatim: 'Both onboarding their individuals and also AIs into an organizational perspective is super powerful.' The use case is not
ClaimClaim_OnboardingMethodologyGeneralizes The three-part seeding methodology generalizes beyond the enterprise customer: any domain expertise codified in historical recordings (sales calls, customer interviews, OKR retrospectives) is a candidate for the same pattern. This is a stat
ClaimClaim_OnboardingModeHighestLeverage Onboarding mode is the highest-leverage product gap for beta. Technical infrastructure works; cognitive bootstrapping does not.
ClaimClaim_OnboardingPromiseSixMonthsTranscripts 'Send us six months of transcripts and get a perspective tomorrow' is absolutely a promise the product can make. The onboarding promise holds because bulk ingestion via AI summarization is a valid, effective method. Scarlet confirmed this e
ClaimClaim_OnboardingPromptLeadsWithTranscripts The Onboarding Mode prompt (ASWRITTEN.md) now asks for call transcripts and AI working-session history first, by name; carries the distill-never-dump-raw rule; the hub-collection method at volume; and the provenance requirement. The protoco
ClaimClaim_OngoingIterationStep3 Step 3 of the three-part seeding pattern: as ARB meetings happen, ingest each transcript with active review. The transcript is the diagnostic, not the memory. When the system fails to answer a question well, the user does a 30-minute voice
ClaimClaim_OnPremNotDefiningFeature On-prem deployment is NOT what defines the Professional tier. It is a requirements-driven add-on available at that tier. The primary differentiator is consultant-led knowledge extraction. On-prem is triggered by compliance requirements (HIP
ClaimClaim_OntologyPrimitivesAreAgnostic Core ontology primitives (narratives, actors, claims, decisions, convictions) are domain-agnostic. A Chinese herbalism perspective has narratives, makes claims with varying conviction, involves actors, supersedes prior thinking — same struc
ClaimClaim_OperationsICP Operations persona as high-value ICP
ClaimClaim_OpinionatedPromiseVerifiedGap Verified gap (Jul 1 search): brand/index.md holds 'opinionated' as definition — 'a perspective is a highly opinionated set of norms, processes, and narratives' and 'the opinionated middle of a category' — and the demo-GIF brief scripts 'a s
ClaimClaim_OpinionedPerspectiveWinsAllThreeMarkets One selector, three markets: hiring (Tony), funding (Cowen's 'applications fail by being normal'), and buying. Opinionated perspective wins all three — and an application or landing page arguing this demonstrates itself. Tyler Cowen's scree
ClaimClaim_OptionsVsDecisionsVerbatim an engineering manager at the customer's verbatim account of the options-versus-decisions distinction that drives the day-to-day re-litigation problem: 'We gave you pros and cons and you were to go away and figure out which one suits and co
ClaimClaim_OrgMemoryDemotedSubUseCase 'Organizational memory' moves from the pitch to a governance sub-use-case. Scarlet: 'I'd take that as a sort of sub-use case, engineering governance or just governance processes generally, and that would be the the enterprise customer ARB u
ClaimClaim_OrgRolesFirstClass From PR 91 work on synthetic representatives: a role (head of marketing, clinical supervisor) is an abstraction that exists independently of the person filling it. The role has a module. It evolves. When someone new steps in, they install i
ClaimClaim_OrgTierCatalogueSetup The [amount] organization-tier setup fee creates the initial roles and packages that populate the internal catalogue — the org's internal marketplace of expertise. This is the scoped engagement version of the conversion service at enterpris
ClaimClaim_OrgTierDedicatedPlane The Org tier is categorically different because it provides a dedicated coordination plane where the organization acts as the 'wizard of oz' behind their own agents.
ClaimClaim_OrgTierThreeCapabilities Org tier subscription includes three capabilities (not separately priced): (1) Internal sandboxed AI with observation — for AI-naive staff who can't be trusted with raw Claude/ChatGPT but need the perspective to do their work; lets non-powe
ClaimClaim_OtterAPIGated Otter — likely the most popular transcription tool — gates its API behind Enterprise contracts; no self-serve API path exists. Any architecture that depends on aswritten holding vendor API integrations hits this wall. This dissolves the Ott
ClaimClaim_OutboundVariantSoftening Scarlet floated aggressive cold outreach: create a perspective on spec, deploy an agent that contacts the prospect as them. Tony softened [37:40]: 'Yeah, I wouldn't make it so scary, but idea it's like, hey, I am the virtual. I've cloned yo
ClaimClaim_OutreachTemplates SMS and email templates saved to docs/sales/outreach-templates.md for inviting experts to interviews. Key elements: what happens (30 min conversation), what they get (free perspective), what's free vs paid (explicit), content angle (publish
ClaimClaim_OutsetaOwnsIdentity Outseta owns identity and billing. GitHub is a connected source, not the identity layer.
ClaimClaim_OwnByDefaultPayToContinue Own-by-default, pay-to-continue is the conversion mechanic for the individual self-serve product. The graph belongs to the user by default; payment gates continued building past the free threshold. This is the individuation curve's monetiza
ClaimClaim_PainFirstDeckTakenToLimit The question-first design is the January pain-first deck decision taken to its limit: the deck does not present capabilities at all until the prospect has answered the three pain questions about their own organization.
ClaimClaim_ParkArchitectureOnlyAdds Park's 25-agent simulated town architecture only adds memories — nothing decays out. Their world ran a day. At longer horizons their memory store would grow unbounded, same shape as the canonical aswritten repo hitting the 256K token compil
ClaimClaim_PassesNotTransactions Passes ≠ transactions. A pass is a compute stage; a transaction is an atomic write. Fan out the compute, write atomically at the edge. This distinction prevents conflating pipeline orchestration with graph mutation.
ClaimClaim_PatientAsGoalHolder Patient pays per consultation with the practitioner's AI, or subscribes for a profile that compounds across consults and verticals. [amount]-100 range per vet was the working number. The patient is the goal-holder for this transaction (thei
ClaimClaim_PaymentResolvesEthicsQuestion Scarlet's resolution to the ethics question of lifting advisor opinions into AI agents: 'Maybe I pay them. Right. Like, but maybe I pay them. Not a lot.' Payment transforms the relationship from extraction to compensation — the advisor is n
ClaimClaim_PeerValidationArchitecture a consultant who ran enterprise discovery engagements for fifteen years independently arrived at nearly the same architecture as aswritten. This call is the strongest peer-validation of the core technical bets to date. a consultant who ran
ClaimClaim_PersonasNotSeparateScenarios Personas are NOT separate scenarios or a selector: 'we just want to make sure there's points in this where any of those personas can see themselves.' Each beat gets at most a couple sentences of commentary — more importantly how it's unique
ClaimClaim_PerspectiveAsCorePrimitive 'perspective' is the chosen tool name replacing 'compile'. The tool name IS the product's core primitive. You don't 'do something to' a perspective — you're in one. Decision reached after brainstorming 17+ candidate words.
ClaimClaim_PerspectiveCreationServiceRename 'Distillation' is renamed to 'perspective creation' for the professional service. Scarlet: 'I don't like the word distillation. Let's just call it perspective creation?' Rationale: 'Perspective creation' is what it literally is — we create
ClaimClaim_PerspectiveDeckAlternativeUnexecuted Status note from Scarlet (June 29, 2026): 'I haven't delivered an AI perspective as an alternative to a deck to a consulting client yet.' The thesis is settled (boulder-level since March 19) but the use case has not been executed in practic
ClaimClaim_PerspectiveHoldsIntent The perspective holds decisions, reasoning, felt intent, and the conceptual how. The implementation holds the literal detail. The line runs inside 'how': directional and relational how belongs in the perspective; specific values, identifier
ClaimClaim_PerspectiveInstallsBehaviorNotJustFacts Perspectives install behavior, values, and self-understanding into agents — not just factual knowledge. An agent that reads from a perspective built from conversations about how to approach search will approach search that way. This is a di
ClaimClaim_PerspectiveNotInformation Modules load a set of perspectives, narratives, approaches, processes, and opinions about how information is used, framed, what it means, and what it shifts in different contexts. This is the exact difference between a human expert and an A
ClaimClaim_PerspectiveStartsDevelopment Scarlet [01:04:07]: 'The perspective starts being the thing that's doing the development instead of you doing it.' This is the culminating insight of the call — the named perspective architecture is not just a knowledge management tool but
ClaimClaim_PerspectiveTracksARBEvolution The system must surface how other ARB sessions during the lag period affect original guidance.
ClaimClaim_PerspectiveVerbConcernResolved The concern that 'perspective' is not a verb ('you can't perspective the perspective') was raised and resolved: nobody says 'perspective the perspective' — they say 'perspective a beta user from outside tech-bluebird' or the AI says 'workin
ClaimClaim_PerTaskOAuthToken a consultant who ran enterprise discovery engagements for fifteen years's formulation: in agentic discussions, the shift is from broad-key OAuth grants to single-task tokens — 'it might be to read a contract, or to increment a salary — but
ClaimClaim_PerUserMethodPrinciple Methodological principle: analyze customer adoption per-user over time, never in aggregate. Aggregate metrics inherit the heaviest user's shape and hide durable-core formation and early-trial attrition.
ClaimClaim_Phase2SharperThroughCapture Phase 2 ARB is sharper because Phase 1 has been captured and the gaps have surfaced.
ClaimClaim_PitchReframeAgentDeployment Pitch reframe from knowledge-management framing to agent-deployment framing. Old pitch: 'We install expert perspectives into AI tools.' New pitch: 'We make your expertise deployable. Internally — AI-naive staff get sandboxed AI you can moni
ClaimClaim_PlanGatedInstructionsFromCompile Feature gating mechanism: compile returns different instructions based on authenticated plan level. Not separate tool sets — different prompts guiding agent behavior per plan. Remember exists at all plan levels but returns upgrade prompt on
ClaimClaim_PlanningIsBottleneck a founder beta user validates Feb 5 thesis: planning is the obvious next thing. Understanding the codebase is the key use case for dev-focused users.
ClaimClaim_POCMSAParallelNotSequential Tony's structural unlock: start MSA/procurement conversations in parallel with the POC, not after. Procurement can take weeks. If you wait for POC success before starting MSA, you add weeks of dead time. Running in parallel means the transi
ClaimClaim_Positioning_1 aswritten positions upstream of agent platforms (Frontier, Cowork, Agent 365) by extracting undocumented knowledge into organizational worldview.
ClaimClaim_PositiveDirectClaims Form claims about what Aswritten does. Make clear positive statements that state what is actually changing, what is happening. Not 'aswritten inverts this' but say what it does instead.
ClaimClaim_PractitionerBetaUserContentAvailability a practitioner beta user immediately said that many podcasts exist plus the forthcoming book and marketing/blog content. Important caveat in his own words: many of the podcasts don't contain direct statements about his own practice — he tal
ClaimClaim_PractitionerBetaUserMarketingPain a practitioner beta user separately lamented that his marketing website is a constant source of frustration and needs work. The session connected this to the perspective-as-evaluation thesis: a perspective becomes a way to evaluate whether
ClaimClaim_PractitionerEarnsConvertingMethodology Practitioner earns by converting unstructured methodology into a reusable, citable AI of themselves with a paying patient already attached. Practitioner sets the rate. Higher tiers add observation and escalation: their AI handles initial co
ClaimClaim_PractitionerICPBridge The discovery wedge sits in the same ICP as a design collaborator's 'deliver an AI perspective instead of a deck' idea — the practitioner whose process is their product (customer-profile #2). This is a potential bridge between the enterpris
ClaimClaim_PreReadDeckRequest a demo-call prospect requests a pre-read deck: visual-first, executive-accessible, not a pitch deck. Framed as 'McDonald picture menu.' Makes meeting productive from minute one.
ClaimClaim_PrimaryUniquenessInvariant Primary sources are globally unique within a single response. The deduplication invariant: unique_citations ≤ supported, where two citations never share a primary. Inline footnote markers ([^1]) may repeat across text; citation entries do n
ClaimClaim_PrivacyConcernDrivesAvoidance Tony revealed a privacy concern that drove his avoidance of aswritten: he didn't want experimental or sensitive work visible to others in the system. This is a distinct adoption barrier from technical complexity — it's about visibility and
ClaimClaim_ProcessSaleNoMaterialNeeded Tony [28:21]: 'So you can take exactly what you're doing now as a process and start selling that already without showing them any material. Right. So you can show them that this process is how you get to good.' The process itself — channel
ClaimClaim_ProductFilesUpdated The perspective vs. implementation principle is now encoded in memory-creation prompts across four files: CLAUDE.md (what makes a good/bad memory), ASWRITTEN.md (same plus Memories intro), resources/tool-descriptions/remember.md (tool descr
ClaimClaim_ProductIsCuratedExpertiseInstallation The product is curated expertise installation — a way to install expertise into your AI and verify where it's coming from. Consistent with Claim_CoreValuePropInstallExpertise (Foundation, March 19 2026 strategy session). This session reaffi
ClaimClaim_ProductIsNotPhoneNumber Explicit correction: the product is NOT a phone number. The prior tagline-level assertion ('the product is a phone number') was generated by the text-for-topic experiment and incorrectly promoted to current worldview by the extraction pipel
ClaimClaim_ProfessionalTierDifferentiator The primary differentiator of the Professional tier is consultant-led knowledge extraction — Scarlet doing the extraction for the client instead of the client doing it themselves. On-prem, unlimited seats, dedicated support, and custom onto
ClaimClaim_PromotionByRewriting Promotion is rewriting memories into main's voice and structure at the moment of decision — not cherry-picking commits or merging the branch. The research branch becomes a source of raw material, not a merge target. This matches the pattern
ClaimClaim_PromptExtensionPattern Base aswritten prompt + customer-specific extension is a real, reusable product pattern for scaffolding agents for specific use cases.
ClaimClaim_PromptIsCompiledNotAuthored The prompt is not authored -- it is compiled from an architecture built on the ontology's own concepts. Build the narrative architecture for the prompt surface first using the ontology as the data model (actors, incentives, behaviors, etc),
ClaimClaim_PromptsInPitches Prompts as ontology compilation targets should be remembered and included in pitches. Along with personal graphs, exemplifies hierarchical/federated knowledge as a strong pitch narrative.
ClaimClaim_ProposalHeadlineFrame Use 'proposal' as the headline value frame. It is structurally accurate across expertise verticals and sidesteps regulatory edges.
ClaimClaim_ProposalRegulatorySafeguard The boundary that keeps the proposal on the right side of regulatory edges: this is not a treatment plan. It describes the shape of the engagement and the practitioner's general methodology for cases like this — not specific clinical decisi
ClaimClaim_ProspectAsGenericActor Prospect as generic actor accepted as useful for modeling sales motion. Accepted as calibrated at fact review.
ClaimClaim_ProvisionalFlagSafetyRail Every memory on the research branch carries a provisional flag (opening line or frontmatter field: 'Provisional — from research/goal-directed-agents exploration branch, not yet promoted to main'). If any memory accidentally leaks into main'
ClaimClaim_ProxyEvolvesThroughConversation The principal can create new variants during real interactions, meaning the proxy learns and sharpens over time. It is a 'clone-that-grows' rather than a static chatbot.
ClaimClaim_PTPractitionerMeetingTwoMonths a practitioner beta user described a meeting he had recently scheduled with a PT to discuss each other's methodology. The meeting took two months to schedule and they were connected by a mutually respected practitioner. Two opportunities: (
ClaimClaim_PublicContentBootstrapOnramp Public content (podcasts, blogs, YouTube) serves as the substrate for perspective-on-spec outbound, bypassing the need for prospect time investment.
ClaimClaim_PublicPerspectiveSiblingRepo Conceptually distinct perspectives live in sibling repos, not subdirectories. Subdirectories are for entity isolation within the same perspective.
ClaimClaim_PublishLiberallyDecision Decision to publish liberally and not try to hold IP back. The line between what to publish and what not to publish is known and understood — you can give a full tour of the castle except maybe the metaphorical security room where all the c
ClaimClaim_PureFunctionSettled Settled design decision: Scarlet articulated it 3+ times, documented in backlog doc-006. Orchestrators own paths and side effects; pure functions return structured data.
ClaimClaim_QAOnboardingKillerDemo QA onboarding scenario identified as 'killer sales demo' by customer. Proposed, has supporting value (concrete timeline, measurable outcome), still moveable pending execution. Stake-level conviction.
ClaimClaim_QualityOfEngagementIsValidation The quality of the engagement IS the product validation.
ClaimClaim_QuantificationOneMetricPerEngagement Quantification is staged honestly: one metric per engagement, instrumented in the calibration readouts; friendlies produce the first real numbers; engagement one's number is engagement two's pitch.
ClaimClaim_QueerRepresentationPersonalGrounding Scarlet is a trans woman. The queer frame runs underneath the universal representation frame: 'this also parallels queerness... there's an opportunity here to do something specifically about queer representation... can I create a perspectiv
ClaimClaim_QuestionFirstDeckStructure Deck structure: (1) Cover question; (2) Optional why-now beat for AI-native buyers only; (3) Three discovery questions as slides 1-3; (4) Bridge — the judgment that fixes all three already exists, it's spoken in your meetings every week, th
ClaimClaim_ReachabilityAsKeystonePromise This framing is a principle candidate that would partially supersede existing category definitions (Curated Expertise, Delegation) by adding a 'promise' layer above the 'mechanism' layer.
ClaimClaim_ReachabilityPhraseOrigin Scarlet [30:26]: 'Narrative I've been playing with is reachability without interruptability.' Tony [30:31]: 'Yeah, I like that. I like that a lot. That's perfect. That is perfect, actually. I really like that.' Scarlet [30:38]: 'It came out
ClaimClaim_ReadPathStable The read path (compile, annotate, introspect) consists of stable, rehearsable operations. The write path has too many moving parts to reliably demonstrate live with unknown content.
ClaimClaim_ReconciliationIsCommentaryLayer Reconciliation is the authored commentary — a write primitive that is read-implemented but not yet written. The read model already separates primary-source grounding from authored reconciliation (the digest renders a position grounded in so
ClaimClaim_ReferralOnlyGrowthRisk Scarlet [39:19]: 'I don't want to paint myself in a corner where like the only expansion pathway over the next however many years is like progressive referral based, you know, person to person, like high touch sales.' Tony [39:44]: 'No, als
ClaimClaim_ReflectionLoopAsAbstractionMechanism Park's threshold-triggered procedural reflection is a documented candidate for task-176's abstraction detection.
ClaimClaim_ReflectionLoopTwoMechanisms Park's threshold-triggered procedural reflection is now the first documented candidate mechanism for task-176's abstraction detection: when sum of importance scores across recent memories crosses a threshold, run a reflection prompt, write
ClaimClaim_RefractBecomesLocalSkill Refract becomes a Claude Code skill (prompt-only, no n8n workflow). Reads recent TXs + snapshot + docs/ folder. Produces list of suggestions and focus areas for document update passes. Can run on cron if desired.
ClaimClaim_RefractDefinition refract takes a document and a TX (or arbitrary prompt) and returns focus areas for shifts in the document; it is change detection + focus generation, not content retrieval
ClaimClaim_RefractNotMVP Refraction is explicitly not MVP — it is a quality tool for publishers (Publisher tier, future). It requires GitHub and is part of the marketplace distribution layer, not the core install-compile-cite loop.
ClaimClaim_RefractRegistryRemoved Refract tool and Registry removed from product. Registry is just the docs folder. Refract becomes a local Claude Code skill. Workflows to delete: graph-refract, graph-scope (scope rolling into introspect). Remove from mcp-generic tool routi
ClaimClaim_RefractRegistryRemovedFromProduct Refract tool and registry removed from product. Registry is just the docs folder. Refract becomes a local Claude Code skill. Eliminates need to separate tools between org and everyone else — now one tool set for everyone. Workflows to delet
ClaimClaim_RegistryNotMVP Registry is explicitly not MVP — it is internal-only aswritten org tooling that serves the org-developing-the-product, not the product itself. Must be cleanly separated.
ClaimClaim_RegistrySimplification Registry is now just paths; no focus or intent metadata; checking for changes against focus is brittle, it should check against actual text
ClaimClaim_RelatedArraysMustDiffer A related source can legitimately appear in multiple constellations under different relevance tags. What is NOT acceptable is two citations whose related[] arrays are identical — that signals flat copies, not distinct constellations.
ClaimClaim_RelationshipGraphMostValuable The relationship graph — what you know about how specific people work together — is potentially the most valuable layer for consensus, though this is an open hypothesis.
ClaimClaim_RelationshipOverDemo Relationship trumps demo; minimum viable call is compile + interview + save one memory
ClaimClaim_RememberReturnThreeFields Three new fields in remember's return value: (1) tx_summary — plain-language summary of what shifted in the worldview; (2) zones — domains/areas affected by the extraction; (3) review_instructions — prompt fragment telling the agent the war
ClaimClaim_RemoveDiscoveries Discoveries fully removed from product, not just disabled.
ClaimClaim_RemoveNotificationsSystem Notifications system removed; synchronous remember eliminates primary use case.
ClaimClaim_RemoveRefractProduct Refract and Registry removed as product features; migrated to local org skills.
ClaimClaim_RemoveStoriesSystem Entire Stories system removed; functions migrated to on-demand skills or direct TX returns.
ClaimClaim_RepoConsentGate User-owned connected repo is the default for rich ingestion. Org-managed repos permitted behind an explicit consent gate: protocol discloses 'this stores verbatim content with us' and proceeds on consent. Philosophy preserved through disclo
ClaimClaim_RepoPerClient Repo per client. The consulting version of the participant loop is 'actually just me' — the salon-model idea with the participant loop deferred. Each client gets their own repo, isolating their perspective graph from other clients and from
ClaimClaim_ReportAsSalesMotion Every paying customer should get a monthly report framed in the brand aesthetic to anchor renewals.
ClaimClaim_ResearchBranchMethodologyDecision The research branch methodology is decision-level (actually in use, twice now) and a candidate for principle-level at a future promotion decision point. When an exploration is substantial enough to reshape main's direction but not yet settl
ClaimClaim_RetireZeroContext Decision: zero-context framing retired from all pitch materials as of 2026-03-05
ClaimClaim_RevenueBeforeRaise Sell a [amount] engagement NOW. Revenue before raise > projections about future revenue. Investment package becomes stronger with even one paid engagement.
ClaimClaim_ReviewableWorldviewSupersedesDemo1 Reviewable Worldview demo is more compelling than Demo 1 After the Call. After the Call is a waypoint within it, not a standalone demo.
ClaimClaim_ReviewEyeChartsSkillPrompt Review eye charts no longer pulled from registry. In Claude Code: review prompt offers documents from project intersecting review domains. In chat: offers documents from conversation or collective memory. In agent: pulls from repo directly.
ClaimClaim_ReviewInstructionsNotBlocking The review_instructions in remember's return should explicitly not block future work if the user declines review. Scarlet: 'We don't want to bombard the user with a giant block of text at the back end. We want a clear summary of the shift a
ClaimClaim_ReviewMemorySemantics Correction 7: Save memory = corrections exist, ends round, PR updates. Merge = no corrections needed, closes PR. A review with corrections always takes at least two rounds.
ClaimClaim_ReviewToolThreeCases Review tool extended to three cases: (1) review PR (team/org) — review a set of transactions; (2) review a named transaction; (3) review arbitrary content. Cases 2 and 3 may be the same. Primary use case: passing a transaction or concept in
ClaimClaim_SalesGuideRequest a demo-call prospect needs a written sales guide plus aswritten module repo — a playbook he can reference and share during direct sales conversations.
ClaimClaim_SalesIsDelivery The sales motion and product delivery are the same thing. Demo 1 IS the sales motion. Every sales call transcript goes into the prospect's repo collective memory before they say yes.
ClaimClaim_SalesMotionIsDelivery The sales motion and product delivery are the same thing — but the delivery happens across two calls, not compressed into one.
ClaimClaim_SalesMotionRecursiveDogfooding Applying goal-directed agents to aswritten's own sales motion is the ultimate dogfooding of the exhausted-expert trigger.
ClaimClaim_SalesPlaybookForDemoProspect A sales playbook for a demo-call prospect is the critical sales enablement piece this week. It should be a play-by-play with GIF descriptions showing what it looks like to take someone from first contact through a demo with deep knowledge e
ClaimClaim_SalonIsFifthPattern Prior taxonomy treated Salon as a fifth org-level deployment pattern alongside Solo, Parallel, Hub, Teams. This claim is now challenged.
ClaimClaim_SalonIsPerMemory Salon is not a fifth deployment pattern — it is a per-memory capability. The reviewers frontmatter on any memory triggers group interview → consensus → merge.
ClaimClaim_SalonReclassifiedAsReviewBehavior The Salon archetype from the original solution archetypes (Solo/Parallel/Hub/Salon/Teams) has been reclassified as the review behavior — collaborative review IS the salon. This supersedes the prior treatment of Salon as a fifth deployment p
ClaimClaim_ScarletBestPitchVerdict Scarlet's verdict, June 10: 'I think this is by far the best pitch and articulation of the product value we've ever had.' This marks the context-server framing as the highest-conviction pitch crystallization to date.
ClaimClaim_ScarletDrivesUserTalks First session is consultant-led; subsequent sessions the user can drive
ClaimClaim_ScarletTemporalityParallelObservation Scarlet independently arrived at the same temporality modeling need on the 2026-05-01 call, before the customer's chief architect described his demo pollution experience. She had been thinking about how to model temporality in the tool — sp
ClaimClaim_SceneGrammarSpecificSubject Scene grammar form: The [subject of the conversation] [conversation-type] [where, with whom]. First-person when the scene is from Scarlet's own life. The subject is the actual content of the conversation, not a tidy summary. Drop client nam
ClaimClaim_ScopeBuildFirst Build order: scope → refract → register → lens → enhanced start_review; scope is the foundation — everything else depends on it
ClaimClaim_ScopeDefinition scope takes arbitrary focus + context (both free text); the text defines the scope; no keywords, no predefined domain buckets
ClaimClaim_ScopeFragility Scope with narrow keywords risks excluding relevant material; passing a full document as context mitigates this; better than compressed worldview
ClaimClaim_ScopeFunctionalityInIntrospect When focus is provided, introspect returns selected_iris, subgraph (filtered TriG), existing_knowledge with provenance, gaps, and suggested questions.
ClaimClaim_ScopeIsNewSeparateTool Scope tool: takes focus and context like introspect, returns windowed graph to full depth in LLM-readable format. Effectively deprecates worldview:domain layers.
ClaimClaim_ScopeIsNotSeparate Scope is not a separate tool, it's a mode of introspect. The 'graph-scope' workflow was deleted.
ClaimClaim_ScopeVsIntrospect Analysis (gaps, tensions) is for picking the lens; content (claims, convictions, provenance) is for drafting from; mixing them makes both harder to use
ClaimClaim_ScrollDrivenAnimationMechanism The mechanism: a curated scroll-driven animation (popularised by NYT Snowfall) where scrolling drives the full pipeline visually. Beats: (1) typed user message; (2) conversations flow in as labeled boxes; (3) each conversation breaks into s
ClaimClaim_SEAL_Ambition SEAL is designed for ambitious long-term businesses with niche targets and deep founder expertise, not lifestyle businesses.
ClaimClaim_SecurityRequired an infrastructure advisor's infrastructure background and cautionary tale elevate this to Boulder; enterprise requirement validated
ClaimClaim_SelfIDAsQualifyingSignal When a customer contact says some version of 'I forget as much as I remember,' they have already diagnosed their own need. an engineering manager at the customer said it unprompted; so did the customer's chief architect. This is the cleanes
ClaimClaim_SeniorPeopleWouldSayYes Tony [35:24]: 'You could get easily all the people you need who are really senior or insightful people or who have a skill in a certain area that you value to do that. I can't imagine they would say no.' And [35:41]: 'And probably much more
ClaimClaim_SessionPatternReplicable The Interview → Save → Draft → Annotate pattern applies to any session that needs to produce a strategic document, content piece, or decision framework. The core insight: the interview elicits thinking not yet in collective memory, the save
ClaimClaim_ShareCodeAsKey Share code is the key, not sender identity. This structural property makes the gifting loop unbounded and forward-shareable by default.
ClaimClaim_SharedAgentMemoryProductFit Scarlet confirms existing product capabilities map to shared agent memory concept. Proposed stake; needs user validation.
ClaimClaim_ShiftsTerminology 'Shifts' is the user-facing term (not 'lenses'). 'Optometrist model' is an internal abstraction for the test-through-reaction review approach. These should remain distinct in documentation — shifts is what users see; optometrist model is ho
ClaimClaim_ShiftTrackingProductMoment the customer's chief architect surfaced a real product moment: the system shows him how a perspective is moving across transcripts (first upload: no idea; second: shifted; third: agreed, with named agreers cited). This is the conviction-evo
ClaimClaim_ShowDontTell Without visible proof of influence, the sales conversation stalls at 'trust me, it works.' a demo-call prospect's core feedback: show how aswritten influences AI behavior; agents should cite which memories shaped their recommendations.
ClaimClaim_SIIFillsProductTermSlot SII fills the '[product term — TBD]' slot the v2 product doc held open for the synthetic-identity meta-goal. Working coinage, not a final lock as of 2026-06-24.
ClaimClaim_SingleClaimFrameReductionist The current cite prompt asks 'what is THE claim that supports this text' — a single-match, reductionist frame that produces thin, wooden citations. The right question is 'what claims, narratives, and historical changes in the graph BEAR ON
ClaimClaim_SingleStreamExpertWorldview Expert plan uses a single-stream worldview, eliminating PR/branch complexity.
ClaimClaim_SMBBiggerThanDevs Scarlet's assessment: 'SMB is a bigger market than devs.' This drives the pivot from developer-first PLG to SMB as primary market thesis.
ClaimClaim_SMBBusinessPlan The business plan a demo-call prospect takes to prospects needs to reflect the SMB pivot. SMB is a bigger market than devs. Current business plan does not match the sales conversation a demo-call prospect is having.
ClaimClaim_SMBPivotReconciliationNeeded Open question: how to reconcile the enterprise-shape primary focus with the March 17 SMB/B2C pivot. The SMB/B2C pivot emphasized volume B2C; this refocus emphasizes enterprise-shape engagement pattern (which may or may not be SMB by org siz
ClaimClaim_SMSAgentsNowViable Self-hosted inference makes SMS/WhatsApp interface to organizational memory and always-on agents holding full perspective in context viable. These were previously cost-prohibitive under per-token API pricing. Fixed server costs change the e
ClaimClaim_SMSEnterpriseOnlyLeaning Leaning (not decision): SMS should be offered only for the enterprise plan, because numbers must be provisioned individually and managed. Scarlet's actual epistemic state: 'We don't know yet how this maps to tiers... I don't want to make th
ClaimClaim_SMSEntryOpenProductIdea The SMS-first entry point (text to start a topic, multiple groups with different reviewers) is currently described in the graph as an 'SMS-first' product, which is an overstatement. It is an open product idea, not the current product. The s
ClaimClaim_SMSFormatHonesty The SMS format makes AI personalization feel honest rather than fake because it removes the 'bot' mental frame associated with chat boxes.
ClaimClaim_SMSIsDistributionChannel The phone number / SMS is a distribution channel. It is one of several. It is the right channel in some contexts (enterprise, trust-signal value) and the wrong channel in others (spam association, prospect reluctance to share number, unjust
ClaimClaim_SMSIsProduct The SMS framing (text for topic / Salon) is actually a product. The git/MCP system is infrastructure. These are distinct layers.
ClaimClaim_SMSOperationalCost SMS is operationally expensive due to fraud maintenance, verification cycles, and ongoing compliance overhead. This cost is a permanent feature of running SMS at scale, not a one-time setup cost. It is one reason SMS may be appropriate only
ClaimClaim_SMSSpamFrictionKRC For the Knowledge-Rich Consumer archetype (a beta user from outside tech canonical), SMS-as-channel adds a trust friction that does not exist for the Knowledge Packager (an early prospect) or Delegator (the customer's chief architect). Spam
ClaimClaim_SMSTrustSignalChannelProperty The trust-signal value of SMS (a phone number doesn't feel like an AI chat) is genuine and not superseded by the channel correction. It is a channel property, not a product identity. Scarlet frames it as 'a bonus; it's an upsell.' Consisten
ClaimClaim_SortVsRankTensionResolved The ordering decision dissolved the long-standing sort-vs-rank tension in the compression-only projection principle. Order is unavoidable — hubs emit in some sequence — but under this decision the order is a fixed property of the graph stat
ClaimClaim_SourceAppendOnlyNeverDedup Source is append-only and never deduplicated. All dedup/merge/supersession lives in the reconciliation layer. This is a pipeline law, not a preference.
ClaimClaim_SourcePrivacyWedge Curated perspective preserves source privacy in ways RAG cannot, providing a concrete competitive wedge.
ClaimClaim_SourceSideScopeGate A source-side scope gate (internal calls only, exclusion rules) is what makes 'send me your transcripts' sayable to a client, given that client calls include their own customers, NDAs, and employees. Without the scope gate, the ask is ethic
ClaimClaim_SpeedRequiresConsensus The faster AI development moves, the more critical it is to maintain consensus on direction. Getting Claude to write a feature is easy; getting it to interact correctly with other features requires shared context.
ClaimClaim_SponsorNotChampionYouAreChampion Scarlet [25:07]: 'the customer's chief architect is interested, but he's not. He may be like ally, but I wouldn't say he's necessarily a champion. You know, he's like, excited about it, but isn't going to ram it through, you know.' Tony's f
ClaimClaim_StakeInGroundStrategy Strategy: Get a stake in the ground with one enterprise customer, then replicate that pattern as many times as possible to build a valuable list for the Second Wave.
ClaimClaim_StaleGraphCitationsWorse The danger with organizational memory: stale graph + citations = worse than having no citations at all. If the graph is inaccurate and someone is delegating work to AI that cites collective memory, the result is work with citations to somet
ClaimClaim_StalenessFail Staleness failure mode: a memory asserting a specific value (e.g., 'the gold is #CDA349') becomes incorrect the moment the source file is edited, causing the perspective to lie without anyone noticing. The file is the source of truth for th
ClaimClaim_StarterPerspectivesWeakTrigger In the delegation framing, generic/starter perspectives have no purchase trigger — nobody delegates a task that requires generic 'good writing.' In the solo professional framing, they might — if the perspective genuinely makes your AI bette
ClaimClaim_StartReviewVsReview 'start_review' renamed to 'review'. Drop the start_ prefix. You review knowledge, you don't 'start a review.' list_open_reviews folds into no-arg review.
ClaimClaim_StaticAssetDependencyModel What consulting clients pay for today: a static asset that makes them dependent on the consultant to execute, or requires an expensive in-house person to do a lossy interpretation and continue forward. This is the status quo the perspective
ClaimClaim_SteeringIsFoundingThesis Steering is also aswritten's founding thesis — narrative as the steering vector for a language model. The genus (steering) connects the product's founding insight to all three use cases, making the frame internally coherent and not merely a
ClaimClaim_SteeringVectorThesis Core thesis traced through entire career arc. Foundational to aswritten.ai positioning.
ClaimClaim_SteeringWeightDiscriminator The discriminator for rich-tier selection is the ontology's own types — claims, narratives, observations, evidence count as conviction-carrying; style-taxonomy record types are ordered later. This is not name-matching or hand-curation. The
ClaimClaim_StickyProductWordOfMouth Tony's assessment: aswritten has a sticky product. If Scarlet can get it into the hands of 20-50 people, the word of mouth will catapult it. The product is ready; the bottleneck is distribution.
ClaimClaim_StopGettingItToSeeYourWay Tony [01:05:22]: 'That's the thing that we need to change. Stop trying to get it to see the way that you see and start getting it to use you to give the perspective on what you see.' This reframes the human-AI relationship: instead of the h
ClaimClaim_StoriesRemoved Entire Stories system removed from product. TX summaries now returned from synchronous remember; blog post ideas, changelog, graph health become on-demand local skills. Workflows to delete: stories-list, story-draft, story-write, story-back
ClaimClaim_StrategyDocAsInternalSourceOfTruth The strategy document is the internal source of truth. From it, multiple documents need to be derived for different audiences: product doc, F&F fundraising doc, VC/Angels doc, intro deck, white paper, sales playbook, one sheet, public docs,
ClaimClaim_SurviveToWin The survival thesis: In technology, the winner is the one who survives. Survival means constantly riding the waves until it is your time to be at the front.
ClaimClaim_SurvivingNextTuesday Decision memory isn't just about surviving turnover; it's about surviving the next week's meeting.
ClaimClaim_SyncIntegrationDriftArgument Drift argument: async integration creates drift between the individual and the actual experience of the individual. The next extraction must wait for the previous to finish or it extracts against a stale state. Synchronous per-message integ
ClaimClaim_SynthesizedOverArtificial 'Synthesized' chosen over 'synthetic'/'artificial' because a synthesizer is an instrument a human plays (agency, participation) and foregrounds the act of synthesis — putting pieces together into a functional representation of what makes an
ClaimClaim_TaglineFinalized Tagline finalized after six iterations: 'Expertise from humans you trust. Cited by AI, aswritten.' Two-beat structure: trust (human source) then proof (AI citation). Scarlet iterated through six variations before landing on this form.
ClaimClaim_TaglineSupersession The tagline 'The product is a phone number.' in the Identity section of the worldview is superseded. Replacement positioning: 'The product is curated expertise installation. The phone number is the distribution channel for contexts where th
ClaimClaim_TeamAsPerspectives Roles are represented by aswritten perspectives contributed to by humans on cadence, rather than traditional full-time hires.
ClaimClaim_TeamPlanAtLaunch Team plan must be available at launch because collaboration prospects exist. Full automation (task-169 repo transfer) stays below the line but manual flow must work.
ClaimClaim_TeamTierGitHubBranch Team tier without GitHub: multiple authenticated users push to the same provisioned perspective on main. Each user reviews their own memories. No structured cross-user review. Scarlet: 'In Team without GitHub, a user reviews their own memor
ClaimClaim_TeamTierSalesCallIsOnboarding The sales call for the Team tier IS the onboarding IS the first knowledge extraction. No separate setup deliverable needed. The call produces the first memory, seeds the worldview, and demonstrates value. Setup fee for Team tier: leaning to
ClaimClaim_TeamVsEnterpriseTierCleanBreak Scarlet [16:20]: Team tier ([amount]/mo) is credit card signup — develop a perspective and have people read from it via MCP. Enterprise tier ([amount]/mo + [amount] setup) requires MSA because of channel provisioning (phone numbers, Slack i
ClaimClaim_TechnicalUserSeparationPrinciple Tony's insight: the best thing for technical users with existing AI workflows is to have a completely new project that hasn't started any repos, and then don't let it touch anything else. Keep them separate. The more involved people are, th
ClaimClaim_TemporalMarkerDesignImplication Four concrete design implications from the customer's chief architect's demo pollution failure: (1) Memories need an explicit speculative/aspirational/now/historical marker that the extraction pipeline preserves. (2) The marker should be ca
ClaimClaim_TextAndEyeChartSameOperation Both text (refraction mode) and eye chart (generation mode) are mechanically identical: run this prompt against the snapshot with this lens applied. Eye chart is a saved/curated prompt; user-provided text is an ad-hoc prompt.
ClaimClaim_TextFirstStagingGatedOnCost The full immersive/voice form of the individual product is gated on cost and model economics — may require frontier-quality reactions against the graph, which is expensive, or waits for open-weight models to reach current-frontier quality.
ClaimClaim_TheyDoItConversionModel Self-service conversion model: the expert creates the module themselves using the remember/review tools. aswritten provides the platform and extraction pipeline. Revenue model: marketplace percentage only — no upfront conversion fee. Scarle
ClaimClaim_ThirdBridgeMarketSegment a consultant who ran enterprise discovery engagements for fifteen years's brother uses Third Bridge-style reports for equity research. a consultant who ran enterprise discovery engagements for fifteen years's framing: 'you become one of the
ClaimClaim_ThreeErasKnowledgeDistribution Three eras of knowledge distribution: Books (static), Search (algorithmic), Curated expertise installation (human-curated perspective at point of need, cited and verifiable). aswritten is the third era.
ClaimClaim_ThreeFixVerifyMergeCycles Three fix-verify-merge cycles ran on the night of July 4 2026, each about thirty minutes. Bugs caught: (1) response-surface clobber, (2) corrupted digest commits (both caught by heartbeat between morning green and evening merge), (3) crash
ClaimClaim_ThreeLayerConvergence Independent validation of three-layer problem decomposition
ClaimClaim_ThreeLayerModel Three-layer model (Creation, Architecture, Infrastructure) independently validated by a founder beta user's execution → tactical planning → organizational memory framing
ClaimClaim_ThreeRankedIngestionMethods Three bulk ingestion methods ranked: (1) Full summarization across the set — what the customer's chief architect did, effective; (2) Per-transcript memory-by-memory ingestion — slow and overwhelming; (3) Target method — batched ingestion fo
ClaimClaim_ThreeTestQuestions Three diagnostic questions: (1) Could this be restated as an intent, decision, or relationship with a person behind it? → Perspective. (2) Is this a numerical value, code identifier, or config that means nothing outside its file? → Implemen
ClaimClaim_ThreeTierSupersedes The Feb 2026 simplified pricing ([amount]/mo per repo, 10 seats included, [amount]/additional seat) was too high for SMB and didn't differentiate between self-serve and consultant-led engagement. The new three-tier model (Individual / Team
ClaimClaim_TierRenameDecided Tier rename decided in earlier session on March 23: Individual→Expert, Publisher→Team. Not yet propagated in code. This is an open item from the session. The upgrade path is: interview → trial with perspective → Free (free forever) → Expert
ClaimClaim_TierStructureNamed Tier structure named in Strategy Document V1: Free (read-only), Individual ([amount]/month), Publisher ([amount]/month, 10 teammates), Organization ([amount]/month + [amount]setup) — renamed from 'Enterprise'. This supersedes the three-tier
ClaimClaim_TighterOnboarding a demo-call prospect requests clearer separation of what are docs vs AI, and the aswritten help repo. New users should not have to figure out which parts are documentation and which are AI-driven.
ClaimClaim_TonyCallUnresolvedEnterpriseB2C A ~Jul 1 call with Tony (per Scarlet; no recording) did not reach a clear conclusion on the broader enterprise-vs-B2C focus. The decision on whether the discovery wedge becomes the committed near-term focus is explicitly deferred. The enter
ClaimClaim_TonyPainFirstContinuity The context-server framing systematizes the January 8 deck decision driven by Tony's pain-first feedback. Tony's words then — 'people have no fucking clue what to do when they join a company' and 'just not having to stop what I'm doing to t
ClaimClaim_TopicAnalysisPlatformFeature The topic analysis step (Step 2) is currently entirely operator-driven and is a high-value activity that could become a Platform feature. The anti-pattern (direct historical ingestion) should be reflected in onboarding documentation and pos
ClaimClaim_TopicAnalysisStep2 Step 2 of the three-part seeding pattern: rather than ingesting six months of ARB transcripts directly, use Claude to perform a topic analysis across the corpus — 'what are these talking about? what are the open questions? what are frequent
ClaimClaim_TranscriptBootstrapOnramp Call transcript bootstrapping (Fireflies/Gemini/Zoom) is a lighter onramp that allows for an immediate working perspective without a live interview.
ClaimClaim_TranscriptDemotedFrictionStory The transcript 'gold mine' is a friction story, not a value story — the market priced transcripts at roughly [amount]per seat; they are inert. Transcripts lead the onramp; pain leads the value story. This refines rather than contradicts the
ClaimClaim_TranscriptNowKnownConstant Meetings shifted from rooms to calls (accelerated by lockdown), making the transcript a known constant — the precondition for aswritten's model. Scarlet: 'the transcript is now a known constant.' a consultant who ran enterprise discovery en
ClaimClaim_TranscriptSurpriseSignal The fact that the enterprise customer was surprised to learn their own historical ARB transcripts were an asset is itself a signal — not just evidence that transcripts are useful, but evidence that aswritten is failing to communicate this a
ClaimClaim_TransformIsBetterVerbThanQuery 'Transform' is a better verb than 'query' because perspective shifts aren't exclusive filters. They change weighting and focus — in some cases removing triples, in others changing their weighting. A query implies exclusion; a transform impl
ClaimClaim_Trigger1SoloProfessional Trigger 1: 'I can't get AI to do my work without babysitting it.' Archetype: a beta user from outside tech. The solo consultant or professional who uses AI daily for complex domain-specific work. Every session, she re-explains the client's
ClaimClaim_Trigger2Delegator Trigger 2: 'I need someone else to work without me.' Archetype: the customer's chief architect. The team lead, practice owner, or senior expert who needs other people (or their AIs) to work with their expertise. Specific moment: hiring some
ClaimClaim_Trigger3ExhaustedExpert Trigger 3: 'I don't want to be their only option for this anymore.' Closes the activation gap by addressing social pain.
ClaimClaim_TriGOverJSON Subgraph format is TriG, not structured JSON. JSON was 1.94x the token cost with no information gain.
ClaimClaim_TrimodalCompLandscape Three-tier competitive landscape for mid-market ICP: Tier A (Light SaaS, what mid-market currently uses) — typical AI tooling stack at 150-person co [amount]-15K/mo, limitation is no sovereignty/citation/on-prem/agent deployments. Tier B (H
ClaimClaim_TwoCallRigidity Two-call structure rigidity flagged at fact review. Hybrid model from an early prospect call not yet captured in TX. Conviction level treated as Stake (flagged, not yet resolved).
ClaimClaim_TwoSidedLegibilityInverse an early prospect's advertising problem (practitioner-side flattening) is the inverse of Scarlet's patient problem (patient-side illegibility). They are the same problem from opposite ends. aswritten makes the practitioner's perspective leg
ClaimClaim_TwoToSixCoreTeam Tony's view: AI-native companies need 2–6 senior people who control AI tooling; AI assistance means two people can handle defining, categorization, and documentation that previously required larger teams.
ClaimClaim_TwoYearWindowTiming Tony's prediction: 'You've got a two year window' before legal/corporate pressure reshapes AI landscape. Settled view based on prior experience (Internet boom pattern).
ClaimClaim_TXIsPrompt A TX is really just a structured description of a change; refract can accept an arbitrary prompt as the change description
ClaimClaim_TXOnlyBundles Perspective shares bundle only .aswritten/tx/*.sparql files, excluding memories and manifest.
ClaimClaim_TXSummariesAutoGenRemoved TX summaries on PR removed as auto-generated feature. Agent provides this from the returned transaction. Keep the PR comment mechanism in tx-post-commit — not hurting anything.
ClaimClaim_TXSummariesOnPRRemoved TX summaries on PR removed as auto-generated feature. Agent provides this from the returned transaction. PR comment mechanism in tx-post-commit kept — not hurting anything.
ClaimClaim_TxSummaryNoLatency The tx_summary is generated during extraction so it adds no latency to remember. It is not a post-processing step — it is produced as part of the extraction pipeline run.
ClaimClaim_TypologyAxisCollision The archetype typology mixed two orthogonal axes — AI-fluency (tinkerer vs consumer) and distribution-motion (packager vs solo-practitioner) — and classified a beta user from outside tech on fluency only. This is the underlying error that p
ClaimClaim_UnderstandingNeverSells Understanding never sells — outputs do. This is the third demotion from the June 10 session, refining the value story away from comprehension and toward deliverable outputs.
ClaimClaim_UnifiedInstrumentPattern The 60-day ICP drift (enterprise → SMB → B2C → transcripts) was not about different customers — it was about different deployments of the same instrument. The Scarce-Expert Filter with Goal-Gated Escalation unifies an engineering manager at
ClaimClaim_UnifyingInsight Claim: enterprise engagement stages map 1:1 onto use-case taxonomy primitives (Pilot=Salon, Full=Teams, Retainer=Parallel). Settled enough to drive prioritization decisions; would require significant counter-evidence to shift.
ClaimClaim_UseCase2Underrepresented Use Case 2 (agentic steering) is currently underrepresented in aswritten's materials — the deck and fundraising brief carry Use Case 1 but not Use Case 2. Foregrounding Use Case 2 is the main thing the steering frame corrects. This is a gap
ClaimClaim_UseCase3MethodologyLayer Use Case 3 is the methodology layer — the shared process underneath many engagements, extracted and aligned so one memory substrate serves every team, human or agent. It addresses the commonalities across different engagements, different cl
ClaimClaim_UserDomainCommentaryRequired The methodology assumes the customer has domain experts who can review and annotate the topic analysis output. The user — not aswritten.ai staff — provides the architectural commentary at Step 2. Without that, the seeding will produce spars
ClaimClaim_ValueDemoPrerequisite Clear, visible proof that aswritten changes how AI works is a prerequisite for raising bigger money. a demo-call prospect's framing: 'Show, don't tell.' Agents should cite which memories shaped their recommendations (task-145).
ClaimClaim_VectorSearchNotUsedInPapers a collaborator's own observation about the production gap: 'Vector search and its classic form was never really used um uh for any of the papers um which I covered. That's kind of was surprising to me.' None of the five agent-simulation pap
ClaimClaim_VerbatimFlaggingAtReview Landmark memories may quote primary sources at length. Wherever ingestion stores source content verbatim, the protocol surfaces that explicitly at review time. Flag mechanics and metadata schema: Workstream 1 spec (backlog).
ClaimClaim_VettingAsBeachhead Vetting (patient vetting, project vetting, client vetting) is the specific entry-point use case — the beachhead. The real product is creating a perspective that represents an individual. The bigger frame is collaboratively configurable/buil
ClaimClaim_VisibilityIsInvariantNotBudget Visibility is guaranteed for every area by the index — it is not a budget decision. Richness (full-digest rendering) is the budget decision. A hub that loses its rich slot is still named, counted, and one expand away. This is the invariant
ClaimClaim_VisualLanguageInfrastructure Simple geometric shapes (circles and lines) represent the 'infrastructure' layer of the technology receding into the background.
ClaimClaim_VoiceDroppedFromRoadmap Voice interviews are explicitly dropped from the roadmap. Native voice-to-text on phones eliminates the need for Retell/Vapi integration complexity.
ClaimClaim_VoiceMemosAndTranscriptsPrimaryInputs Scarlet herself uses call transcripts and voice memos as her two primary inputs into aswritten. The fact that the customer's chief architect had to ask, and Scarlet had to confirm, reveals this is not being communicated as the default usage
ClaimClaim_VouchEmailMotionTemplate The two Vouch emails are a template for a motion, not for words. The four steps: (1) person-first outreach through named human threads, honest account of situation, small ask; (2) call that runs on values resonance, listening for the other
ClaimClaim_VouchTextProcessPrecursor Scarlet recognised a design collaborator's process as the same thing she built at Vouch with text: a bunch of files representing how the team was thinking — the equivalent of design tokens for product language — put at the top of the chat t
ClaimClaim_VulnerabilityPersonFirstOutreach The tenor of the Vouch outreach emails was 'very very different from everything I've been sending about aswritten lately — much more vulnerable about my personal situation, much more connected to individuals related to the company.' The fir
ClaimClaim_WarmStartSkipsIntro Warm-start entry (from remember's return, same session) skips: intro (just explained in the offer), TX summary (already shown in shift summary), zone identification as separate step (already shown). Test document: agent suggests from sessio
ClaimClaim_WedgeDensityFirst Density first, geography second. The goal is 50 active practitioners in one referral neighborhood, not 50 scattered practitioners across cities. an early prospect and a practitioner beta user Barber opening a collaborator together, with a r
ClaimClaim_WeDoItConversionModel Professional conversion model: aswritten takes the expert's book, course, training materials, or consulting package; runs extraction; builds the module; curates the perspective; hands them a living, installable, citable module ready for dis
ClaimClaim_WhatWasSaidExceedsWhatWasDecided 'What was said' deliberately exceeds 'what was decided' because it is not just decisions — notions, open arguments, and half-formed positions steer thinking too. This is what conviction levels exist to carry. The distinction is the objectio
ClaimClaim_WhitePaperAIFiltered Tony's insight: if someone sends him a white paper, he'll put it into Claude and ask for a summary rather than read it. The implication: the white paper format is wrong for initial executive engagement. GIF-based approach first, white paper
ClaimClaim_WorldviewAsPrimaryProduct Central to platform positioning; hard to move; requires multi-party consensus
ClaimClaim_WorldviewBottleneck The next bottleneck in AI productivity is not LLM capability but the absence of a versioned, coherent organizational context.
ClaimClaim_WorldviewBuilding Works with you to understand your org... comprehensive worldview. Foundation-level from competitive analysis session (Feb 5, 2026).
ClaimClaim_WorldviewDomainLayersDeprecated worldview:domain layers are rarely useful and have only been called a few times. Scope with arbitrary focus effectively deprecates them.
ClaimClaim_WorldviewEncodesAIBehavior Common AI criticisms — 'good catch', 'you're right', collaborative flatness — are the kind of behavior that worldview encoding can address. This is the concrete behavioral case for modules beyond information delivery.
ClaimClaim_WorldviewSummaryIsAlreadyLens compile(layer=worldview) is ~4-8K tokens, already editorially compressed — the everything-weighted-equally-and-compressed lens. Shifting a summary shifts lossy compression, not knowledge. The lens needs full-depth content for the domains it
ClaimClaim_WritingForAIs Writing for the AIs is a strategic move to shape how LLMs perceive and trawl intellectual work.
ClaimClaim_YC_Application_1 Frontier launch strengthens YC application 'Why now?' narrative with concrete same-week validation of AI agent shared context need.
ClaimPrinciple_AppropriatenessGate A perspective-grounded agent must refuse to act if the principal's principles forbid the action (e.g., refusing cold outreach if warm-intro-only is a principle).
ClaimPrinciple_DisclosureDefault Agent outreach uses disclosure by default, with the principal visibly present in the coordination substrate able to take over as themselves.
Observation (532)
ObservationActor_CustomerArchitect_TinkererRef the customer's chief architect referenced as the canonical Tinkerer archetype example (C1). Specifically needs cross-repo memory he owns at rest with compliance control. His existing CLAUDE.md-based solution is a barrier to adoption — somet
ObservationActorPain_CodingAgents Coding agents (Claude Code, Cursor) rely on brittle personas and scattered docs; AI agents forget context between sessions, docs go stale.
ObservationActorPain_ContentGeneration Content generation produces generic output disconnected from actual organizational perspective.
ObservationDeckUpdate_DevOpsPersona Add DevOps persona to 'Connected AI Agents' section; emphasize deployment strategies, environment management, rollback procedures as use case
ObservationDeckUpdate_GitLabSupport Mention GitLab support alongside GitHub; address platform assumptions for broader market
ObservationDeckUpdate_GTMFraming Strengthen 'How we do it' with 'train the trainers → here's the solution' framing; clearer than current consulting-first language
ObservationDeckUpdate_PropagationMechanics Clarify propagation mechanics (git submodule hierarchy); currently hand-waved as 'all materials update automatically'
ObservationDeckUpdate_SecurityFiltering Add content filtering/security scanning to 'Under the Hood'; address inbound API key/credential detection before enterprise sales
ObservationFocusSequencing_ADR002 Developer-first adoption (PLG motion): Solo developers / small teams using Claude Code, Cursor, Windsurf. Hook: 'Add collective memory to your repo, never write README again'. Expansion: Once developers see value, expand to executive/conten
ObservationMoatLeverage_ADR002 Unique Value: Same collective memory snapshot backs all three use cases. When sales learns a customer pain point, that memory informs product roadmap, feature implementation, and marketing content. Eliminates strategy-execution disconnect.
ObservationObs_ABACFiltering a consultant who ran enterprise discovery engagements for fifteen years's named-graph ABAC filtering mechanism
ObservationObs_ActorTransformTemporalityQuote Scarlet's statement on what actor transforms should surface — temporality over conviction-settled items.
ObservationObs_AdvisoryStaffingModelQuote Scarlet's verbatim articulation of the advisory staffing model structure — monthly fee, call cadence, decision review, agent audit.
ObservationObs_AgentificationMDashQuote The specific observable symptom of agentification homogenization: M-dash, quoting, 'load-bearing' as the converging register.
ObservationObs_AgentsCiteMemories a demo-call prospect's core insight: agents should cite which memories shaped their recommendations. This is the value demonstration problem (task-145). The product needs to show how aswritten influences AI behavior — visible proof of influ
ObservationObs_AIAsNewHireQuote Scarlet's bridging insight connecting both triggers through the new hire metaphor.
ObservationObs_AIAsShip Scarlet's product moment articulation: AI as ship, human as guide holding the tiller. Core metaphor for the product vision.
ObservationObs_AIOpinionQuote Scarlet's articulation of AI's inherent but singular opinion — the problem modules solve.
ObservationObs_AltitudeSwitchingQuote Scarlet's verbatim identification of altitude-switching as the specific cognitive limit — distinct from context-switching.
ObservationObs_AmbientBehaviorCorrection Scarlet's correction establishing the on-demand vs. ambient distinction for product behavior.
ObservationObs_AnnotationCoverage 80 claims analyzed in strategy document, 64 supported by collective memory. 80% aggregate coverage across 14 source memories spanning Jan 2025–Mar 2026. 20% uncommitted = general market framing or new strategic assertions.
ObservationObs_AntiHallucination Anti-hallucination via citation requirement
ObservationObs_AppendOnlyLog Conviction-over-time query capability from append-only log
ObservationObs_AreYouTalkingAboutTheseQuote Scarlet's articulation of 'are you talking about these?' as the legibility differentiator
ObservationObs_ArtifactsLoseWhy Scarlet's verbatim articulation of the core thesis
ObservationObs_AsWritten 'As written' is both the company name and the editorial policy: ship things as they are, process content over polished content.
ObservationObs_AsyncDriftArgumentQuote Scarlet's drift argument for synchronous integration
ObservationObs_AudienceMapping Four distinct audience resonance frames: Enterprise buyer (governance), Series A CEO (contractor autonomy), Ops lead (trust), Technical leader (accountability via citations and review).
ObservationObs_AuthoritativeSpaceCyclicalQuote a design collaborator's framing of the authoritative space as a cyclical credibility mechanism
ObservationObs_AutoUpdateQuote Scarlet's verbatim articulation of the auto-update mechanism as a core module property.
ObservationObs_B2CAngleQuote Tony's articulation of the B2C angle — the single people building whole companies on their own as the primary target.
ObservationObs_B2CSellingStyleQuote Scarlet's self-diagnosis: B2B default is a career-history artifact, not a considered fit; her selling style is native to B2C
ObservationObs_BaseCampQuote Scarlet's canonical base-camp metaphor: 'I had a bit of an internal sit the fuck down at the finish line kind of pit. Now I'm looking up, and there's still a mountain to climb to get into base camps, a long track, but this is definitely bas
ObservationObs_BootstrappingGuidance Scarlet's live onboarding guidance on what to ingest first — the content that should be automated in an onboarding mode.
ObservationObs_BottleneckProcessesQuote Scarlet extending the bottleneck beyond individuals to processes — the founding evidence for Claim_BottleneckBeyondIndividuals.
ObservationObs_BoundedDomainQuote Scarlet's verbatim articulation of the bounded-domain fidelity argument — the structural case for expertise modules over org memory.
ObservationObs_BrandingCharacterQuote a design collaborator's 'branding 101' argument: give the business a human character people can connect to
ObservationObs_CascadeReaction Knowledge graph propagation demo: slogan change cascading through graph. a founder beta user saw non-obvious downstream implication: 'It's less intuitive that you have to update job descriptions.'
ObservationObs_CEOContextServerQuote Scarlet's canonical articulation of the context-server pain — the founding evidence for Narrative_CEOContextServerReframe.
ObservationObs_ChatNotCodingSession Scarlet's guidance on first-use framing: treat aswritten as a chat session, not a coding session.
ObservationObs_ChineseHerbalismLayeringExample Concrete example of layering: three distinct expertise perspectives compiled together in a single chat session.
ObservationObs_CitationGroundingError The citation tool, when processing a draft that quoted an engineering manager at the customer's self-description ('I forget as much as I remember'), grounded the quote to the customer's chief architect — because the perspective held the cus
ObservationObs_ClaudeCodeJSONL Claude Code persists full session transcripts as plaintext JSONL, directly readable by the agent. Scarlet: 147 sessions across 19 projects. Zero new infrastructure needed for session-history ingestion.
ObservationObs_CollaborationTaxonomyList The four collaboration shapes: Individual→Self, Individual→Team, Individual→Many (publisher), Team→Many. Missing from the strategy doc's three-archetype model.
ObservationObs_CompetitiveLandscape_Frontier OpenAI Frontier: semantic layer for enterprise; could be DOWNSTREAM of aswritten; assumes knowledge exists in systems
ObservationObs_CompetitiveLandscape_MultiPlayer Conversation-side vs. contribution-side multi-player AI distinction; aswritten is contribution-side
ObservationObs_CompetitiveLandscape_Notion Notion/Confluence graphs: indexes existing pages; goes silent when strategic knowledge lives in founder's head
ObservationObs_CompetitiveLandscape_RAG RAG pipelines: retrieves existing docs, no architectural opinion, cannot extract undocumented knowledge or steer vs. dilute
ObservationObs_CompilationLayers Compilation layers serve right depth of context: 4K-token worldview bootstrap to full 100K-token domain expansion
ObservationObs_CompileDeprecated compile tool deprecated in practice
ObservationObs_CompilerSentenceQuote Identified in draft flags as the cleanest one-sentence mechanism statement yet written. Should survive every edit.
ObservationObs_ConsultingFramingQuote Scarlet's verbatim articulation of the core framing decision — the founding evidence for Narrative_ConsultingIsMeUsingProduct.
ObservationObs_ContentCommentaryKeystoneQuote Scarlet's articulation of the content/commentary keystone
ObservationObs_ContentCreatorPublishingQuestion The reframing of the publishing question for the AI age — from content creation to AI access.
ObservationObs_ContextHemorrhage a consultant who ran enterprise discovery engagements for fifteen years on meetings as context hemorrhage
ObservationObs_ConvergenceAsymptotic Convergence is asymptotic accumulation of review evidence, not a binary threshold. More honest to how organizational knowledge actually works.
ObservationObs_ConvictionDecayQuestion a founder beta user's spaced-repetition framing for conviction decay — a feature request for temporal graph analysis.
ObservationObs_ConvictionLabelPlainEnglish Conviction level rename rationale: notion and foundation already worked as plain English; stake and boulder were metaphors requiring a key to decode. New labels: notion (unchanged), claim (was stake), decision (was boulder), principle (was
ObservationObs_ConvictionLevels Conviction levels as product feature: Notion → Stake → Boulder → Foundation signal how settled each claim is
ObservationObs_ConvictionTemporalityQuote Scarlet's canonical statement distinguishing conviction from temporality as orthogonal axes.
ObservationObs_CoreFrustrationCompile Scarlet's opening statement identifying the core problem with 'compile' as a tool name.
ObservationObs_CoreNarratives_1 Core narrative arc: enterprise pilot = Salon, full engagement = Teams, retainer = Parallel. Same primitives, different initiator. This is the unifying insight that could resolve the two-motion tension.
ObservationObs_CoreNarratives_DemoMapping Three product demos map to escalating solution complexity: Solo (After the Call), Salon (Text Your Topic), Teams (The War Room).
ObservationObs_CorePrinciple The call is explicitly not a product demo but a consulting engagement session
ObservationObs_CoreThesisQuote1 Scarlet's governing principle for all scoping decisions: focus on bounded behaviors done extremely well
ObservationObs_CoreThesisQuote2 Apple comparison: opinionated workflow, not a tool that does everything
ObservationObs_CPCPricingQuote Scarlet's verbatim articulation of the CPC-on-goal-URL pricing model — the founding evidence for the new pricing narrative.
ObservationObs_CreatePerspectivePitchQuote Scarlet's verbatim articulation of the create-a-perspective pitch and the Fireflies integration as a strong demo.
ObservationObs_CuratedDroppedQuote Scarlet's verbatim rationale for dropping 'curated' from the category name.
ObservationObs_CustomerArchitectCadence the customer's chief architect's nightly ARB transcript ingestion is live proof the protocol works before it was formalized. The first enterprise user already runs this loop by hand.
ObservationObs_CustomerArchitectCloneMeQuote the customer's chief architect's verbatim articulation of the Trigger 2 (delegator) pain.
ObservationObs_CustomerArchitectCrossRepoQuote the customer's chief architect's articulation of the cross-repo memory problem that nobody is solving.
ObservationObs_CustomerArchitectCuratedMemory the customer's chief architect articulates the curated-memory wedge argument against code-grounded AI
ObservationObs_CustomerArchitectDailyTranscript the customer's chief architect describes his daily ARB transcript ingestion cadence
ObservationObs_CustomerArchitectDemoPollutionQuote the customer's chief architect's verbatim account of the demo pollution failure — the founding evidence for Claim_CustomerArchitectDemoPollutionProductEvidence.
ObservationObs_CustomerArchitectEnterpriseCustomerFitsPattern Scarlet's verbatim observation that the the customer's chief architect/the enterprise customer use case fits the same structural pattern as an engineering manager at the customer's — the founding evidence for the unified instrument claim.
ObservationObs_CustomerArchitectFullLoopPattern the customer's chief architect's full-loop pattern: perspective → introspect → remember → cite, distributed across all four core tools. The 'expert' archetype — captures, asks, reads, and grounds.
ObservationObs_CustomerArchitectNotComfortableQuote the customer's chief architect's closing statement reinforcing the curated-not-dumped principle — the champion's own words validating the same thesis as the CTO.
ObservationObs_CustomerCTOAISlopQuote the customer's CTO's verbatim articulation of the AI slop problem — the economic buyer independently naming the exact problem aswritten solves.
ObservationObs_CustomerCTOClosingMarketing the customer's CTO's closing reaction to the the enterprise customer description demo — 'marketing department' line
ObservationObs_CustomerCTOExcitedCuratedKnowledge the customer's CTO's primary excitement: curated knowledge accessible to customer solution architects without exposing source code.
ObservationObs_CustomerCTOInformationCommodity the customer's CTO's articulation of the core value of perspective and collaboration over information and code — independently arriving at aswritten's central thesis.
ObservationObs_CustomerCTOPRMetaphorQuote the customer's CTO's verbatim articulation of the PR-with-unit-tests metaphor for memory ingestion — the founding evidence for Claim_CustomerCTOPRMetaphorCTOValidation.
ObservationObs_CustomerCTOUsageVisibilityRequest the customer's CTO asked for 'usage visibility' during the 2026-05-07 kickoff. The response was not a dashboard but an MCP tool — validating the agent-as-interface thesis.
ObservationObs_CustomerEngManagerForgetQuote1 an engineering manager at the customer's verbatim self-identification as bottleneck candidate — the founding evidence for Claim_CustomerEngManagerSelfIDBottleneck.
ObservationObs_CustomerEngManagerForgetQuote2 an engineering manager at the customer's extension of his self-diagnosis — the volume and specificity of what the ARB facilitator must hold.
ObservationObs_CustomerEngManagerLag24h an engineering manager at the customer pushes hardest on the 24-hour transcript lag operational pain
ObservationObs_CustomerEngManagerNotepadValue an engineering manager at the customer volunteers demo value unprompted
ObservationObs_CustomerEngManagerPilotDemoQuote Scarlet's verbatim statement on the an engineering manager at the customer pilot — establishing that it is not really a demo but a real deployment.
ObservationObs_CustomerEngManagerUseCasesQuote The three an early prospect use cases that establish the unified instrument pattern: patient vets fit, referring practitioner decides appropriateness, student filters questions.
ObservationObs_CustomGPTCorrection Scarlet's correction distinguishing Custom GPTs from aswritten's perspective-based approach.
ObservationObs_DatabaseConsistency Datomic transactor lineage: database consistency as audit-log property
ObservationObs_DataLineQuote Core data-philosophy statement: aswritten is processor not store. Any 'give us your data' framing is wrong and should be corrected on sight.
ObservationObs_DavidArchaeologyQuote David Nolen's articulation of the decision-archaeology use case — the founding evidence for Narrative_MoneyAnswerDecisionArchaeology.
ObservationObs_DavidConvictionQuote David Nolen asked for conviction unprompted — the founding evidence that the conviction/temporality distinction resonates with technical users.
ObservationObs_DavidKnowsWhatItDoesntKnowQuote David Nolen on the knows-what-it-doesn't-know feature — anti-hallucination value landing.
ObservationObs_DavidMicroservicesBlindersQuote David Nolen's microservices/blinders argument — the audience question pushed three times; founding evidence for Claim_DeveloperBlindersAudienceQuestion.
ObservationObs_DavidNolenArchaeologyEmotionalPeakQuote David's emotional peak — the 'how did we end up here' moment, self-generated
ObservationObs_DavidNolenArchaeologyQuote David's self-constructed archaeology query use case
ObservationObs_DavidNolenIntrospectQuote David's independent validation of introspect-as-escalation / 'it knows what it doesn't know'
ObservationObs_DavidNolenProvenanceDatesQuote David's unprompted validation of provenance and dates as differentiators
ObservationObs_DavidNolenWhiteSpaceNonDeterministicQuote David sharpening the white-space coinage
ObservationObs_DavidProvenanceQuote David Nolen's reaction to the provenance/citation feature — the primary value-landing moment.
ObservationObs_DeckIsArtifact Scarlet's articulation that the deck is a derived artifact and the real consulting value is the perspective shift — the module the org can install.
ObservationObs_DeliEmailOpinionatedQuote Scarlet's articulation of why the deli email worked — opinionated perspective as the hiring signal
ObservationObs_DeliStoryQuote1 Core line from the deli story email: the recognition framing that rebranded Vouch.
ObservationObs_DeliStoryQuote2 The pivot line from the deli story email: reframing Vouch's positioning from power/control/security to respect/comfort/invitation.
ObservationObs_DeliStoryQuote3 The tagline that closed the deli story email — the rebranding of Vouch in three words.
ObservationObs_DemocratizedConsensus Scarlet's framing of knowledge democratization as a counter to political influence in organizational decision-making.
ObservationObs_DemoProspectCommissionGapQuote Scarlet's framing of the value-driven upgrade from Team to Professional — consultant-led extraction as the pull factor.
ObservationObs_DemoProspectMaterialPriorities a demo-call prospect's six material priorities in order: (1) pre-read deck, (2) tighter onboarding, (3) sales guide, (4) demonstration of value, (5) updated business plan (SMB), (6) updated landing page. These are a demo-call prospect's dir
ObservationObs_DesignCollaboratorAIOutputQuote a design collaborator on the output not reading as generic AI
ObservationObs_DesignCollaboratorAmplificationQuote a design collaborator's verbatim articulation of the amplification framing — AI amplifies craft, not uniformity.
ObservationObs_DesignCollaboratorLensIrreducibleQuote a design collaborator naming the lens as the irreducible, defensible core
ObservationObs_DesignCollaboratorTokenProcessQuote a design collaborator's description of his extraction→tokens→surfaces process
ObservationObs_DifferentiatorFolded Differentiator folded into the 'what we do' paragraph rather than isolated as a separate section. 'Narrative architecture' replaced with what it actually means in human terms.
ObservationObs_DigestBrowserUIQuote Scarlet's description of the digest browser surface
ObservationObs_DirectIngestionDriftQuote Scarlet's verbatim articulation of why direct historical ingestion introduces drift — the founding evidence for the anti-pattern claim.
ObservationObs_DiscoveriesRemovalQuote Discoveries fully removed, not just disabled
ObservationObs_DiscoveryConsultantAgentBoundaryQuote a consultant who ran enterprise discovery engagements for fifteen years's articulation of the agent/human boundary problem — the organizing thread of the call.
ObservationObs_DiscoveryConsultantBobAnnaCharlieQuote a consultant who ran enterprise discovery engagements for fifteen years's verbatim articulation of the provenance value in discovery reports — Bob, Anna, Charlie as named stakeholders whose perspectives are preserved.
ObservationObs_DiscoveryConsultantBreadcrumbsQuote a consultant who ran enterprise discovery engagements for fifteen years's articulation of the breadcrumbs criterion — the re-runnable, zero-state property of the cold sub-agent test.
ObservationObs_DiscoveryConsultantChineseWhispersQuote a consultant who ran enterprise discovery engagements for fifteen years's verbatim Chinese Whispers framing — the founding evidence for Narrative_ChineseWhispersProvenance.
ObservationObs_DiscoveryConsultantClipWingsQuote a consultant who ran enterprise discovery engagements for fifteen years's verbatim 'clip the wings of the raven' articulation — the founding evidence for Claim_DiscoveryConsultantClipWingsRaven.
ObservationObs_DiscoveryConsultantDiscoveryToolQuote a consultant who ran enterprise discovery engagements for fifteen years's verbatim articulation of the discovery-tool use case — the founding evidence for Narrative_DiscoveryToolGTMWedge.
ObservationObs_DiscoveryConsultantFalsifiableQuote a consultant who ran enterprise discovery engagements for fifteen years's falsifiability framing — the Boolean property of the cold sub-agent test.
ObservationObs_DiscoveryConsultantFortuitousTimingQuote a consultant who ran enterprise discovery engagements for fifteen years's combustion-engine/oil analogy for fortuitous timing — the founding evidence for Narrative_FortuitousTimingPair.
ObservationObs_DiscoveryConsultantHATEOASQuote a consultant who ran enterprise discovery engagements for fifteen years's verbatim HATEOAS/navigation framing — the founding evidence for Claim_DiscoveryConsultantHATEOASParallel.
ObservationObs_DiscoveryConsultantOAuthScopeQuote a consultant who ran enterprise discovery engagements for fifteen years's OAuth-scope formulation of per-task agentic permissions.
ObservationObs_DiscoveryConsultantOnlyTestQuote a consultant who ran enterprise discovery engagements for fifteen years's canonical articulation of the cold sub-agent test as the only test that counts.
ObservationObs_DiscoveryConsultantPlayback a consultant who ran enterprise discovery engagements for fifteen years played back the processor-not-store framing exactly, adding the two-pipe model and trial/exit narrative
ObservationObs_DistillationQuote Scarlet's verbatim rationale for the distillation → perspective creation rename.
ObservationObs_DocumentEcosystemTable The full document ecosystem table: nine derived documents from the strategy doc, each with audience and relationship to strategy. Includes product doc, F&F doc, VC/Angels doc, intro deck, white paper, sales playbook, one sheet, public docs,
ObservationObs_DomainsNotRealQuote Scarlet's articulation of why domain-focused compiles fail in practice — humans don't live in single domains.
ObservationObs_EarlyProspectHubDeployment an early prospect (a collaborator): small business owner + assistant. Hub deployment pattern. Healthcare context raises HIPAA/on-prem questions (open question). Some SMB will need on-prem deployment — HIPAA compliance is an open question fo
ObservationObs_EarlyProspectInboundBottleneck Scarlet's verbatim observation that an early prospect fields inbound inquiries constantly from clients and prescribing practitioners — the same inbound-bottleneck pattern as the enterprise customer, despite being a 5-person org.
ObservationObs_EarlyProspectLivingModuleQuote an early prospect as the canonical example of the living module pitch — book already written, could become an evolving expertise package.
ObservationObs_EnterpriseCustomerARB the enterprise customer ARB use case: after ingesting the ARB, anyone can ask their AI a question, get an answer that cites the architect
ObservationObs_EnterpriseCustomerTranscriptSurpriseQuote Scarlet's verbatim account of the enterprise customer's surprise — the founding evidence for the pitch gap diagnosis.
ObservationObs_EnterpriseShapeDefinition Scarlet's verbatim articulation that enterprise shape is the engagement pattern (inbound bottleneck + stakeholder access), not org size — the same shape applies to a 5-person practice and a 150-person consultancy.
ObservationObs_EscalationVocabulary Four perspective-grounded escalation types
ObservationObs_EVConsensusAnswerQuote The committed voices statement that answers EV's consensus-view prompt — the specific sentence that locks the EV application's answer to Cowen's 'start from a consensus you actually agree with' prompt.
ObservationObs_ExpansionPathway_1 Consumer play could generate bottom-up enterprise demand, prove product without consulting labor, and create content pipeline automatically.
ObservationObs_ExpertReachCorrection Scarlet's correction reframing the expert relationship — expansion not replacement.
ObservationObs_ExpertSingleStreamQuote Expert plan: single-stream worldview, memories save to main by default
ObservationObs_ExportableDataBlade The aphoristic compression of the ownership argument: exportable data ≠ ownership; right to be forgotten ≠ freedom. Two parallel antitheses that sharpen the scanner essay to a blade.
ObservationObs_ExportableDataOwnershipQuote The ownership spine in two sentences — the sharpest distillation of the SII political argument.
ObservationObs_ExtractionEvalsQuote Extraction quality declared the most important part of the product
ObservationObs_ExtractionScopeQuote Scarlet's correction of 'domain' framing — the Professional tier builds organizational worldview, not a siloed domain index.
ObservationObs_FableAntiLockInPolitics The rhetorical question that radicalizes the anti-lock-in moat into a political claim about means of production.
ObservationObs_FeatherArrowsClose The closing metaphor of the flocking piece — archery/arrows as the productivity argument. Draft flag: the EV version should open out to the voices question rather than close on productivity.
ObservationObs_FeedbackHelpQuote Feedback/help tool as primary churn-prevention mechanism
ObservationObs_FirefliesWebhookQuote Scarlet's verbatim description of the Fireflies/webhook pipeline that should move from 'previously existing idea' to 'actively pitch.'
ObservationObs_FirstThingYouInstall This is going to be the first thing that people install when they start a new company. Like you would never consider starting a software company without GitHub and git
ObservationObs_FiveBeatArc Five-beat arc: (1) Fear/hook, (2) Process/differentiator, (3) Proof/citation, (4) Loop/write-path+governance, (5) Payoff/compounding value. Each beat maps to a demo segment on Call One.
ObservationObs_FlockingPieceVerbatim Verbatim opening of the flocking piece (primary source, Scarlet draft). Establishes the flocking origin story as the EV application opener.
ObservationObs_FormulaNotNumbers Numbers can work for you or go against you. The formula is what you need, not the numbers.
ObservationObs_FounderFitB2CQuote Scarlet's verbatim founder-fit rationale — the clearest statement of why B2C is fit, not just market size.
ObservationObs_FounderPowerUserQuote Scarlet's self-diagnosis as a poor user example — the power user par excellence whose graph stays current in ways typical users' will not.
ObservationObs_FounderShiftPivotQuote a design collaborator's explicit statement that Scarlet is the durable through-line across pivots
ObservationObs_FreeTrialQuote Free plan IS the trial — converting the sale right then.
ObservationObs_FrictionFailure Root cause of sales friction identified: lack of clarity on exactly what is being sold.
ObservationObs_FrictionFailure_1 Consumer SMS path introduces friction: SMS costs, abuse prevention, and onboarding without a consultant as identified failure modes.
ObservationObs_GIFHookFeature Tony's articulation of the GIF sales strategy: one hook GIF to grab attention, then feature GIFs. Hook first, features after.
ObservationObs_GIFLooksRealQuote Tony's articulation of why GIFs work as sales materials — they look real, not like a concept or engineering skunk works.
ObservationObs_GitNativeNotTagline I'm not buying it because it's git native. That may be how you make it happen for me, but that's not why I'm buying it.
ObservationObs_GitNativeReaction Git-native versioning resonated: 'the provenance of goals, the evolution of goals, is really key.' Validates technical architecture choice.
ObservationObs_GovernanceGraveyardQuote Scarlet's hard constraint on the word 'governance' as a product category.
ObservationObs_GraphPlacementDecision Explicit design decision: pragmatism over architectural purity; personal memory placed in org graph pending personal graph submodule; migration documented as intentional.
ObservationObs_GraphWalkerCompile Scarlet's graph-walker compile approach and a consultant who ran enterprise discovery engagements for fifteen years's big takeaway
ObservationObs_GreenButtonJIRA a consultant who ran enterprise discovery engagements for fifteen years's green-button/JIRA problem as illustration of context loss
ObservationObs_GTMFlexibility Founder rejects timing-window risk; infrastructure flexibility makes medium-term outlook strong.
ObservationObs_HaveAnOpinionPromiseQuote Scarlet's verbatim articulation of the customer-facing product promise — the missing marketing claim
ObservationObs_HowItWorksGap the application has the problem (clear), the benefit (clear), and the tech architecture (clear) — but is missing how the actual workflow within an organization works.
ObservationObs_IDParamRedundancyQuote Scarlet's verbatim request to rename the perspective parameter to id, identifying the redundancy created by the compile→perspective rename.
ObservationObs_InformationVsPerspective Scarlet's information vs. perspective distinction
ObservationObs_IngestDiscussion Ingest the discussion around artifacts, not just the raw content
ObservationObs_InsideJokeReference 'getting there' with a note saying 'inside joke reference from previous company.'
ObservationObs_InterviewMapsToOnboarding Interview question order intentionally mirrors CLAUDE.md onboarding mode priority: vision/mission → architecture/decisions → current priorities → team → recent decisions
ObservationObs_IntroEmailBestEverQuote a design collaborator re-anchors the original introductory email as the founding artifact of their working relationship, arguing it matters more now than it did then
ObservationObs_IRILeaves IRI-expansion-at-the-leaves: stated design goal, not yet built
ObservationObs_KeepingBeatsStartingQuote The conversion model flip: keeping beats starting as a close. The install-the-tool model ends by asking the client to start something; this model ends by asking them to keep something that already works.
ObservationObs_KeyConstraintPerspective Scarlet's key constraint that drove selection of 'perspective': the AI becomes the perspective, not just consults it.
ObservationObs_KeyPhrase_AttentionDilution Attention dilution vs. steering — core architectural differentiator phrase
ObservationObs_KeyPhrase_ContextEngineering Context engineering named as THE industry problem; used to reframe competitive landscape
ObservationObs_KeyPhrase_Hallucinating The hallucination question as the universal hook — replaces enterprise framing as the opening.
ObservationObs_KeyPhrase_SteeringVector Founding thesis phrase: narrative as steering vector for language models
ObservationObs_KeyPhrase_ThinkLikeYourOrg 'Think like your org' — the outcome promise of the worldview-building process. Positions Aswritten as an org-level intelligence layer.
ObservationObs_KeyPhrase_UnderneathTheText 'Underneath the text' is the core differentiator phrase — signals depth beyond document indexing without using technical jargon.
ObservationObs_KeyPhrase_WorldviewAsCode Worldview as code — the infrastructure layer's defining phrase
ObservationObs_ManualLensTestPR91 Manual lens test against PR #91 (Synthetic Representatives voice memo): compile worldview → introspect → draft centrist org model at three scales → re-draft through investor lens → delta table. Exposed the lossy compression problem.
ObservationObs_MarketContext_1 Two distinct go-to-market motions identified: enterprise consultant-led ([amount]–[amount]) and consumer/viral SMS (undefined revenue model). Currently unconnected in strategy and backlog.
ObservationObs_MarketplaceOfWorldviews A skill is limited context and a set of instructions about how to do something. We're talking about worldview — an entirety of process and methodology and code
ObservationObs_MeetingsThrowaway IRI-expansion-at-the-leaves as stated-but-unbuilt design goal for aswritten
ObservationObs_Mission Implied mission: extract undocumented knowledge, architect it to steer, make it infrastructure
ObservationObs_MoatLeverage Proprietary system, 7 org case studies, deep deployment knowledge cited as durable moat even in failure scenario.
ObservationObs_ModelAsOSMetaphor The supporting metaphor for the AGI reclaim: expecting the general-purpose model to handle every case is like expecting your OS to ship with all programs preinstalled. The task-adapted perspective is the program.
ObservationObs_ModulesCapabilities Three extraction moments: onboarding mode, decision-point extraction during work, consolidation and cross-referencing during review
ObservationObs_MonocultureFiveVoicesQuote Scarlet's articulation of the monoculture risk: five frontier-lab voices vs. the weird crackpot with a deli story.
ObservationObs_MoonshotPlainStatementQuote Scarlet's plainest statement of the moonshot: not an AI agent for everything, but as many distinct voices as there are humans.
ObservationObs_MoonshotSIIQuote The moonshot statement — the clearest articulation of the SII vision and the ownership thesis.
ObservationObs_MultiFormatApplicability Pitch simultaneously serves: elevator pitch, McDonald's picture memo/sales deck intro, and demo intro on Call One. Works at every format because every sentence is about prospect experience.
ObservationObs_MultiRepoCompilationQuote Scarlet's verbatim articulation of multi-repo compilation as the technical basis for layering.
ObservationObs_NamedPerspectiveQuote Scarlet's canonical articulation of the named perspective architecture — the ship metaphor and sub-agent framing.
ObservationObs_NarrativeArchLanded Tony's reaction to the fundraising materials — the Narrative Architecture white paper landed best, with no notes on any of the drafts.
ObservationObs_NarrativeOntologyExtraction Scarlet's tool extracts narratives from raw transcripts into an RDF knowledge graph using a narrative architecture ontology.
ObservationObs_NatureOfExpertiseQuote Scarlet's argument for permanent human irreplaceability: expertise is the selection of what to do and when, based on prior experience — exactly what humans provide and AI cannot substitute.
ObservationObs_NotificationsRemovalQuote Notifications removed because GitHub already handles review requests
ObservationObs_NumbersWorkB2C Tony's volume argument for B2C — the numbers work because every single person is now their own team.
ObservationObs_OneProtocolSpineQuote Defines the hybrid output shape: synthesis memories (cross-corpus recurrent themes + user framing) and landmark memories (specific sessions/docs annotated directly). Human reconciliation is the gate.
ObservationObs_OnPremHIPAAQuote Scarlet's framing of the on-prem upgrade path from Team to Professional via HIPAA compliance — the canonical requirements-driven upgrade example.
ObservationObs_OntologyAsSpeechActs Scarlet's verbatim description of her ontology modeling approach
ObservationObs_OntologyGeneralizationQuote Scarlet's articulation of the ontology generalization question — org-ready origins vs. expertise-writ-large scope.
ObservationObs_OpenQuestions Four open design questions: document registry schema, eye charts per ontology domain, refraction-annotate skill interaction, compile weight thresholds
ObservationObs_OptionsVsDecisionsQuote an engineering manager at the customer's verbatim articulation of the options-vs-decisions distinction — the sharpest customer-language version of the re-litigation problem aswritten has captured.
ObservationObs_OrgChangeManagement a founder beta user identifies organizational change management as the core friction in staffing AI adoption — not the technology itself.
ObservationObs_OrioleReaderGrounderPattern a collaborator's reader-grounder pattern: perspective → cite, repeatedly. Pulls captured context, grounds her work against it via citation. Zero authoring (yet).
ObservationObs_OutsideTechBetaUserChatWidget a beta user from outside tech's expressed preference for chat widget over SMS.
ObservationObs_OutsideTechBetaUserChatWidgetRequestQuote a beta user from outside tech's verbatim request for a chat widget — the direct trigger for the share-page widget architecture sketch.
ObservationObs_OutsideTechBetaUserGitHubOpacityFeedback a beta user from outside tech's March 4 feedback on GitHub opacity (~01:01:10): committing, pulling, and backend terminology were completely foreign. The above-the-line abstraction work is motivated in part by exactly this cognitive cost. S
ObservationObs_OutsideTechBetaUserSMSSpamCodedQuote a beta user from outside tech's verbatim insight on SMS trust tax in low-trust contexts — the founding evidence for the channel scoping decision.
ObservationObs_OutsideTechBetaUserSpamPushback a beta user from outside tech's pushback on SMS, paraphrased by Scarlet: spam association in lower-trust contexts.
ObservationObs_OutsideTechBetaUserToolLanguageFeedback a beta user from outside tech's March 4 prescription for clearer tool-language prompts after getting stuck post-connector-install. Scarlet named the pattern: 'Feel like this is actually one of the hardest parts about this thing. Someone's l
ObservationObs_OwnFromDayOneStructure Own-from-day-one structure: client pays for Team plan, adds Scarlet as team member. Consulting continues or stops independently of repo ownership. The consulting relationship and the product subscription are separate decisions.
ObservationObs_PerspectiveInversionQuote Tony's canonical statement of the development inversion — stop trying to get AI to see your way, start using your perspective as the development driver.
ObservationObs_PerspectiveOrderFourSections The four-section canonical ordering of the perspective response, as decided 2026-07-01.
ObservationObs_PerspectiveVsInfoQuote Scarlet's canonical statement of the perspective vs. information distinction — the core differentiator from search and RAG.
ObservationObs_PipelineSteps Six-step pipeline: memory commit → GitHub Action → SPARQL transaction → PR with comment → story draft regeneration → PR diff as telltale.
ObservationObs_PitchDirectionSolo The crystallized pitch direction for Trigger 1 (solo professional) — the line that bridges both triggers.
ObservationObs_PlainLanguageDifferentiator Plain-language differentiator retained for the 'how is this different from uploading docs to ChatGPT?' objection.
ObservationObs_PlanGatedQuote Plan-gated instructions from compile: different prompts per plan level, not separate tool sets.
ObservationObs_PlanningBottleneck a founder beta user validates the planning-as-bottleneck thesis from the Feb 5 discovery call.
ObservationObs_PositioningShift Pivot from enterprise/org-knowledge-retention framing (mechanism) to hallucination hook (pain). Scarlet's insight: current pitch describes the mechanism rather than the pain.
ObservationObs_PositioningThesis_1 Enterprise positioning: product-as-service, consultant-led knowledge extraction. Differentiator is Scarlet's interview and synthesis capability.
ObservationObs_Primitives_1 Five shared primitives identified as the atomic units for both GTM motions: branch management, reviewer assignment, interview-to-memory pipeline, conversational review/approval, notification on merge.
ObservationObs_PriorRaise Founder's prior fundraising experience (Populace, 2015) grounds the mechanics as known quantities.
ObservationObs_PrivacyConcernQuote Tony's privacy concern that drove avoidance — not wanting experimental or sensitive work visible to others in the system.
ObservationObs_ProcessorNotStore Scarlet's articulation of processor-not-store positioning and zero data retention
ObservationObs_ProductDeliverable Crystallized product definition: MCP server + 3 interviews = working collective memory system. [amount] price point. Three conversations with CEO, CTO, senior engineer.
ObservationObs_ProfessionalNamingQuote Scarlet's rationale for choosing 'Professional' over 'Enterprise' — the tier is defined by what you get, not who you are.
ObservationObs_PsychologyTodayNegativeExample Scarlet's verbatim critique of Psychology Today as the canonical negative example of property-based matching — the founding evidence for the matching-search reframe.
ObservationObs_RDFDatomicRegret a consultant who ran enterprise discovery engagements for fifteen years's verbatim expression of RDF-over-Datomic regret
ObservationObs_RDFMotherTongue a consultant who ran enterprise discovery engagements for fifteen years's rationale: RDF is the LLM's native tongue
ObservationObs_RecognitionPrecedesCreationQuote Scarlet's verbatim resolution: recognition through old association precedes creative reading.
ObservationObs_ReconciliationSetTransactionQuote Scarlet's definition of reconciliation as a transaction about a set
ObservationObs_RefractionPattern Two-stage PR pipeline: (1) refraction — agent introspects existing docs, surfaces coverage delta, re-annotates; (2) proposed edits — agent drafts actual changes as suggested edits against real documents
ObservationObs_RefractRegistryQuote Refract and registry removed from product; refract becomes a local skill
ObservationObs_RefractSkillQuote Refract as local skill: reads last TX, snapshot, docs folder, produces suggestions for update pass.
ObservationObs_RepricingHumanTime a consultant who ran enterprise discovery engagements for fifteen years's repricing-of-human-time thesis
ObservationObs_ReviewIsRememberQuote Scarlet's canonical statement that review corrections are just another memory — not a special flow.
ObservationObs_ReviewNamedGraphModel Reviews as RDF named graphs: each review carries reviewer identity, timestamp, lens used, feedback, and triples in scope. Enables accumulation, not state transition.
ObservationObs_ReviewToolQuote Review tool extended to three cases
ObservationObs_RichTierCompositionResult The real-graph result of applying steering-weight selection: positioning thesis, friction/failure region, narrative spine, strategy overview, and core workflows entered the rich tier.
ObservationObs_RiskGuardrails_1 Existing strategic guardrail: avoid consumer products and developer-first PLG. The SMS/topic product is identified as a third category that doesn't fit either exclusion cleanly.
ObservationObs_SalesHardPart Scarlet's verbatim statement on sales as the hardest part
ObservationObs_SalesMotionBreakthrough Sales call IS the product demo IS the delivery. Fireflies transcript routed to prospect repo automatically. Free trial embedded in sales process.
ObservationObs_ScarletAgentBoundsQuote Scarlet's articulation of agent self-knowledge and escalation — the extension of a consultant who ran enterprise discovery engagements for fifteen years's cold sub-agent test into bounds-awareness.
ObservationObs_ScarletAgentIncentivesQuote Scarlet's enumeration of agent incentives including accountability
ObservationObs_ScarletAppleAnalogy Scarlet's Apple analogy for product positioning from the a design collaborator call.
ObservationObs_ScarletBackPressureQuote Scarlet's articulation of the AI code review back-pressure problem and the PR volume it creates.
ObservationObs_ScarletBranchAtRootQuote Scarlet's decision to use branch-at-root over subdirectory scoping — the boring-and-works principle applied to temporal isolation.
ObservationObs_ScarletBranchPolicyQuote Scarlet's statement on branch defaulting to main with context-dependent guidance.
ObservationObs_ScarletBrandAIIdentityQuote Scarlet's articulation of brand-as-AI-identity
ObservationObs_ScarletCalmCompanyQuote Scarlet's verbatim articulation of the Calm Company / one-round-ever thesis, contrasting with the stale VC-stepping-stone framing in the fundraising brief.
ObservationObs_ScarletCantAutomateQuote Scarlet's verbatim framing of the human review requirement — the product-language version of the same argument the customer's CTO then made in engineering language.
ObservationObs_ScarletCitationEngineQuote Scarlet's verbatim correction establishing aswritten as citation engine not synthesis engine — the founding evidence for Narrative_CitationEngineNotSynthesis.
ObservationObs_ScarletCocaColaTargetStateQuote Scarlet's Coca-Cola example as the target state for brand-as-AI-identity
ObservationObs_ScarletCorrectionVerbatim Scarlet explains the extraction failure: text-for-topic was an exploration, not a product definition.
ObservationObs_ScarletDelegationQuote Scarlet's articulation of delegation as the core use case from the a design collaborator call.
ObservationObs_ScarletDemoLegibilityQuote Scarlet's candid articulation of the demo-legibility problem
ObservationObs_ScarletDemoProspectPathwayQuote Scarlet's summary of a demo-call prospect's two-step sales pathway covering both individual/SMB and org-tier targets.
ObservationObs_ScarletDemoVsProductQuote Scarlet's review note drawing the line between the share demo and the product goal.
ObservationObs_ScarletDenseProductLanguageQuote Scarlet on dense product language as intentional narrative architecture reinforcement
ObservationObs_ScarletDeveloperLikeMeQuote Scarlet's load-bearing ICP narrowing under pressure from David
ObservationObs_ScarletDifferentiatorDefinitionQuote Scarlet's in-memo answer to 'how is this different from another knowledge graph?' — stated deliberately at 1.76 w/s with a long pause before 'conversational events', word-searching on the definition but content arriving intact.
ObservationObs_ScarletEconomyPrincipleQuote Scarlet's verbatim statement of the economy principle
ObservationObs_ScarletEnrollmentQuote Scarlet's verbatim statement of the enrollment decision
ObservationObs_ScarletExecutionStrategyQuote Scarlet's articulation of the execution-to-strategy shift and how AI has restored IC status for everyone.
ObservationObs_ScarletExpertGitHubQuote Scarlet's articulation of Expert tier + GitHub behavior — main as default for single-user case.
ObservationObs_ScarletFeltVsShownQuote Scarlet's articulation of the felt-vs-shown crux — founding evidence for Claim_FeltVsShownOwnDataIsDeepener.
ObservationObs_ScarletFirstAppointmentQuote Scarlet's articulation of the category change vs. price comparison — the founding evidence for Narrative_FirstAppointmentNotAlternative.
ObservationObs_ScarletGoalDefinition Is part of what we're missing here is the agent + perspective having a goal?
ObservationObs_ScarletHallucinationReframeQuote Scarlet's hallucination reframe — the founding evidence for Narrative_HallucinationReframe.
ObservationObs_ScarletHumanDrivenQuote Scarlet's foundational statement of the human-driven development principle.
ObservationObs_ScarletIncredibleQuote Scarlet's reaction to the full annotated thesis — the product demonstrating itself through the process of producing the strategy document.
ObservationObs_ScarletInstallableExpertise Scarlet's reframe on the memory word, from the a design collaborator call.
ObservationObs_ScarletInterviewSalesMotionQuote Scarlet's articulation of the interview sales motion as a forcing function for human-driven development, with the free/paid split.
ObservationObs_ScarletLandingPageAnswerArchitecture Scarlet's verbatim answer architecture for the landing page: case studies that map to real pilots, showing current way, how aswritten does it, product behaviors, narrative architecture for each.
ObservationObs_ScarletLandingPageFocusQuote Scarlet's verbatim statement on picking consulting discovery as the clear landing page focus.
ObservationObs_ScarletLeanStartupQuote Scarlet's reframe of the Lean Startup principle for the AI era — humans-first, not just test-first.
ObservationObs_ScarletMetaLayerQuote Scarlet's reaction when shown the v1 map composition — the genesis of the complete area index requirement.
ObservationObs_ScarletMethodologyQuote Scarlet's verbatim correction establishing methodology density as the real requirement — the founding evidence for Narrative_MethodologyDensityOverContent.
ObservationObs_ScarletMoatQuote Scarlet's articulation of the technical moat from the a design collaborator call.
ObservationObs_ScarletModularSynthesizerQuote Scarlet's articulation of the emergent quality of the interview-driven strategy session — the synthesizer metaphor captures how the AI guides without prescribing.
ObservationObs_ScarletNarrativeArchitectureMethodQuote Scarlet's verbatim statement of the narrative architecture method
ObservationObs_ScarletNarrativeDefinitionQuote Scarlet's definition of a narrative as a statement true beyond the words it describes
ObservationObs_ScarletNarrowingTemptationQuote Scarlet's articulation of the narrowing temptation — the over-narrow pitch she resists.
ObservationObs_ScarletNavigationScalingQuote Scarlet's verbatim articulation of the scaling reason for navigation over full-graph-dump — the founding evidence for the attention-dilution problem at scale.
ObservationObs_ScarletNotRightTimeQuote Scarlet's explicit rejection of ontology and repo structure experiments mid-launch — the rationale for deferring narr:appliesAs.
ObservationObs_ScarletPerspectiveSteeringQuote1 Scarlet's first articulation of perspective-as-agent-steering — the mechanism by which accumulated conversation installs behavioral norms.
ObservationObs_ScarletPerspectiveSteeringQuote2 Scarlet's second articulation — the alternative model framing, steering through human and human-agent conversation.
ObservationObs_ScarletPhoneNumberFeels Scarlet on the genuine trust-signal value of the phone number format.
ObservationObs_ScarletPrimaryInputsQuote Scarlet's verbatim articulation of her two primary inputs — the evidence that the default usage pattern is not being communicated.
ObservationObs_ScarletProfitableByDefaultQuote Scarlet's verbatim statement that the business is built on a model that expects repeatability and profitability early — the founding evidence for the 'profitable by default' reframe.
ObservationObs_ScarletProposalQuote Scarlet's articulation of the proposal as meta-contract — the founding evidence for the proposal-artifact framing.
ObservationObs_ScarletPurchaseTriggerQuote Scarlet's articulation of the core finding: understanding ≠ sale.
ObservationObs_ScarletQuote_Constellation Scarlet articulates the constellation frame: richer citations from the interaction between many claims and narratives.
ObservationObs_ScarletQuote_PrimaryUnique Scarlet's refinement: primary must be unique within a single return, but constellations can have intersecting related sets.
ObservationObs_ScarletQuote_RealDifference Scarlet's refined distinction: the how that belongs in a memory is directional and relational (we did it this way, I asked this person, here is the why). Specifics of code, color values, layout are implementation details.
ObservationObs_ScarletQuote_Reductionist Scarlet's diagnosis of the single-claim reductionist frame in the cite prompt.
ObservationObs_ScarletQuote_SameClaim Scarlet's direct diagnosis of the deduplication problem.
ObservationObs_ScarletQuote_TensionHere Scarlet's initial framing: the perspective is the source of truth for reasoning, not for values. Specific values in the perspective risk confusing canonical sources as they drift.
ObservationObs_ScarletQuote_TextHierarchy Scarlet's positive example of what IS appropriate: the nature of a text hierarchy, ratio between elements, emotional impact, feeling — these are perspective. Specific pixel values or direct multiples are not.
ObservationObs_ScarletRAGCritiqueQuote Scarlet's critique of the status quo — RAG-based architecture documents miss the conversational why. The founding evidence for Claim_ConversationalSteeringAlternativeToRAG.
ObservationObs_ScarletReferralQuote Scarlet's articulation of the gift-economy referral mechanic — the founding evidence for the referrer-stays-in-gift-economy claim.
ObservationObs_ScarletRejectsAIPhrasing Scarlet's verbatim rejection of the coined compression 'speaks from the record' — the founding evidence for the conversational phrasing rule.
ObservationObs_ScarletReturnValueShiftQuote Scarlet's canonical statement of the product-wide shift toward prompting in return values.
ObservationObs_ScarletRevenueModelQuote Scarlet's articulation of the dual revenue model — upfront fee and/or percentage — and the viability of both.
ObservationObs_ScarletReviewKickoffQuote Scarlet's verbatim articulation of the desired review kickoff behavior after remember returns.
ObservationObs_ScarletScrollAnimationQuote The generative moment of the memo — the scroll-driven animation idea arriving fast and fluent at 4.7 w/s, the idea outrunning the sentence.
ObservationObs_ScarletSetupFeeQuote Scarlet's verbatim articulation of the setup fee as conversion service for single/small team publishers.
ObservationObs_ScarletShowDontTellQuote Scarlet's core reframe on the landing page failure mode — stated as a settled conclusion at 2.0 w/s, energy below baseline, conf 0.96.
ObservationObs_ScarletSiloingFeatureNotBugQuote Scarlet conceding David's siloing point as a real design boundary, not a misunderstanding
ObservationObs_ScarletSMSBonus Scarlet frames SMS trust-signal as bonus/upsell, not core product identity.
ObservationObs_ScarletSMSEnterpriseLeaning Scarlet's leaning toward SMS as enterprise-only, stated as preference not decision.
ObservationObs_ScarletSMSFraudCost Scarlet on SMS operational cost: fraud maintenance and verification overhead.
ObservationObs_ScarletSMSNotPhoneNumber Scarlet's corrected positive position on SMS as channel, verbatim.
ObservationObs_ScarletStaircaseQuote Scarlet's verbatim articulation of what the staircase actually is — the founding evidence for Narrative_ProfitabilityAsMathCheck.
ObservationObs_ScarletStoreDecisionProcessQuote Scarlet's articulation of storing the decision process vs. the half-baked artifact
ObservationObs_ScarletTargetMarketQuote Scarlet's articulation of the target market from the a design collaborator call.
ObservationObs_ScarletTeamWithoutGitHubQuote Scarlet's articulation of Team without GitHub — users review their own memories against shared perspective.
ObservationObs_ScarletTemporalityQuote Scarlet's verbatim articulation of the parallel temporality observation — the founding evidence for Claim_ScarletTemporalityParallelObservation.
ObservationObs_ScarletTextForTopicQuote Scarlet's verbatim articulation of the text-for-topic precedent and the canon-too-soon drift mechanism — the founding rationale for the research branch methodology.
ObservationObs_ScarletThreeUseCasesVerbatim Scarlet's verbatim taxonomy of the three enterprise use cases, distinguishing Discovery, Onboarding, and Governance/Organizational Memory.
ObservationObs_ScarletTierUncertain Scarlet's actual epistemic state on tier mapping: deliberately undecided.
ObservationObs_ScarletTopicAnalysisQuote Scarlet's verbatim articulation of the topic analysis approach as the preferred alternative to direct ingestion of historical transcripts.
ObservationObs_ScarletUnexecutedUseCaseQuote Scarlet's status note: the perspective-as-deck-alternative has not been executed with a consulting client
ObservationObs_ScarletUnsureArchetypes Scarlet's verbatim articulation of uncertainty about continuing to push other user archetypes vs. focusing on enterprise-shape.
ObservationObs_ScarletVerdictQuote Scarlet's verdict on June 10 — the highest-conviction pitch validation statement in the perspective.
ObservationObs_ScarletVettingQuote Scarlet's articulation of why patient-pays-per-vet is genuinely additive — the founding evidence for the patient-as-goal-holder claim.
ObservationObs_ScarletVoiceMemoQuote Scarlet's verbatim articulation of the ongoing iteration step — voice memo as gap-filling mechanism after each ARB.
ObservationObs_ScarletVouchPermissionsQuote Scarlet connecting the agent/human boundary to Vouch's granular permissions work — the founding evidence for Narrative_PerTaskScopedPermissions.
ObservationObs_ScarletVouchPrecursorQuote Scarlet recognising a design collaborator's process as the Vouch text process — the product's precursor
ObservationObs_ScarletWebNotMappingQuote Scarlet's correction: the architecture is a web, not a 1:1 mapping
ObservationObs_ScarletWhiteSpaceQuote Scarlet's coinage of 'white space' as the mechanism for how the perspective fills in model defaults
ObservationObs_Segmentation_1 Enterprise buyer persona: CEO/CTO of Seed/Series A technical companies. Consumer buyer: anyone with a phone.
ObservationObs_Segmentation_ICP ICP validated by an infrastructure advisor: 10-50 person dev shops at [amount]-50M revenue; sophisticated enough to already have context solutions
ObservationObs_SeparationPrincipleQuote Core insight: conflating org automation with product features
ObservationObs_SeparationSolutionQuote Tony's prescribed solution for technical user adoption — start completely fresh, keep aswritten separate from existing repos.
ObservationObs_SMBOpenQuestions Three open questions for SMB market: (1) HIPAA / zero-data-residency on-prem deployment option for healthcare SMB; (2) How to abstract GitHub from signup (high priority — SMB users won't have GitHub accounts); (3) Custom ontology for non-te
ObservationObs_SMBProductImplications Product implications of SMB pivot: modules for grantwriting, decks, LinkedIn posts, graph coverage goals; target audiences beyond tech (arts orgs, cultural orgs, early entrepreneurs, artists); business plan and investor materials should lea
ObservationObs_SoakTestDefinition Quality definition: living-with-it testing over days/weeks, not passing unit tests.
ObservationObs_SPARQLUpdateChoice a consultant who ran enterprise discovery engagements for fifteen years's unprompted convergence on SPARQL Update as log primitive
ObservationObs_StaleGraphDanger The stale graph + citations danger: worse than generic AI with no citations. Key risk for org memory use case.
ObservationObs_StatelessDedupDesign Scarlet chose stateless dedup over server-side session tracking to keep the read path a pure function (ZDR-pure, no session store, composable). The reliability comes from the return carrying the manifest forward, not from trusting the agent
ObservationObs_SteeringGenusQuote Scarlet's canonical articulation of the genus/species distinction between steering and governance.
ObservationObs_StickyProductQuote Tony's assessment of the product's stickiness and the word-of-mouth potential once it reaches 20-50 users.
ObservationObs_StoriesRemovalQuote Clear decision to remove Stories entirely
ObservationObs_StructureContextMemory Tony's observation that 90% of what people are chasing is structure for context and memory — the core JTBD for aswritten's B2C market.
ObservationObs_StyleVoiceEncoding The ontology encodes style, voice, diction, key phrases, and context around those phrases.
ObservationObs_SynthesizerNamingRationale The naming rationale for SII over AGI: synthesizer connotes human-guided music-making; artificial connotes lab food flavoring. The word choice encodes the relational, human-guided nature of SII.
ObservationObs_SyntheticIdentityDefinitionQuote Scarlet's definition of synthetic identity / SII from the walk voice-memo — the most complete technical definition in the session.
ObservationObs_SyntheticIdentityForkFull Full verbatim of the ownership fork — the whole SII positioning in miniature. Individually owned and carried vs. patchwork corporate networks forever asking to prove access.
ObservationObs_TaglineGap 'AI native teams' is WHO, not WHAT or WHY. Missing the 'oh, I want that' factor.
ObservationObs_TaglineIterationProcess Scarlet iterated through six tagline variations before landing on the two-beat structure. The final form encodes the product's dual promise: human provenance (trust) and AI citation (proof).
ObservationObs_TeamTierOnboardingQuote Scarlet's canonical statement that the Team tier sales call is simultaneously the onboarding and the first knowledge extraction — no separate setup deliverable.
ObservationObs_TechnicalChaosQuote Tony's articulation of the technical user's adoption barrier — 10 repos on the go, worried about what breaks or crosses over.
ObservationObs_ThemeIsInterpretiveQuote Scarlet's verbatim articulation of the theme-is-interpretive trap and the anti-reductionist resolution.
ObservationObs_TimeToValue Three interviews to working system. Explicitly contrasted with 'six-month implementation'.
ObservationObs_TimingThesis Raise framing: raw innovation with traction → specific product with sales motion → angel/seed from case studies and unit economics. Summer 2026 as inflection point.
ObservationObs_TinkererBarrierCorrection Scarlet's correction to the AI's hypothesis about the Tinkerer archetype — their existing solution is a barrier, not an accelerant to adoption.
ObservationObs_TonyAITrapQuote Tony's diagnosis of the AI action trap mechanism — every response prompts another action, answer is always yes.
ObservationObs_TonyExperienceAsAsset Tony's validation that experience and strategy are the real assets in the AI era, not execution capacity.
ObservationObs_TonyPainFirstQuote Tony's pain-first words from January 8 — the context-server pain in the buyer's own voice, already in the perspective.
ObservationObs_TonyPipitAdoptionBarrier Tony's self-diagnosis of why he hasn't adopted aswritten via a collaborator — not being all-in on a collaborator means the product can't get full traction.
ObservationObs_TonyValidationAdvisoryQuote Tony's validation that senior people would not say no to the advisory model — the access unlock argument.
ObservationObs_ToolNameConversationalWord Scarlet demonstrating how 'perspective' becomes a natural conversational word, and 'aswritten' lands as provenance marker.
ObservationObs_TransactionSourceOfTruth a founder beta user's key moment of understanding: the transaction/memory graph is the durable source of truth; docs are transient.
ObservationObs_TranscriptionToolsSurvey Six-tool survey (Fireflies, Otter, Google Meet/Gemini, Microsoft Teams, Zoom, Krisp) verified against vendor developer docs. Key finding: native API grain already matches aswritten's philosophy — nearly every tool fires a notification and l
ObservationObs_TransformVerbQuote Scarlet's canonical statement of the transform-over-query distinction.
ObservationObs_TrustedInformationQuote Scarlet's canonical market positioning statement — the future of trusted information in the AI age.
ObservationObs_TSABodyFlagQuote The TSA scanner as the concrete instantiation of the AI individuality problem — a centralized norm flagging the divergent body.
ObservationObs_TwoCallStructure Pitch is spoken on Call One before the read-path demo. Enterprise framing (senior expert bottleneck, knowledge retention, compliance) becomes Call Two language.
ObservationObs_TylerCowenScreeningQuote Tyler Cowen's screening principle cited as structural parallel: boring applications fail across hiring, funding, and buying
ObservationObs_UndocumentedKnowledgeQuote Scarlet's canonical statement of the undocumented knowledge thesis — why distillation and remember matter.
ObservationObs_UnifierQuote The unifier statement — each use case is a steering loop plus a memory substrate.
ObservationObs_UpgradePathQuote Scarlet's framing of the two upgrade path types — value-driven and requirements-driven — as the architectural principle baked into the tier design.
ObservationObs_UseCase1Quote Scarlet's verbatim articulation of Use Case 1 — governing a human team.
ObservationObs_UseCase2AgenticScalesQuote Scarlet's verbatim articulation of Use Case 2 — agentic steering at two nested scales.
ObservationObs_UseCase3Quote Scarlet's verbatim articulation of Use Case 3 — steering the shared process across engagements.
ObservationObs_UserAgentAssumptionQuote Foundational assumption enabling the protocol model: rich-content onboarding users can be assumed to have Claude Code or equivalent agentic environment.
ObservationObs_UserOwnedRepo Scarlet's product stores all data in the user's own repository — no backend storage — positioning data sovereignty as a trust differentiator.
ObservationObs_V1UXVision Scarlet's articulated V1 UX: MCP endpoint + notifications; GitHub is infrastructure, not interface.
ObservationObs_VisionManifestoPrescriptionQuote a design collaborator's prescription for the Scarlet Vision manifesto
ObservationObs_VocabularyConsolidation Vocabulary consolidation: telltales retired (mixed metaphor); lenses, eye chart, refraction check, optometrist model established as canonical terms
ObservationObs_VulnerabilityTenorQuote Scarlet's characterization of the tenor of the Vouch emails vs. her current aswritten outreach — the structural difference she identified as the template.
ObservationObs_WhatItsNot_ZeroContext Explicit retirement of zero-context framing; defines what aswritten is NOT positioning as
ObservationObs_WhitePaperAIFilter Tony's articulation of the white paper AI filter problem — executives will summarize with AI rather than read.
ObservationObs_WhiteSpaceCollapse White-space-collapse metaphor for organizational perspective steering LLM output
ObservationObs_WhoCaresIfAIEmailQuote Scarlet's restatement of the product thesis through the amplification lens: the question is not whether AI can produce an email but whether it produces a meaningful, unique one.
ObservationObs_WorkflowTaxonomy Six new workflows created across four architectural roles: Flows (orchestration with side effects at edges), LLM Pure Functions (no side effects), Comms (external communication), Entrypoints (webhook normalization).
ObservationObs_WorldviewEncodingCriticismsQuote Scarlet's claim that worldview encoding addresses the common AI behavior criticisms — tone-deaf responses, collaborative flatness.
ObservationObs_WorldviewEngine Scarlet's core product framing: the worldview engine / narrative engine. SMS is something you can do with it.
ObservationObservation_1 Narrative steering as worldview bootloading
ObservationObservation_4 The steering IS the product
ObservationObservation_ActiveSentencesQuote Scarlet's direct quote on active sentences and no X-not-Y
ObservationObservation_ActiveVoiceQuote Scarlet's direct quote on active voice
ObservationObservation_AdditiveQuote This is something she does not have or cannot do is additive to the company or her personally.
ObservationObservation_AITricksQuote Scarlet's direct quote on AI-language tricks
ObservationObservation_ApproachPattern_1 Conversational AI Interview for Knowledge Extraction
ObservationObservation_ApproachPattern_PilotNetworkEffect Canonical flow: Start with one team (5-10 people), add second role mid-pilot to demonstrate network effects, expand org-wide after proof.
ObservationObservation_ArchetypeTitle_AdvisorLedPilot Solution archetype: Advisor-led warm intro → CEO pitch → 30-day pilot (Strategy or CTO team) → org-wide expansion.
ObservationObservation_AswrittenCapQuote Scarlet's direct quote on Aswritten capitalization rule
ObservationObservation_AsWrittenProduct As Written: git-native organizational memory tool that compiles a worldview graph and exposes MCP tools for content introspection and generation.
ObservationObservation_AugmentFirst Augmentation principle: annotate first, complete the request, present annotation results alongside — provenance awareness layered onto every agent action.
ObservationObservation_Beat1_ColdDelegation Beat 1: Cold delegation. Ask the AI to do real work that would normally require an hour of briefing. The prospect evaluates output quality directly. The AI did not need briefing because it already has the worldview.
ObservationObservation_Beat2_CitationLayer Beat 2: Citation layer (annotate). Shown as a quality attribute of the work product, not a separate feature. Claims trace to specific conversations and decisions, not hallucination.
ObservationObservation_Beat3_GapHonesty Beat 3: Gap honesty (introspect). Ask the AI something it does not know. It tells you what is missing and who to ask. Demonstrates epistemic integrity.
ObservationObservation_BetaCallPattern Five-step beta call pattern: (1) Interview them — call IS the first memory. (2) Walk through onboarding mode together. (3) Upload documents. (4) Do ingestion and narrative strategy. (5) Solicit feedback in use.
ObservationObservation_BreadBakingGifter my friend just asked me about bread baking, call me and we'll talk it through and then text them and now they have my perspective
ObservationObservation_CascadingEffects Strategic changes in As Written should cascade automatically to all downstream AI-generated content, analogous to code propagation.
ObservationObservation_CaseResults_1 Complementary vs. Competitive Positioning
ObservationObservation_ClientMindset Traditional clients often want a 'tool to fix all problems'; consultants must redirect toward understanding AI workflow integration.
ObservationObservation_CodeInferenceMoment The system inferred a code change from past discussions even without a specific captured claim.
ObservationObservation_CompetitiveLandscape Named competitors explicitly positioned as org-only, establishing the personal-to-org compile layer as a novel competitive moat
ObservationObservation_CompetitiveLandscape_1 OpenAI Frontier as Agent Orchestration Platform
ObservationObservation_CustomerEngManagerBridgeRole an engineering manager at the customer acts as the bridge between ARB decisions and implementation teams, using aswritten to maintain coherence.
ObservationObservation_CutCruftQuote Scarlet's direct quote on cutting cruft
ObservationObservation_ElephantCollective Bootstrapped from [amount]to [amount]/event in 6 months. ~20 artists-in-residence, 1,000–2,000 person events, 100–200 volunteer crew recruited via mass SMS.
ObservationObservation_EmDashQuote Scarlet's direct quote on em-dash prohibition
ObservationObservation_EnterpriseCustomerDemoReframe Reframes the the enterprise customer demo from 'goal-directed agent primitive' to 'handing you back peace'.
ObservationObservation_ExamplesContextQuote Scarlet's direct quote on examples needing context
ObservationObservation_ExtractionExplainer Scarlet describes the 'as written' extraction process: breaking down conversational memories into a graph of concepts, observations, and narratives.
ObservationObservation_FlockingInsight The incentives around the behavior of what seem like small actors within larger systems drive emergent behavior in larger networks in a way that creates incredibly complex, powerful and beautiful interactions.
ObservationObservation_FocusSequencing_PilotExpand Pilot with Strategy or CTO team (5-10 people) for 30 days; add second role mid-pilot to demonstrate network effects; expand org-wide within 60 days.
ObservationObservation_FutureLensTool Future: formalized aswritten/lens tool taking focus, perspective_weights, and target; returns re-weighted compilation. Lens configurations become saveable and reusable named configurations. Prompt-based version to be validated first.
ObservationObservation_GitHubBranchTrunk Scarlet explains GitHub's 'branch and trunk' model to a beta user from outside tech as a conceptual analogy for version-controlled knowledge management.
ObservationObservation_HandWavedQuote Scarlet's direct quotes on hand-waved closings
ObservationObservation_HumanReadableIDs the id should be human readable and generated from the tx set.
ObservationObservation_IncentivesConstraints_DecisionSpeed Incentives: Seed/Series A CEOs can make purchase decision quickly (no lengthy procurement), can mandate org-wide adoption, have budget but care about ROI.
ObservationObservation_InflectionPoints_1 Enterprise Budget Reallocation to AI-Native Platforms
ObservationObservation_InvertedComplement The document explicitly frames introspect and annotate as inverted complements operating on the same graph from opposite directions — a key architectural design tradeoff.
ObservationObservation_JohnDyessExample Recurring conflation where devs mistake 'options' for 'decisions' leading to re-litigation.
ObservationObservation_LandingThePlaneQuote Scarlet's direct quote rejecting the 'X, not Y' landing-the-plane pattern
ObservationObservation_LeadingIndicators_AdvisorResponse Leading indicators: Advisor responses (10-15 outreach → 5-8 positive), CEO intro calls (5-10 scheduled), pilot commitments (2-3 companies).
ObservationObservation_LensDraftingProcess 10-step Lens Drafting process: (1) compile worldview, (2) introspect for competing perspectives, (3) draft centrist version, (4) identify 2-3 lens shifts, (5) draft full section through each lens, (6) show delta table, (7) present and ask w
ObservationObservation_LLMContinuation LLMs are continuation machines, not thinking machines. They follow narrative vectors embedded in training data.
ObservationObservation_MCPStack As Written uses MCP (Model Context Protocol) as the integration layer for organizational memory tools.
ObservationObservation_MoatLeverage_IPProtection Compounding advantage: git-native worldview protects IP while enabling AI; network effects multiply value across roles.
ObservationObservation_NorthStarMetrics_TimeToExecution North-star metric: Time from strategic decision to implementation (weeks/months → days); correlates with durable value creation.
ObservationObservation_NotificationPersuasion Scarlet's elaboration collapsing notification and persuasion via personal graph
ObservationObservation_OpenQuestions Four open design questions: privacy/consent model, UX for multi-repo, join timing (compile vs query vs notification time), competitive position
ObservationObservation_OutcomesProof_MetricsDelivered Expected metrics: Time from idea to updated materials (days→15 min), time from vision shift to implementation (weeks/months→days), consensus building (ongoing chase→built into PR review).
ObservationObservation_ParkMemoryMechanism The creators of this paper, they suggested a mechanism based on three different parameters. Recency, importance, and relevance to the current context.
ObservationObservation_PeaceAsBenefit The word 'peace' is the felt benefit that experts will pay almost anything for.
ObservationObservation_Pipeline Five-step pipeline: Memory→extraction→provisional triples; TX commits to branch→PR opens (no story gen); Reviewers notified→start_review in Claude; Review IS optometrist: lenses exercise triples, iterate to convergence; Merge→reviewed tripl
ObservationObservation_PositioningThesis_SeedSeriesA Primary target: Seed/Series A CEOs via advisor network; pain is acute, decision authority + budget, network effects kick in fastest.
ObservationObservation_PositiveClaimsQuote Scarlet's direct quote on positive direct claims
ObservationObservation_ProblemContext_1 Knowledge Locked in Expert Heads
ObservationObservation_ProblemContext_IPLeakageAlignment Triggering conditions: Heavy AI adoption + IP leakage risk + strategy-execution misalignment + small team velocity needs.
ObservationObservation_RememberAsConcept Remember is introduced on call one as a concept, not a live feature. The framing bridges the demo to the next call.
ObservationObservation_ReportAesthetic cream paper background, Source Serif headlines, Source Sans labels, the brand palette accents per tool
ObservationObservation_ScarletLensDefinition Scarlet's canonical definition of the lens tool: competing priorities + alternative perspective weights in the graph → different drafts → delta. Humans read drafts but need help identifying where shifts occurred.
ObservationObservation_SceneGrammarQuote Scarlet's direct quote on scene grammar specificity
ObservationObservation_Segmentation_HighIPTechnical Target verticals: AI/ML companies, developer tools/infrastructure, SaaS with proprietary methodology, infrastructure/platform companies—where IP is the moat.
ObservationObservation_SeniorExpertBottleneck The senior expert bottleneck as a special case of the personal graph value proposition
ObservationObservation_Signals_1 AI Coworker Framing Dominance
ObservationObservation_StaticVsLiveKB an engineering manager at the customer contrasts aswritten's live information against static-data knowledge bases.
ObservationObservation_StructuralIndex Graph shape without content: entity counts, relationship types, no actual triple content. Generated by GitHub Action alongside extraction when new memories are committed.
ObservationObservation_Task240Phase1Real Audit table writes are anonymized at the proxy boundary
ObservationObservation_TaxonomyRevision Revised taxonomy: four org-level deployment patterns (Solo, Parallel, Hub, Teams) plus one per-memory capability (Salon). Full cohort coverage achieved.
ObservationObservation_TimingThesis_1 AI Context Fragmentation Crisis Validated
ObservationObservation_TimingThesis_AIAdoption Converging forces: heavy AI adoption creates acute IP leakage pain NOW; companies investing in AI are systematically transferring competitive moat.
ObservationObservation_ToolHierarchy Tool hierarchy: Primitives vs Composers
ObservationObservation_TrustContractPrincipledEscalation Principled escalation (knowing limits) is the trust contract that makes delegation real.
ObservationObservation_TwoCallStructure Two-call structure: Call 1 = read-path demo on own org + consultative interview. Between calls = write path offline. Call 2 = read-path demo on prospect's org.
ObservationObservation_TwoPassPattern an engineering manager at the customer articulates the Phase 1 (exploratory) and Phase 2 (finalization) pattern of ARB sessions.
ObservationObservation_TwoQuestions The two questions that currently collapse into one — the core product gap
ObservationObservation_TXScaleFlow Four-step scale management for large TXs: summary, zone ID, targeted shifts, fact review.
ObservationObservation_ValueExchange_1 Senior Expert Bottleneck Value
ObservationObservation_ValueExchange_MetricsROI Value exchange: CEO gives budget + org mandate; receives quantified IP protection, time from pivot to execution (weeks→days), consensus building without overhead.
ObservationObservation_VerificationCycle Three-step verification cycle: Introspect identifies undocumented knowledge, Remember fills gaps, Annotate confirms text reflects what was documented.
ObservationObservation_VocabDefinitions Controlled vocabulary: Lenses, Optometrist model, Story drafts, Telltales — four terms redefined or retired
ObservationObservation_VouchSSoT Created 'narrative source of truth' manually at Vouch — seeding AI conversations with organizational context and sanctioned narrative statements as steering vectors.
ObservationObservation_WhatDoesItDo_1 Upstream Knowledge Extraction Value Proposition
ObservationObservation_WorkedExampleAfter After version of Architecture beat: belief, specifics, belief shape. Contrast emerges from the walk through what we believe and what actually crosses the wire.
ObservationObservation_WorkedExampleBefore Before version of Architecture beat showing multiple AI-tells: 'from day one', 'not a configuration', 'not the conversation you had'
ObservationObservation_WorldviewTool Same as layer0 compile. Code-based, free, always called first. Identity and routing. No change.
ObservationObservation_WritePathOffline Between calls: Scarlet ingests the call transcript into the prospect's repo, runs extraction, sets up examples based on their actual workflow. Bugs are private. No time pressure.
ObservationOpportunity_ADR002 Current AI tools are context-starved across coding, executive, and content domains. Each domain uses separate tools with no shared context, leading to strategy-execution gaps.
ObservationOutcome_EliminateStrategyExecutionGap Eliminates the [amount] mistake in startups where engineering builds features customers don't want by ensuring same collective memory snapshot backs product, engineering, and marketing.
ObservationPersona_SoloDeveloper Solo developers / small teams using Claude Code, Cursor, Windsurf. Pain: AI agents forget context between sessions, docs go stale.
ObservationPositioningThesis_ADR002 Worldview layer for AI agents (not semantic memory API). Shared and collaborative worldview backs AI agents across coding, executive, and content use cases.
ObservationTagline_ADR002 Worldview layer for AI agents
ObservationTimingThesis_ADR002 Strategy-execution gaps caused by separate tools with no shared context create urgency for unified worldview layer.
ObservationTool_ClawdBot Clawd bot (lobster claw spelling): connects to 20+ LLMs, hooks into WhatsApp/Telegram/Slack, gives bot access to Gmail/filesystem; research for competitive analysis
ObservationTool_HappyCoder Happy coder (capi coder): wrapper for Claude code, enables iPad/iPhone access to Claude instance on laptop; research for mobile UX patterns
ObservationTool_MCPInstances Various MCP instances and skills/plugins doing shared memory 'not necessarily as complicated... or maybe as robust'; an infrastructure advisor offered to send details
ObservationUseCase_CodingAgents AI-Assisted Development
ObservationUseCase_ContentGeneration Speeches, presentations, blog posts generated from organizational worldview. Brand voice consistency (narrative architecture encodes style/perspective). Automatic updates when worldview changes (strategies compile into stories).
ObservationUseCase_ExecutiveAssistance Strategy discussions become queryable memories. AI agents maintain organizational worldview (goals, values, market position). Cross-team alignment (sales, product, engineering work from same context). Meeting notes and decisions persist as
ObservationWhatIsIt_ADR002 collective memory provides a shared and collaborative worldview that backs AI agents across three primary use cases: AI-Assisted Development, Executive Assistance, Content Generation.
Narrative (429)
NarrativeaboutNode AI-native teams lack collaborative memory and source of truth
NarrativeInsight_PremiseMismatch Enterprise AI adoption reality
NarrativeNarrative_1 Flocking Origin (2012)
NarrativeNarrative_10xGapIsFeature The 10x Price Gap Between Team and Professional Is Intentional
NarrativeNarrative_2 Core Technology Insight
NarrativeNarrative_3 Narrative Architecture
NarrativeNarrative_4 Version, collaborate, manipulate, review, deploy perspective—treating narrative like software.
NarrativeNarrative_5 Feather arrows so first attempt goes further in the right direction; adjust heading, don't chase missed shots.
NarrativeNarrative_6 Layer Design Principle
NarrativeNarrative_ActivationGap The Activation Gap: Investment Model vs. Tool Model
NarrativeNarrative_ActivationGapClosed The agent deployment layer closes the activation gap by providing a concrete application for the technology.
NarrativeNarrative_AdaptiveToolsUX Single MCP endpoint with context-driven adaptive tools eliminates per-repo server config UX friction
NarrativeNarrative_AdditiveVsIncremental Additive vs. Incremental Value
NarrativeNarrative_AdvisoryStaffingAlternativeToHiring Advisory Staffing as Structural Alternative to Hiring
NarrativeNarrative_AgentAsAnalyticsInterface Agent-as-Interface Is the Right Shape for Analytics
NarrativeNarrative_AgentBoundsAndEscalation Agent self-knowledge: defining the bounds of task, intent, and world limits so an agent knows when it cannot finish and must escalate
NarrativeNarrative_AgentComputerInterface Agent Computer Interface: the agent-facing analog of human-computer interface
NarrativeNarrative_AgenticCoherence Coherence across agentic outputs—preventing conflicting information across documents and agents—is the next critical bottleneck in AI workflows
NarrativeNarrative_AgentificationHomogenization AI systems converge on a single register as they get agentified — M-dash, 'load-bearing,' uniform voice — making monoculture the default trajectory
NarrativeNarrative_AI_Steering Narrative as steering mechanism for AI systems
NarrativeNarrative_AIActionTrap The AI Action Trap: Every Response Prompts Another Action, Answer Is Always Yes
NarrativeNarrative_AIAsNewHireBridgingInsight AI as New Hire: The Bridging Insight Across Both Triggers
NarrativeNarrative_AIBackPressureCodeReview AI Code Back-Pressure: Volume Outpaces Human Review Capacity
NarrativeNarrative_AINativeOrgSize AI-native companies need only 2–6 senior core people who control AI tooling. Roles: Architect, PM (technical), Senior DevOps, Product-literate coder/strategist, Sales/partnerships, Designer.
NarrativeNarrative_AIofPractitioner The product is an AI of the practitioner that produces proposals
NarrativeNarrative_AItoAIProvenance AI-to-AI provenance as moat
NarrativeNarrative_AltitudeSwitchingCognitiveCost Switching altitude (building vs. vision-pitch) is more cognitively expensive than switching projects — explains why a year in the mechanics left the vision pitch undeveloped
NarrativeNarrative_AntiHallucinationViaCitation Anti-hallucination enforced by requiring LLM to cite every claim with provenance
NarrativeNarrative_AppendOnlyConvictionLog Append-only log enables conviction-over-time queries across claims and actors
NarrativeNarrative_April1SoftLaunch April 1 Soft Launch: Beta Done, Public Marketing Begins
NarrativeNarrative_ArchaeologyQueryUseCase Archaeology query: show how it was originally done and all the conversations that led to this moment
NarrativeNarrative_ArchitectureGap Dumping context into flat RAG creates attention dilution; narrative ontology with conviction levels makes knowledge steer rather than dilute
NarrativeNarrative_ArchitectureTeamsBeachhead Architecture Teams as Beachhead
NarrativeNarrative_ARRepoIdentity Every aswritten repo IS a perspective — there is no special category of perspective repo
NarrativeNarrative_ArtifactsLoseTheWhy Artifacts lose the reasoning behind their creation
NarrativeNarrative_Aspen Language as code: a collaborator analytics platform
NarrativeNarrative_aswritten aswritten.ai: automating the manual process
NarrativeNarrative_AswrittenIsMemorySubstrate aswritten Is the Memory the Steering Runs On — Never the Loop
NarrativeNarrative_AsyncSharedMemory Shared persistent memory solves the broken synchronization problem: decisions currently fragment across Slack, email, Teams, recordings. Shared memory enables async work with real progress and no synchronous overhead.
NarrativeNarrative_B2CPivot B2C as Primary Market: Every Individual Is Their Own Team
NarrativeNarrative_BaseCampNotCamp4 Base Camp, Not Camp 4: Real Inflection but Long Climb
NarrativeNarrative_BetaAsConsulting Full-service consulting engagement as the mechanism for beta product validation
NarrativeNarrative_BetaFeedbackLoopIsSalon The product tests itself: beta feedback loop is itself a Salon
NarrativeNarrative_BoundedDomainFidelityArgument Bounded-Domain Fidelity: Expertise Modules Are Higher Fidelity Than Org Memory
NarrativeNarrative_BrandAsAIIdentity AI identities have the opportunity to be the brand's identity — an AI shaped to think, design, and talk like the brand, deployable across the org
NarrativeNarrative_BrandMaturitySignal Restrained color palettes and sedate tones signal maturity and credibility early in a startup's lifecycle.
NarrativeNarrative_BuilderTrap Builder trap is not building instead of selling—it's not knowing what you're selling
NarrativeNarrative_BulkIngestionValidMethod Bulk Historical Transcript Ingestion via AI Summarization Is a Valid Onboarding Path
NarrativeNarrative_BusinessPlanAsSource Full Business Plan as Source of Truth
NarrativeNarrative_BYOKFeature BYOK (Bring Your Own Key) reframed from limitation to enterprise selling point
NarrativeNarrative_BYOKOpenRouterLimit BYOK Limitation: OpenRouter Only, Azure/Copilot Gap
NarrativeNarrative_CalibrationAsStaffWork Calibration Calls Are Staff-Work, Not Friendship
NarrativeNarrative_CareerArc Career progression from physics to NLP to organizational memory
NarrativeNarrative_CarfaxSubstackAnalogy Carfax for patients, Substack for practitioners
NarrativeNarrative_CascadingValue Cascading Collaborative Value
NarrativeNarrative_CaseStudyArchitectureInAswritten Building each case study's narrative architecture in aswritten dissolves the meta-loop and produces org-shaped example material
NarrativeNarrative_ChannelIsDistributionNotIdentity Channel is distribution, not product identity
NarrativeNarrative_ChatFirstCodeLater Chat-first adoption path: populate worldview before coding
NarrativeNarrative_CitationEngineNotSynthesis Citation Engine, Not Synthesis Engine: The Core Moat Over Generic AI
NarrativeNarrative_CiteConstellationModel Citations are constellations, not single claims
NarrativeNarrative_CiteDeduplicationInvariant Deduplication invariant: unique_citations ≤ supported; no two citations share a primary
NarrativeNarrative_CiteIntrospectAnalyticsGoldmine Cite and introspect tool pair is the highest-value analytics signal
NarrativeNarrative_CodeAsCompilation Worldview as source, code as output
NarrativeNarrative_CodeAsCompilationOutput Code as Compilation Output of Worldview
NarrativeNarrative_CodingAgentsContext Coding agents (Cursor, Claude Code, GitHub Copilot) query witness while implementing, aligning features to org direction automatically
NarrativeNarrative_ColdSubAgentTest Cold sub-agent test: spawn a zero-knowledge sub-agent with only a goal and URL; the only test that counts
NarrativeNarrative_ConciergeToProduct Concierge MVP as Product Validation
NarrativeNarrative_ConsensusAsSMS SMS-first consensus product: text a topic, name contacts, AI interviews each asynchronously, updates flow as contributions arrive, resolution produces a shared state object that merges into the worldview.
NarrativeNarrative_ConsensusEngine Consensus requires reaching each person through their worldview, not a generic prompt
NarrativeNarrative_ConsultingAsOnboarding Consulting engagement disguised as onboarding
NarrativeNarrative_ConsultingIsMeUsingProduct Consulting Is Me Using the Product, Not Getting Them to Use It
NarrativeNarrative_ContentAsReengagement Content & Newsletter as Re-engagement Tool
NarrativeNarrative_ContentCommentarySplit Content/commentary split: per-message integration is the content layer; reconciliation is the authored commentary layer
NarrativeNarrative_ContentDistributionAIAge Content Distribution in the AI Age: Modules as the Publishing Answer
NarrativeNarrative_ContentFlywheel Interview as Content Flywheel: Supply Side + Marketing + Demo
NarrativeNarrative_ContextArch_2 Propositions (subject-predicate-object) as the common element bridging language, databases, and logic
NarrativeNarrative_ContextArch_3 Compile-to-context is brute-force and hitting its ceiling; manifest + context pattern replaces it
NarrativeNarrative_ContextArch_4 Organizational memory works like human memory: directional, contextualized, sometimes imprecise, but attributable and auditable
NarrativeNarrative_ContextHemorrhage Organizations hemorrhage context by not enabling conversation at the point of record
NarrativeNarrative_ConvergenceAsSharpness The goal of review is sharpness on the branch's intent, not exhaustive coverage
NarrativeNarrative_ConvergenceNotCompleteness The goal of review is convergence on the branch's intent, not exhaustive completeness
NarrativeNarrative_ConvictionDecay Conviction should decay over time when claims are not reinforced
NarrativeNarrative_ConvictionFramework Conviction Levels as Organizational Memory
NarrativeNarrative_ConvictionLevelRename Conviction Level Rename: Metaphorical to Plain English
NarrativeNarrative_CoreConflationProductVsTooling Core Conflation: Product vs. Org-Development Tooling
NarrativeNarrative_CoreProductLoop Memory-to-worldview-shift pipeline as the core product loop
NarrativeNarrative_CorrectBeforeBuilding Foundational misunderstandings must be corrected before any building begins
NarrativeNarrative_CPCOnGoalURL CPC-on-Goal-URL for External, Subscription for Internal
NarrativeNarrative_CreationGap The most consequential organizational knowledge was never written down and cannot be retrieved — it must be extracted
NarrativeNarrative_CustomerArchitectCuratedExpertisePitch Sponsor independently reproduced the curated-expertise GTM framing
NarrativeNarrative_CustomerArchitectPitch Enterprise buyer value articulation
NarrativeNarrative_DataHandlingSubstantiated Data-handling posture is now substantiated by working code (Task-240 Phase 1)
NarrativeNarrative_DataOwnership Client owns memories, transactions, custom ontology extensions; base ontology licensed; repo under client org from day one
NarrativeNarrative_DayToDayRelitigation Surviving Next Tuesday: The Day-to-Day Re-litigation Problem
NarrativeNarrative_DelegationAsValue The value of collective memory is delegation — enabling someone who knows nothing about your business to do work that would normally require your most senior person
NarrativeNarrative_Deliverable_OrgChangelog Quantified organizational change, human readable org changelog
NarrativeNarrative_DemocratizationContext Democratization of Strategic Context
NarrativeNarrative_DemocratizedConsensus Democratizing knowledge collection creates balanced organizational representation and reduces political influence
NarrativeNarrative_DemoGIFStory The Money GIF Story: Before and After Perspective Installation
NarrativeNarrative_DemoScriptOutsideTechBetaUser Demo Script Finalized: a beta user from outside tech / Bluebird Learning Partners Scenario
NarrativeNarrative_DemoVsProductOnramp Demo vs. Product: The Share Demo Is a Demo; the Product Is Your Own Perspective
NarrativeNarrative_DensityInversion Density inversion: for AI, conceptual input density has no upper bound and more dense is always better; for humans it is the opposite. Dense thinkers become gold mines for AI-mediated communication.
NarrativeNarrative_DensityInversionAtScale Personal graphs enable density inversion at scale: dense thinkers' communication liability becomes an asset
NarrativeNarrative_DesignByRejection Architecture emerges through successive rejection of conflation and metadata drift
NarrativeNarrative_DesignTokenPerspectiveParallel Design-token extraction process is structurally identical to perspective extraction: both tokenise a lens, apply it through AI to surfaces, and break without the lens
NarrativeNarrative_DevOpsUseCase Untapped DevOps Use Case
NarrativeNarrative_DigestAsRecognitionAid Digest as Recognition Aid Over Lossless Store
NarrativeNarrative_DigestBrowserProductSurface The product surface is the digest browser — the human as consumer of the same collapse/expand navigation that extraction and citation already use
NarrativeNarrative_DinosaursVsAINatives Old companies are dinosaurs facing an asteroid; AI-native startups will move Amazon-fast
NarrativeNarrative_DirectIngestionAntiPattern Direct Historical Transcript Ingestion Is a Trap
NarrativeNarrative_DirectVsInfluenceSales a demo-call prospect does direct sales (face-to-face with executives, needs pre-read and demo materials); Tony does influence sales (warm intros, one-sheet as leave-behind). Explicit division of labor: opening doors vs. providing depth. a d
NarrativeNarrative_DiscoveryConsultantThreeFramings a consultant who ran enterprise discovery engagements for fifteen years's three framings of the aswritten/knowledge-graph overlap
NarrativeNarrative_DiscoveryOnboardingGovernanceSalesArc Discovery / Onboarding / Governance as a single-customer sales arc
NarrativeNarrative_DocumentAsSelfDescribing The document speaks for itself; metadata is brittle and prone to drift
NarrativeNarrative_DocumentEcosystem Document Ecosystem: Strategy Doc as Internal Source of Truth for All Derived Documents
NarrativeNarrative_DryRunFirst Preview before commit
NarrativeNarrative_ElephantCollective Elephant Collective: narrative as founder's weapon
NarrativeNarrative_EnterpriseAsRevenueNotPositioning Enterprise Consulting as Revenue Path, Not Positioning
NarrativeNarrative_EnterpriseComplianceGate Enterprise compliance as adoption gate
NarrativeNarrative_EnterpriseCustomerPatternRepeatable The the enterprise customer pilot validates a repeatable go-to-market shape worth codifying
NarrativeNarrative_EnterpriseCustomerWeekOneConvergence the enterprise customer ARB Job-to-be-Done Converged in Week One
NarrativeNarrative_EnterpriseShapePrimaryFocus Enterprise-Shape as Tentative Primary Sales Focus
NarrativeNarrative_EverythingHasAConversation Everything is a URI with a conversation attached — the Tim Berners-Lee dream applied to organizational records
NarrativeNarrative_ExecutionToStrategyShift AI Restored IC Status: The Real Work Is Curation and Strategy, Not Execution
NarrativeNarrative_ExhaustedExpertTrigger The Exhausted-Expert Purchase Trigger
NarrativeNarrative_ExpansionTracks Three expansion tracks: Core (Product/Eng), Enterprise (On-prem/Compliance), Executive (Founders/Board)
NarrativeNarrative_ExpertiseOnDemandReframe Expertise on Demand, Not Ambient Worldview
NarrativeNarrative_ExpertiseProvenancePositioning Positioning on Expertise and Provenance
NarrativeNarrative_ExpertSingleStreamWorldview Expert plan uses single-stream worldview: no branching, no PRs
NarrativeNarrative_ExtractionScopeOrgWorldview Extraction Scope: Organizational Worldview, Not Domain Index
NarrativeNarrative_FableWakeUpCall Fable 5 takedown radicalizes anti-lock-in moat from feature into means-of-production politics
NarrativeNarrative_FilteringDivergenceABAC Productive divergence: a consultant who ran enterprise discovery engagements for fifteen years uses named-graph ABAC filtering; Scarlet uses graph-walker compile
NarrativeNarrative_FirstAppointmentNotAlternative The First Appointment Is Not the Alternative — It's a Category Change
NarrativeNarrative_FocusAsGoverningPrinciple Radical product focus: bounded behaviors done extremely well
NarrativeNarrative_FocusOvertooling Radical product focus: bounded behaviors done extremely well over feature breadth
NarrativeNarrative_ForgettingBottleneckIsRole The Forgetting Bottleneck Is a Role, Not a Person
NarrativeNarrative_FortuitousTimingPair Fortuitous timing: meetings shifted to calls (accelerated by lockdown), making the transcript a known constant — the precondition for aswritten
NarrativeNarrative_FounderAsAuthorityLayer The founder's personal vision is a distinct authority layer that cyclically builds credibility for both the person and the business
NarrativeNarrative_FounderFitB2CTurn Founder-fit B2C turn: B2C is not just the larger market but the channel that plays to the founder's actual edge — narrative, writing, content
NarrativeNarrative_FramingLayerDeflection Framing-layer over-resolution and exit-ramp generation
NarrativeNarrative_FrontHITLSubstrate Front App as the Human-in-the-Loop (HITL) coordination substrate
NarrativeNarrative_FullEngagementUnlocksTesting Withholding engagement from unpaid relationships undermines product validation; inverting this is the beta strategy
NarrativeNarrative_GenerativeSystems Generative systems as conceptual foundation
NarrativeNarrative_GIFBasedSalesMaterials GIF-Based Sales Materials: Hook GIF, Story Page, McDonald Picture Menu Deck
NarrativeNarrative_GitHubAnalogy Collective memory as infrastructure, not feature—category-creation move
NarrativeNarrative_GitHubAsInfrastructure GitHub-native architecture is infrastructure, not product. The product is the channel-agnostic interface on top — non-engineers never see GitHub; they interact via simple natural language triggers.
NarrativeNarrative_GitHubAsInstitutionalMemory GitHub, abstracted through the 'as written' interface, enables version-controlled institutional knowledge management for non-programmers.
NarrativeNarrative_GitHubIsInfraNotIdentity GitHub is infrastructure, not identity; Outseta owns identity and billing
NarrativeNarrative_GoalDirectedPrimitive The Goal-Directed Agent Primitive
NarrativeNarrative_GovernanceAsSpeciesOfSteering Governance Is One Species of Steering — Not the Category
NarrativeNarrative_GraphInMemoryStorage The graph can live in memory, stored nowhere by default — download at session end, upload to resume; GitHub becomes optional storage, not a required third party
NarrativeNarrative_GraphLLMMoat The Technical Moat: Graph Store + LLM Fuzzy Traversal
NarrativeNarrative_HallucinationReframe Hallucination reframe: long-term aswritten use makes hallucination a foreign concept — the LLM becomes a language tool to navigate a graph of provenance, not a generator
NarrativeNarrative_HarmonyLabs Harmony Labs: narrative analysis at scale
NarrativeNarrative_HeroVideoFormat Hero Video Format: Human Moment, Not Screen Recording
NarrativeNarrative_HowToTryItSection How to Try It Section
NarrativeNarrative_HubCollectionMethodAtVolume Hub-Collection Method at Volume: One Memory Per Hub, Not Per Transcript
NarrativeNarrative_HubMemoryProvenanceNonNegotiable Provenance Survives Summarization: Hub Memories Must Contain Primary Source Material
NarrativeNarrative_HumanDrivenDevelopmentPrinciple All Software Development Must Be Directly About People and Their Needs
NarrativeNarrative_ICPRefinedTwoSubsegments ICP Refined: Functional Volume Across Two Subsegments
NarrativeNarrative_ICPValidation an infrastructure advisor validated target customer profile: 5-150 people, [amount]-50M revenue dev shops; noted even large orgs have small core teams (10-15 people at Google Search)
NarrativeNarrative_IdentityOnboardingGap Identity Onboarding Gap
NarrativeNarrative_ImplicitGoalForcingFunction Goal naming as product value
NarrativeNarrative_IndividuallyOwnedAGIReclaim The reclaim: labs build the general (commodity substrate); what you own is the particular — the perspective installed on the model that makes it think as a distinct individual
NarrativeNarrative_IndividuationCurveConversionMechanic Engagement = individuation curve (generic → first divergence → accretion → ownership → expansion); conversion mechanic is own-by-default, pay-to-continue
NarrativeNarrative_InformationVsPerspective Information is the artifact; perspective is the why behind it
NarrativeNarrative_InfrastructureGap Organizational worldview as git-native code at the core of CI/CD enables agentic automations, multi-player contribution, and composable distribution
NarrativeNarrative_InfrastructureVsProduct Infrastructure vs Product: SMS as the product layer
NarrativeNarrative_IngestionIsProtocol Ingestion is a protocol inside the user's agent, not a connector farm built by aswritten
NarrativeNarrative_InPersonSalesAdvantage Physical presence enables organic live demos that convert to warm referrals in ways remote calls cannot replicate
NarrativeNarrative_InterviewForcingFunction Interview Sales Motion as Forcing Function for Human-Driven Development
NarrativeNarrative_InterviewIsTheBusiness The Interview IS the Business
NarrativeNarrative_IPProtection Moat Protection Through Integration
NarrativeNarrative_KeepingBeatsStarting Keeping Beats Starting: The Conversion Model Flip
NarrativeNarrative_KnowledgeExtractionBeachhead Knowledge Extraction as Enterprise Entry Point
NarrativeNarrative_KnowledgePackagingMoment Knowledge Packaging as Second Moment of Value (an early prospect Aha)
NarrativeNarrative_LandingPageEpistemicArchitecture The landing page's information architecture mirrors the product's own epistemics: primary source material plus commentary and framing
NarrativeNarrative_LaunchFlightControlArc The Launch Flight Control Arc: March 19 – April 14, 2026
NarrativeNarrative_LeanStartupUpdatedForAI Lean Startup Updated for AI: Start with Humans, Work Back
NarrativeNarrative_LivingEvolvingModules Living Modules: Auto-Updating Expertise vs. Static Books
NarrativeNarrative_LLMsContinuation LLMs as continuation machines (2019 Clojure talk)
NarrativeNarrative_LongTailPerspectiveAutomation Long-Tail Automation: AI Agents Farm Perspectives at Scale
NarrativeNarrative_LossyCompressionProblem A compressed worldview summary is already a lens — shifting it shifts lossy data, not knowledge
NarrativeNarrative_ManagedRepoPerUser Every user gets an owner/repo managed entirely by aswritten — the repo is invisible infrastructure
NarrativeNarrative_MatchingSearchReframe Perspectives as Matching-Search: The Category Reframe
NarrativeNarrative_MCP_InstallationBarrier MCP installation as adoption barrier
NarrativeNarrative_MCPPositioning MCP as Standardization Point
NarrativeNarrative_MCPToolSurface_FiveTools The five user-facing MCP tools are perspective, cite, introspect, remember, and review
NarrativeNarrative_MemoriesAsCollaborativeNarrative Memories as collaborative narrative-building, not assessment
NarrativeNarrative_MemoryAsCollaboration Remember is collaboration — across teams and across chats — not a save action
NarrativeNarrative_MemoryAsSourceOfTruth Transactions and memories are the source of truth; generated docs are transient and regenerable
NarrativeNarrative_MemoryAsWitness User watches their own words become structured memory in real time
NarrativeNarrative_MemoryIngestionAsCodeReview the customer's CTO reframed AI memory ingestion as code review with PRs and unit tests
NarrativeNarrative_MetaphorClarity Internal metaphors (optometrist/eye chart) useful for system design but confusing for end users; present process in plain language instead.
NarrativeNarrative_MethodologyDensityOverContent Methodology Density Over Content Density: The High-Value Practitioner Case
NarrativeNarrative_MetricReframe Reframe value in ratios/time units so buyers apply own context
NarrativeNarrative_ModelAsOS Model as operating system; perspective as application that runs on top
NarrativeNarrative_ModelAsOSNotApplication Model as operating system, not model as application — the program (perspective) is what matters, not the hardware (model)
NarrativeNarrative_ModuleLayeringComposability Module Layering: Compounding Perspectives in a Single Chat
NarrativeNarrative_MonetizationAuthorConversion Monetization principle: paywall converts recipients into authors
NarrativeNarrative_MoneyAnswerDecisionArchaeology The money answer is a real concrete question whose answer comes back grounded in dated cited conversational provenance — strongest form is the decision-archaeology answer
NarrativeNarrative_MultiTenantHub Directory inheritance enables multi-tenant Hub deployment for consultants working across multiple clients
NarrativeNarrative_MultiUserRepoPromptChallenge Multi-User Repo System Prompt Challenge
NarrativeNarrative_MVPScopeDecision MVP Scope: Core vs. Future vs. Internal-Only
NarrativeNarrative_NamedPerspectiveArchitecture Named Perspective Architecture: Advisors as Sub-Agents
NarrativeNarrative_NamingConstraints Naming Constraints and Direction
NarrativeNarrative_NarrativeAsVector Narrative as velocity/acceleration vector in embedding space
NarrativeNarrative_NavigationReplacesEmbeddedSnapshot Navigation replaces embedded snapshot: every consumer reads through the digest
NarrativeNarrative_NeurosymbolicKGLLM Neurosymbolic marriage of knowledge graphs and LLMs as shared community framing
NarrativeNarrative_NoCustodyDataPhilosophy aswritten is a processor, not a data store — users own their transcripts
NarrativeNarrative_NonTechnicalAccessibility Accessibility for non-technical users
NarrativeNarrative_NoSecretsOnboarding Client repos require no secrets for onboarding
NarrativeNarrative_NotificationAsEpistemicAlignment Notification collapses into persuasion when driven by personal worldview graphs
NarrativeNarrative_Offering_PnD AI-Native Narrative Strategy consulting service
NarrativeNarrative_OnboardingGap The gap between technical connection and first value is the core onboarding problem
NarrativeNarrative_OnboardingUseCase Onboarding humans and AIs into an organizational perspective
NarrativeNarrative_OnPremAsRequirementsDriven On-Prem as Requirements-Driven Add-On, Not Tier Differentiator
NarrativeNarrative_OntologyAsSpeechActs Ontology modeled as actors, speech acts, and documents — not facts and entities
NarrativeNarrative_OntologyDomainAgnostic Ontology Scope: Core Primitives Are Domain-Agnostic
NarrativeNarrative_OpinionatedPromiseMissingArticulation The customer-facing promise 'have an opinion and here's why' is the missing marketing articulation — present as description in brand docs but absent as a dare
NarrativeNarrative_OrgKnowledgeAsCode Organizational worldview should be versioned, collaborative, and graph-structured like software—enabling cascading updates across all AI-generated content
NarrativeNarrative_OrgMemoryDemotedToGovernanceSub Organizational memory demoted from the pitch to a governance sub-use-case
NarrativeNarrative_OrgTierAgentDeploymentValue Org Tier Value Prop: Three Capabilities, Not Just EDOs
NarrativeNarrative_Outcome_1 New hires ask AI instead of leadership; strategy-to-docs time: weeks → 15 min; branchable narrative testing
NarrativeNarrative_OutputBeforeSystem Prospects must experience output (content reflecting their input) before understanding the system.
NarrativeNarrative_OwnFromDayOne Own-From-Day-One: Client Owns Repo via Team Plan
NarrativeNarrative_PersonalGraphFederation Personal graphs federated with org repos act as both filter and expander on worldview changes
NarrativeNarrative_PersonalGraphLayer Personal worldview graphs as a missing product layer beneath organizational memory
NarrativeNarrative_PerspectiveAsAgentSteering Perspective as agent steering: install behavior, values, and self-understanding through accumulated human-agent conversation rather than RAG
NarrativeNarrative_PerspectiveDrivenDevelopmentInversion Perspective-Driven Development Inversion: From User→AI to Perspective→User
NarrativeNarrative_PerspectiveOnSpec Perspective-on-spec as AI-mediated ad spend
NarrativeNarrative_PerspectiveResponseOrderedByVolatility Perspective response ordered by volatility: stable skeleton leads, per-call focus appended last
NarrativeNarrative_PerspectiveShares_Distribution Perspective shares as self-contained worldview bundles
NarrativeNarrative_PerspectiveVsImplementationLine Perspective holds decisions and reasoning; implementation holds literal detail; both are knowledge but belong in different places
NarrativeNarrative_PerspectiveVsInformation Perspective vs. Information: The Core Distinction Between Modules and RAG
NarrativeNarrative_PerTaskScopedPermissions Per-task scoped permissions: OAuth scope narrowed to single-use grants for agentic operations
NarrativeNarrative_PerUserAdoptionAnalysis Pilot health requires per-user cohort analysis over time, not aggregate metrics
NarrativeNarrative_Phase1Phase2Pattern The Two-Pass ARB Pattern
NarrativeNarrative_PilotOffer Consulting pilot as entry point for collective memory adoption
NarrativeNarrative_PipelineSelfDocumentation_20260704 The product documents itself: cutover decision saved through new production pipeline, citation offsets slice-exact on first try
NarrativeNarrative_PitchFrameEvolution Value frame evolution from matching to proposal
NarrativeNarrative_PlanGatedInstructions Compile returns plan-gated instructions as the feature-gating mechanism
NarrativeNarrative_PodcastInterviewSalesMotion The Podcast/Interview as Sales Motion: Three Outputs from One Conversation
NarrativeNarrative_Populous Populous: medium-agnostic communication system
NarrativeNarrative_PostPilotSelfSufficiency Post-pilot team autonomy: no ongoing dependency
NarrativeNarrative_PractitionerPaymentReframed Patient offers to pay for practitioner perspective
NarrativeNarrative_PractitionerThreeRoles The Practitioner's Three Roles in the System
NarrativeNarrative_PRasConsensus Git PR as organizational consensus metaphor: memory = PR, AI-mediated interview = automated review, approval = merge into worldview. Scales to any novelty requiring group consensus.
NarrativeNarrative_PrepackagedModuleConcept Prepackaged Expertise Modules as Answer to Activation Gap
NarrativeNarrative_PreReadDeckConcept a demo-call prospect's request: a visual-first deck executives can absorb before a call. Not a pitch deck — a pre-read that makes the meeting productive from minute one. a demo-call prospect's framing: 'Explain like I'm five for executives,
NarrativeNarrative_PreRegisteredEvalDiscipline Pre-Registered Eval Battery as the Bar for Validated
NarrativeNarrative_PriceTheAccumulation Price the Accumulation, Not the Conversation
NarrativeNarrative_Process_1 Biweekly cadence, team interviews, narrative evolution, structured meeting minutes
NarrativeNarrative_ProcessorNotStore aswritten is a processor, not a data store — enabling zero data retention for high-compliance environments
NarrativeNarrative_ProductAsProof The YC application itself was generated by AI using collective memory—'one shot and usable'—demonstrating the product's capability to produce human-quality, contextualized output.
NarrativeNarrative_ProductKnowledgeSecondGraph Compile returns two graphs: org worldview AND aswritten product knowledge graph — self-updating, no static artifacts
NarrativeNarrative_ProductSurfacesOwnRoadmap The Product Surfaces Its Own Roadmap Through Use
NarrativeNarrative_ProductVision Product as Platform, Not Infrastructure
NarrativeNarrative_ProfessionalOverEnterprise 'Professional' Tier Name Over 'Enterprise'
NarrativeNarrative_ProgressionPrinciple Give the human more and more concrete material to react to until they are only correcting, not generating
NarrativeNarrative_PromptAsCompilationFromArchitecture Prompts are compiled from narrative architecture, not authored
NarrativeNarrative_PromptsAsOntologyTargets LLM prompts as ontology compilation targets: versioned, scoped, and recompiled when the worldview shifts
NarrativeNarrative_ProposalArtifactPractitioner The Proposal: What the Patient Receives from the Practitioner's AI
NarrativeNarrative_ProtocolEnrollmentOverCommand Protocol must earn model trust through enrollment, not command
NarrativeNarrative_ProvenanceCitationMoment Provenance/Citation as Primary Moment of Value
NarrativeNarrative_ProxyInversionHumanInsulation Proxy inversion: SII acts on the human's behalf so humans are insulated from AI's constant interruptions and oversight demands
NarrativeNarrative_PublicPerspectiveModule Public Perspective as First Module
NarrativeNarrative_PureFunctionArchitecture Pure function discipline applied to workflow orchestration: LLM workflows are pure (input → output), Flows handle all side effects at the edges
NarrativeNarrative_PureFunctionDesign Pure function architecture separates orchestration from computation
NarrativeNarrative_QuantifiedOrganization The quantified/reified organization as an introspectable asset with value over time
NarrativeNarrative_QuestionReframe_MVPToUnlocked From 'What Is the MVP?' to 'What Did We Just Make Possible?'
NarrativeNarrative_RDFOverDatomicRegret Shared regret: Rich Hickey was close to making Clojure/Datomic web-native RDF but stopped short
NarrativeNarrative_ReachabilityWithoutInterruptibility Reachable without being interruptible
NarrativeNarrative_ReadPathFirst Demonstrate the product at full maturity (read path) before showing how it is built (write path)
NarrativeNarrative_ReconciliationAsSetTransaction Reconciliation is a transaction about a set of transactions or nodes — the design of correction and commentary about the past
NarrativeNarrative_ReferrerGiftEconomy Trusted Referrers Stay in the Gift Economy — No Money in the Loop
NarrativeNarrative_RegisterInformNotInstruct Inform about the system; do not instruct the agent
NarrativeNarrative_RelationshipCadenceAsset The relationship cadence is a strategic asset for renewal
NarrativeNarrative_ReportAsDeliverablePattern The monthly curated usage report as a core sales motion pattern
NarrativeNarrative_RepricingHumanTime Human time will be repriced upward as AI scales — meetings and transcripts become more valuable
NarrativeNarrative_RetireZeroContext The zero-context framing is a liability with sophisticated buyers who already have context solutions; replacement framing acknowledges the industry and positions against it
NarrativeNarrative_RevenueBeforeRaise Sell first, raise from strength rather than from projections
NarrativeNarrative_RevenueTradeoff Strategic Focus Over Near-Term Revenue
NarrativeNarrative_ReviewableWorldview The Reviewable Worldview demo supersedes Demo 1 and shows the full product: capture, collaborative review, validation, and merge
NarrativeNarrative_ReviewLoopIsRemember Review Is Just Another Remember: The Natural Loop
NarrativeNarrative_ReviewMemoryFlow Saving a review memory ends a round and triggers extraction; merge signals no corrections needed and closes the PR — corrections always require at least two rounds.
NarrativeNarrative_ReviewProcess Iterative review process design requires establishing baseline before introducing shifts, labeling intent explicitly, and managing scale through chunked comparisons.
NarrativeNarrative_ReviewReturnValuePattern Prompting in Return Values: Product-Wide Shift in Agent Communication
NarrativeNarrative_RichnessVsVisibilitySeparation Richness is a budget decision; visibility is an invariant — the principled separation that makes tier weighting compatible with compression-only
NarrativeNarrative_RichTierBySteeringWeight Rich-tier slot allocation by steering weight: conviction-carrying children ranked over style-taxonomy record types
NarrativeNarrative_SalesAsHardPart Sales is the hardest part: the value proposition is technically oriented but buyers are not
NarrativeNarrative_SalesCallIsOnboarding Team Tier: Sales Call = Onboarding = First Knowledge Extraction
NarrativeNarrative_SalonAsCapability Salon is a per-memory capability, not a fifth deployment pattern
NarrativeNarrative_ScalingGap Only 33% scaling AI across organization—rest stuck in silos
NarrativeNarrative_ScarceExpertFilterGoalGated Unified Instrument: Scarce-Expert Filter with Goal-Gated Escalation
NarrativeNarrative_SearchVsAswritten Disambiguation: Search vs. aswritten
NarrativeNarrative_SecondWaveBridge aswritten as the Bridge to the Second Wave
NarrativeNarrative_SecurityFiltering Security Filtering as Enterprise Requirement
NarrativeNarrative_SelfServeDemoOnOwnPerspective The self-serve demo should run on aswritten's own perspective from a trial session: solves cold-start, is self-demonstrating, and the honesty beats are the least-fakeable material
NarrativeNarrative_SeparationOfConcerns Separation of concerns as architectural virtue: LLM workflows must not commit files, send notifications, or perform I/O — those responsibilities belong exclusively to orchestrating Flows
NarrativeNarrative_SeparationPrincipleProductVsOrg Separation of product features from org automation: what ships vs. what Scarlet runs internally
NarrativeNarrative_SeparationProductVsOrgAutomation Product features vs. org automation must be separated
NarrativeNarrative_ShareDataModel Goals are properties of a share, not the perspective. This prevents instance-specific goals from contaminating the reusable perspective.
NarrativeNarrative_ShowDontExplain Let the output speak; resist the over-explain trap
NarrativeNarrative_ShowVsDemoProof Value is proven by showing clearly; their-own-data is the conversion deepener, not the proof
NarrativeNarrative_SIC_Behaviors_1 Compile (replay transactions to snapshot), Extract (RDF from input), Diff (semantic comparison), TX (generate transaction), Commit (append to Git), Evolve (operation on snapshot).
NarrativeNarrative_SIC_CaseStudies_1 Planned demo for Crooked Media (Pod Save America): auto-ingest transcripts/show notes, generate updating stories, enable perspectival operations on podcast narrative.
NarrativeNarrative_SIC_DataModelLifecycle_1 Append-only transaction log in .aswritten directories. Hierarchical compilation across ancestor directories. Immutable transaction files with provenance metadata.
NarrativeNarrative_SIC_Flows_1 Individuation pipeline: extract → diff → TX → user review → commit. Evolved variant: snapshot + message → transaction. Story generation: compile → template → model → output.
NarrativeNarrative_SIC_Mission_1 Extend software development rigor into strategy, content, and marketing product organizational operations.
NarrativeNarrative_SIC_MoatLeverage_1 RDF-based narrative architecture with Git-native versioning, SHACL validation, and semantic diff operations. Marketplace of reusable perspective snapshots.
NarrativeNarrative_SIC_Opportunity_1 AI context quality problem
NarrativeNarrative_SIC_OrgByDomain_1 Synthetic Identity Co (SIC) is legal entity. aswritten.ai is user-facing brand. Collective memory is RDF narrative source of truth (internal). STORYBASE is capital-case internal identifier.
NarrativeNarrative_SIC_PrimaryActors_1 Programming literate entrepreneurs, designers, developers, consultants with Git and basic AI coding fluency.
NarrativeNarrative_SIC_Primitives_1 RDF transactions (currently SPARQL, migrating to TriG named graphs), .aswritten directory hierarchy, .story files with YAML front matter, collective memory snapshots in Turtle.
NarrativeNarrative_SIC_ProductOverview_1 Initial prototype in n8n with tools: aswritten/compile, collective memory/ontology, aswritten/extract, aswritten/diff, aswritten/tx, commit. Hosted at aswritten.ai with Open WebUI.
NarrativeNarrative_SIC_Tagline_1 aswritten.ai — AI that tells you a story as written
NarrativeNarrative_SIC_WhatDoesItDo_1 Replaces brittle role prompts with deep narrative source of truth that can be clearly understood and operated on. Uses organizational metadata/configs to steer lightweight prompt execution.
NarrativeNarrative_SIC_WhatIsIt_1 AI memory that is versionable, collaborative, branchable, so Git native, and narrative driven. Encodes style and conviction and other metrics.
NarrativeNarrative_SII_SynthesizedIndividualIntelligence Synthesized Individual Intelligence (SII) as the owned, portable, relational alternative to lab-owned AGI
NarrativeNarrative_SMBPivot a demo-call prospect's business plan targets SMB — described as a bigger market than devs. The business plan a demo-call prospect takes to prospects needs to reflect the SMB pivot. Prior targeting was developer-first (PLG motion per ADR-002
NarrativeNarrative_SMBPrimaryMarket SMB as Primary Market
NarrativeNarrative_SoakTestQuality Quality validation requires sustained real-world use, not unit tests
NarrativeNarrative_SPARQLUpdateAsLogPrimitive SPARQL Update as the log primitive — independently chosen by both Scarlet and a consultant who ran enterprise discovery engagements for fifteen years
NarrativeNarrative_StaffingAIAdoptionWall AI that works too well can create internal organizational conflict, blocking adoption
NarrativeNarrative_SynchronousPerMessageIntegration Integration is synchronous, per-message, and silent — async breaks the individual and the deliberate feedback-in-chat decision
NarrativeNarrative_Tagline_ConsultingPitch Collective Memory for AI-Native Teams
NarrativeNarrative_TamarackDemoAsLiveSalesMotion The demo with a collaborator is explicitly framed as the same sales motion used with any SMB prospect — not a special investor demo. Live knowledge extraction, product walkthrough, account setup. a collaborator is technically savvy but not
NarrativeNarrative_TechnicalAdoptionFriction Technical User Adoption Friction: Contamination Fear, Not Product Difficulty
NarrativeNarrative_UpdateProblem_Evolution The Update Problem: Knowledge Evolution and Abstraction
NarrativeNarrative_UseCase1GoverningHumanTeam Use Case 1: Governing a Human Team
NarrativeNarrative_UseCase2AgenticSteering Use Case 2: Agentic Steering (Currently Underrepresented)
NarrativeNarrative_UseCase3SharedProcessSteering Use Case 3: Steering the Shared Process Across Engagements
NarrativeNarrative_UserAgentIngestionModel One protocol, many sources: transcripts, AI sessions, codebases, docs, email all ingested via user's local AI context
NarrativeNarrative_UserFacingVsInternalLanguageSplit User-facing tool names and internal system terms are intentionally different layers with a proxy mapping between them
NarrativeNarrative_V0_ChannelAgnostic V0 primitive is channel-agnostic
NarrativeNarrative_V0_SMS_DeepLink V0 Scope: SMS Deep Link Demo
NarrativeNarrative_ValueDemonstrationGap a demo-call prospect's core feedback: the product needs to show how aswritten influences AI behavior. Agents should cite which memories shaped their recommendations. This is the 'value demonstration' problem (task-145). Without visible proo
NarrativeNarrative_VCChannel VCs get better portfolio visibility (queryable, not just reported), faster board prep, cross-portfolio learning
NarrativeNarrative_VCPivot_OrganicScale Shift away from venture capital toward organic scale
NarrativeNarrative_VerifyAtUserSeam Verify at the seam the user touches, not one layer above it
NarrativeNarrative_ViralSMSModel Natural viral loop: first topic free, each named recipient experiences product free, post-resolution invitation triggers 1-in-5 conversion. Cost model: expensive models (Claude/Opus) for extraction, Gemini Flash for everything else.
NarrativeNarrative_VisibilityInvariantAreaIndex Complete area index: every hub named in every response, one-line entry with child count, one expand away
NarrativeNarrative_VoiceMemoAsInput Long rambling voice memos are high-quality memory input for the extraction pipeline
NarrativeNarrative_Vouch Vouch: the manual prototype
NarrativeNarrative_VouchIdentityForAgents Vouch.io Identity Infrastructure for Agent Security
NarrativeNarrative_WarmStartVsColdStart Warm-Start vs Cold-Start Review: Two Entry Points into Shared Core
NarrativeNarrative_WatcherAgent The watcher agent concept: instead of one person intentionally saving memories, an agent watches the evolving worldview and proactively notifies affected parties when someone else's contribution changes their domain. Cross-domain notificati
NarrativeNarrative_WedgeCashPayNYC Wedge: Cash-Pay Alternative Care, NYC, Density First
NarrativeNarrative_WhatIsIt_ConsultingPitch Collaborative AI with a single source of truth (collective memory)
NarrativeNarrative_WhitePaperAIFilter White Paper AI Filter: Busy Executives Won't Read, They'll Summarize
NarrativeNarrative_WhiteSpace AI Fills Unspecified Space with Training Defaults
NarrativeNarrative_WhiteSpaceCollapse Organizational perspective collapses the white space of LLM output toward the right answer
NarrativeNarrative_WhiteSpaceMechanism White space: the organizational perspective fills in what the model decides by default
NarrativeNarrative_WhyYC YC as best market partner: biggest network to most organizations
NarrativeNarrative_WorldviewAsProduct Worldview as Primary Product
NarrativeNarrative_WorldviewReview Review Worldview Deltas, Not Outputs
NarrativeSalesDecks Demo repo needed to show completed collective memory without exposing source
RubricAssessment (197)
RubricAssessmentAssess_Accuracy_cm Precise type signatures; accurate code examples; consistent terminology; references to prior memory (collective memory Data Model & API Architecture Discussion); SHACL validation schema included
RubricAssessmentAssess_Accuracy_collective_memory Precise technical specifications; explicit corrections documented ('Technical Corrections Made' section); consistent type definitions; accurate sequence diagrams; verifiable claims about system behavior
RubricAssessmentAssess_Cadence_cm Varied sentence length; mix of short declarative statements and longer explanatory sentences; good rhythm for technical documentation
RubricAssessmentAssess_Cadence_collective_memory Mix of short declarative sentences and structured lists; bullet points create scannable rhythm; sequence diagrams break up dense technical prose
RubricAssessmentAssess_Cadence_RevenueModelTAM Bullet lists create punchy rhythm; parallel 'No company' structure; avg sentence length 22 words (moderate, not short).
RubricAssessmentAssess_Flow_RevenueModelTAM Logical progression: revenue model → metric reframe → TAM expansion → contract tiers → expansion path. Section headers guide; transitions implicit but clear.
RubricAssessmentAssess_Novelty_cm Novel architecture patterns (State abstraction, Memory decomposition); some standard technical documentation phrases; Mermaid diagrams add visual novelty
RubricAssessmentAssess_Novelty_collective_memory Novel integration pattern (two-step save); creative use of Git as narrative log; fresh framing of 'Memory' as ephemeral input vs 'Transaction' as durable change; avoids generic SaaS architecture clichés
RubricAssessmentAssess_Novelty_RevenueModelTAM 'Remove dollar signs' reframe is fresh positioning move; GitHub analogy is established but well-executed; TAM expansion sourced (not boilerplate).
RubricAssessmentAssess_Phrasing_collective_memory Domain-specific terminology (Memory, Transaction, Snapshot) used consistently; clear naming conventions (mcp/ prefix for tools); technical idioms appropriate for audience
RubricAssessmentAssess_Phrasing_RevenueModelTAM Stock phrase 'remove dollar signs' encodes design principle; 'collective memory' canonical; GitHub analogy is signature move.
RubricAssessmentAssess_Register_cm Formal technical register; authoritative tone; appropriate for architecture documentation; no casual markers
RubricAssessmentAssess_Register_collective_memory Technical register with clear structure; appropriate formality for architecture documentation; consistent use of second-person in user-facing descriptions
RubricAssessmentAssess_Resonance_cm Uses concrete examples (test cases, code snippets, diagrams); analogies implicit in architecture patterns; could add more narrative framing
RubricAssessmentAssess_Resonance_collective_memory Concrete examples (DC37 training story, email domain migration); sequence diagrams illustrate workflows; use cases ground abstract patterns in real scenarios
RubricAssessmentAssess_Resonance_RevenueModelTAM GitHub analogy is high-resonance for technical founders; 'context leaves with departures' is visceral pain point; McKinsey 33% stat validates urgency.
RubricAssessmentAssess_Strategy_cm Directly aligned with Architecture domain; supports Strategy (roadmap, sequencing), Product (capabilities), and Organization (team responsibilities); clear implementation phases
RubricAssessmentAssess_Strategy_collective_memory Strong alignment with narrative architecture principles: accessibility (non-technical users), provenance (Git as source of truth), collaboration (two-interface model), auditability (append-only transactions)
RubricAssessmentAssess_Tailoring_cm Tailored for technical audience (engineering, architecture teams); includes code examples, type signatures, implementation details; assumes domain knowledge
RubricAssessmentAssess_Tailoring_collective_memory Explicitly addresses multiple personas (nonprofit coordinators, sales execs, educators); separates user-facing from internal tools; includes both technical and non-technical stakeholder concerns
RubricAssessmentAssess_Tailoring_RevenueModelTAM YC application context explicit; ratio-based metrics designed for buyer self-application; TAM sourced for investor audience (Census, McKinsey).
RubricAssessmentAssessment_Accuracy_CustomerArchitect Factually accurate named entities (the enterprise customer, USPS, Österreichische Post, Poland Post); precise pricing signal ([amount]-150K); concrete timeline (late Feb beta); no unsupported claims.
RubricAssessmentAssessment_Accuracy_IPCrisis Accurate references to ADR-002; explicit 'uncommitted' status for exploratory claims; preserved verbatim quotes; acknowledges open questions; no false precision
RubricAssessmentAssessment_Accuracy_TaxonomyBlockers Specific metrics cited (243K tokens, [amount]+/day, 15-minute batch timer). Task references (task-070, task-072) are verifiable. No external citations but internal cross-references are present and specific.
RubricAssessmentAssessment_Accuracy_YC Factual: dates, names, entities, product details all verifiable. No unsupported claims. Citations implicit (e.g., 'YouTube video' reference).
RubricAssessmentAssessment_Cadence_CustomerArchitect Mix of short punchy sentences and longer explanatory clauses; natural speech rhythm; effective use of em-dashes and parentheticals.
RubricAssessmentAssessment_Cadence_IPCrisis Varied sentence length; strategic use of short punchy sentences for key claims; longer analytical passages; good rhythm variation
RubricAssessmentAssessment_Cadence_TaxonomyBlockers Effective short-punchy pattern: section headers are crisp ('Less May Be Better', 'Four Complementary Solutions'). Arrow-chain code block creates visual rhythm. Some mid-document paragraphs run long.
RubricAssessmentAssessment_Cadence_YC Short, punchy sentences; colon structures; fast cuts in demo storyboard. Rhythm supports urgency and clarity. Supporting observation: narr:StyleObs_NotionCadence
RubricAssessmentAssessment_Flow_IPCrisis Clear section structure; logical progression (problem → solution → implications); effective use of headers; transitions explicit; 'Open Questions' section acknowledges uncommitted areas
RubricAssessmentAssessment_Flow_TaxonomyBlockers Excellent structure: Problem \u2192 Solution \u2192 Related for each blocker. Summary table and ordering section provide strong closure. Arrow-chain diagram makes solution relationships explicit. Minimal formulaicness.
RubricAssessmentAssessment_Novelty_CustomerArchitect Fresh framing ('senior expert bottleneck' vs. 'AI memory lock-in'); some domain-standard phrasing ('signs the checks', 'locked in heads'); overall avoids cliché.
RubricAssessmentAssessment_Novelty_IPCrisis Fresh framing (IP crisis vs developer convenience); novel terminology (compilation targets, cultivation trap, context squishing); avoids generic AI hype language
RubricAssessmentAssessment_Novelty_TaxonomyBlockers 'Less May Be Better' inversion is genuinely novel framing. 'Semantic coordination problem' is precise and non-obvious. Some standard tech-doc scaffolding reduces novelty slightly.
RubricAssessmentAssessment_Novelty_YC Fresh phrasing ('worldview behind the skill', 'separated at birth', 'I know Kung Fu'). Avoids boilerplate. Conviction framework (notion/stake/boulder/grave) is novel taxonomy.
RubricAssessmentAssessment_Phrasing_CustomerArchitect Domain-specific verbs ('extracts', 'query', 'bootstrap'); stock phrases ('signs the checks'); idiomatic turns ('locked in people's heads').
RubricAssessmentAssessment_Phrasing_TaxonomyBlockers Strong idiolect: 'introspect call', 'worldview shift', 'layer0Plus1', 'squash/rebase pass' are domain-specific stock phrases that signal insider fluency. Highly distinctive.
RubricAssessmentAssessment_Phrasing_YC Strong idiolect: 'separated at birth', 'snappy, snappy', 'I know Kung Fu'. Domain verbs ('onboards'). Stock phrases signal identity and chemistry. Supporting observation: narr:StyleObs_SeparatedAtBirth
RubricAssessmentAssessment_Register_CustomerArchitect Conversational, direct, second-person ('you'); active voice dominant; enterprise-appropriate formality with accessible analogies.
RubricAssessmentAssessment_Register_IPCrisis Authoritative strategic register; second-person address to reader; active voice dominant; technical precision without jargon overload
RubricAssessmentAssessment_Register_YC Conversational, direct, second-person; active voice; sharp and confident. Appropriate for founder-to-founder working session.
RubricAssessmentAssessment_Resonance_CustomerArchitect Strong use of organizational in-joke ('duplicate the customer's chief architect's brain'), concrete analogies (pricing engine as test case), and relatable pain points (senior expert bottleneck).
RubricAssessmentAssessment_Resonance_IPCrisis Strong metaphors (existential crisis, deepest lever, state secrets); concrete examples (sales/eng value flow); git-native analogy resonates with technical audience
RubricAssessmentAssessment_Resonance_TaxonomyBlockers Git analogy is well-chosen and extended consistently. 'The viral product' framing adds aspirational resonance. Concrete numbers ([amount] tokens, [amount]+/day) ground abstract claims.
RubricAssessmentAssessment_Resonance_YC Strong analogies ('Git for software', 'I know Kung Fu'), founding story, conviction framework. Resonates with YC values: technical depth, founder chemistry, clear vision. Supporting observation: narr:StyleObs_GitMetaphor
RubricAssessmentAssessment_StrategicAlignment_CustomerArchitect Directly surfaces positioning thesis, value exchange model, actor incentives, and product-market fit insights; aligns with Strategy > Positioning Thesis and Opportunity > Actor Incentive Analysis.
RubricAssessmentAssessment_StrategicAlignment_IPCrisis Directly addresses core strategy shift; explicit connection to ADR-002; worldview-as-product aligns with platform positioning; IP protection moat
RubricAssessmentAssessment_StrategicAlignment_TaxonomyBlockers Explicit mapping of blockers to demos to revenue paths (enterprise consulting vs. consumer). Ordering section directly ties solutions to strategic priorities. Strong alignment to Focus & Sequencing.
RubricAssessmentAssessment_StrategicAlignment_YC Entire call aligns to YC application strategy: founding story, product vision, differentiation, why YC. Every narrative element reinforces core thesis: collective memory as foundation for AI-native teams.
RubricAssessmentAssessment_Tailoring_CustomerArchitect Explicitly tailored to enterprise buyer (CFO, CTO, CEO); addresses the enterprise customer-specific context (RFPs, pricing engine, JIRA workflow); the customer's chief architect's 30-second pitch is a model of audience fit.
RubricAssessmentAssessment_Tailoring_TaxonomyBlockers Clearly written for a technical co-founder audience. Named participants (Scarlet, Claude). References specific tasks (task-070, task-072). Assumes deep product context — appropriate for the audience.
RubricAssessmentRubric_Accuracy_CustomerArchitect Factually precise: named entities correct (the customer's chief architect, the enterprise customer, GitHub Copilot, GPT5.2 Codex, Opus), dates accurate (Feb 9 2026, Jan 27, Feb 14 vacation), technical details verifiable (60 markdown files,
RubricAssessmentRubric_Accuracy_FounderBetaUser Specific revenue figure ([amount]); precise date (Feb 5, 2026); direct quotes with timestamps; named entities (Press Warner Ventures, Google, LiveKit, Sierra); clear attribution of claims to a founder beta user; pending action items noted.
RubricAssessmentRubric_Accuracy_PilotPlan Specific pricing ([amount] pilot, [amount]/mo license); concrete deliverables (2-3 experts, 2-4 weeks); named tools (Claude, Cursor, MCP); Open Questions section flags uncertainties; no false claims detected.
RubricAssessmentRubric_Accuracy_ScarletDame Specific dates, numbers, names. Explicit corrections ('Do NOT list Cognitect'). Transparent about uncertainties ('Pending Questions'). Attribution clear ('Scarlet's words'). High factual rigor.
RubricAssessmentRubric_Accuracy_YC_Call Specific dates (Feb 5, 2026), names (Scarlet Dame, Tony Maley, the customer's chief architect, an infrastructure advisor), metrics (6 users → 12 next week, ~68 minutes, 3 years ago). High factual precision.
RubricAssessmentRubric_Accuracy_YCDemoBeats Accurate references to Decision_YCPositioning_20260203, Proof_DemoCriteria, Feature_OnboardingMode; gap names match Tony's identification; timing constraints (60s, 20s per beat) are precise; tool names correct
RubricAssessmentRubric_Assess_Register Conversational, direct, second-person; high active voice; appropriate for strategic planning session.
RubricAssessmentRubric_Assess_Resonance Strong use of analogies (Matrix chips, a collaborator the mainframe guy); founding story as narrative proof.
RubricAssessmentRubric_Assess_StrategicAlignment Entire conversation is strategic alignment exercise; vision, positioning, and narrative anchor all addressed.
RubricAssessmentRubric_Assess_Tailoring Tailored to YC application context; audience frames (economic buyer, technical, end user) explicitly discussed.
RubricAssessmentRubric_Cadence_CustomerArchitect Short, punchy sentences for emphasis ('It worked.', 'This is significant:'). Average 18.3 words/sentence with variation. Strong rhythm through mix of short declaratives and longer explanatory clauses.
RubricAssessmentRubric_Cadence_FounderBetaUser Varied sentence length; some long analytical passages balanced by punchy quotes; section headers provide rhythm; could benefit from more short declarative statements in analysis sections.
RubricAssessmentRubric_Cadence_PilotPlan Mix of short declarative sentences and longer explanatory clauses; em dashes for emphasis; bullet lists create rhythm; triadic structures (rule of three) throughout.
RubricAssessmentRubric_Cadence_ScarletDame Short, punchy sentences. Telegraphic style in places ('Bootstrapped from [amount]to [amount]/event in 6 months'). Effective use of em-dashes and arrows. Strong rhythm variation.
RubricAssessmentRubric_Cadence_YC_Call Short, punchy sentences; repetition for emphasis ('Snappy, snappy'); rule of three. Strong rhythm.
RubricAssessmentRubric_Cadence_YCDemoBeats Mix of short punchy beats (20-second structure) with longer explanatory sections; good rhythm variation; bullet lists create visual cadence
RubricAssessmentRubric_Flow_CustomerArchitect Natural progression: context → validation → use cases → compliance → product feedback → next steps. Section headers clear. Transitions effective ('This is significant:', 'the customer's chief architect identified', 'Scarlet connected'). Min
RubricAssessmentRubric_Flow_FounderBetaUser Clear section structure (Who/Three-Layer/Gods/Operations/etc.); logical progression from context → validation → implications; transitions between sections are natural; some sections could be tightened.
RubricAssessmentRubric_Flow_PilotPlan Clear progression: Goal → Pitch → Offer → Cost → Process → Expansion → Terms → Email → Open Questions; parallel structure in expansion tracks; natural transitions between sections.
RubricAssessmentRubric_Flow_ScarletDame Chronological career arc structure. Clear section headers. Natural transitions between roles. 'Pending Questions' section signals incompleteness transparently. Minimal formulaicness.
RubricAssessmentRubric_Flow_YC_Call Natural progression from founding story → product vision → YC strategy → demo storyboard → open gaps. Coherent structure; transitions clear.
RubricAssessmentRubric_Flow_YCDemoBeats Clear three-act structure; natural progression from problem→solution→propagation; transitions between beats are logical; Why This Works section provides coherent summary
RubricAssessmentRubric_Novelty_CustomerArchitect Fresh insights: AI-to-AI provenance pattern, token cost sensitivity in enterprise, 'killer use case when it doesn't know', collaborative steering of coding agents. Some domain-standard phrases ('cross-platform validation', 'enterprise compl
RubricAssessmentRubric_Novelty_FounderBetaUser 'Gods of Olympus' metaphor is fresh; 'ambient adoption' framing is distinctive; operations persona angle is new; avoids generic startup/SaaS clichés; some standard discovery call structure.
RubricAssessmentRubric_Novelty_PilotPlan 'Collective memory' framing distinct; 'compounds value' fresh; 'Herself' signature unique; some consulting boilerplate (phases, retainers) but applied to novel domain.
RubricAssessmentRubric_Novelty_ScarletDame Fresh phrasing ('continuation machines', 'narrative source of truth'). Avoids startup clichés. Arrow notation distinctive. Some standard bio language but overall original.
RubricAssessmentRubric_Novelty_YC_Call Distinctive framing ('worldview vs. skills', 'code as compilation output', 'superset framing'); avoids generic startup clichés. Fresh positioning.
RubricAssessmentRubric_Novelty_YCDemoBeats Fresh framing (three beats with aha moments); avoids generic demo clichés; specific tool callouts (Claude Desktop vs Code); some template structure (decision doc format) is intentional
RubricAssessmentRubric_Phrasing_ScarletDame Domain-specific phrasing ('steering vector', 'narrative source of truth', 'continuation machines'). Consistent terminology. Arrow notation for process flows.
RubricAssessmentRubric_Phrasing_YC_Call Strong idiolect ('separated at birth', 'deli email'); domain verbs ('onboards'); simple analogies ('code as compilation output'). Distinctive phrasing.
RubricAssessmentRubric_Phrasing_YCDemoBeats Strong use of established key phrases (senior expert bottleneck, institutional knowledge at cursor); domain-specific verbs (propagates, committed); consistent terminology
RubricAssessmentRubric_Register_CustomerArchitect Professional-conversational register. Direct second-person address, active voice (82%), technical precision without jargon overload. Appropriate for enterprise B2B context.
RubricAssessmentRubric_Register_FounderBetaUser Professional but conversational; technical precision without jargon overload; appropriate for founder-to-founder context; direct quotes preserve speaker voice.
RubricAssessmentRubric_Register_PilotPlan Conversational, direct, second-person throughout; active voice dominant; professional but approachable; fits consulting sales context perfectly.
RubricAssessmentRubric_Register_ScarletDame Professional, direct, conversational. Mix of imperative ('Do NOT list') and declarative. Active voice dominant. Appropriate for internal strategic memo.
RubricAssessmentRubric_Register_YCDemoBeats Professional yet direct; second-person in implementation notes; active voice throughout; appropriate formality for decision doc
RubricAssessmentRubric_Resonance_CustomerArchitect Strong resonance through concrete use cases (QA onboarding, CEO knowledge transfer, code quality), real failure modes (hallucinated behavior, 46/112 bugs), and direct quotes from customer. Test framework as interface analogy lands well. Thr
RubricAssessmentRubric_Resonance_FounderBetaUser 'Gods of Olympus' metaphor is vivid and memorable; Google example provides concrete story; operations persona pain ('75 decisions a day') is specific and relatable; demo reactions show real-time validation.
RubricAssessmentRubric_Resonance_ScarletDame Strong central metaphor (steering vector). Concrete examples (Elephant Collective, Vouch SSoT). Personal story (five-year wait) adds human resonance. Could use more analogies.
RubricAssessmentRubric_Resonance_YC_Call Strong use of founding story ('deli email' as proof of chemistry); analogies (code as compilation output); metaphors (twins as subset). Resonant storytelling.
RubricAssessmentRubric_Resonance_YCDemoBeats Concrete scenario (YC application work) is relatable; aha moments clearly articulated; WITH vs WITHOUT comparison is powerful framing; voice memo detail adds authenticity
RubricAssessmentRubric_StrategicAlignment_PilotPlan Directly articulates positioning thesis (collective memory as org-level brain), pricing model, expansion pathways, delivery guarantee, and data ownership — complete strategic artifact.
RubricAssessmentRubric_StrategicAlignment_ScarletDame Entire document is strategic alignment artifact. Traces core thesis ('narrative as steering vector') through career arc to current product. Explicit YC application framing. Perfect alignment.
RubricAssessmentRubric_StrategicAlignment_YC_Call Directly addresses YC application strategy; founding story aligns with mission/vision; product vision ties to organizational authority narrative. Strong alignment.
RubricAssessmentRubric_Tailoring_ScarletDame Tailored for YC application context. Explicit decisions about what to include/exclude. Addresses YC-specific questions (system hack, most impressive thing). Audience-aware.
RubricAssessmentRubric_Tailoring_YC_Call Tailored to YC audience (references YC questions, founder video, demo video); addresses enterprise buyers and technical reviewers. Good audience fit.
RubricAssessmentRubric_Tailoring_YCDemoBeats Dual-persona approach (Tony/Scarlet) addresses gap_demo_persona_tailoring; YC partner audience implicit; implementation notes tailored for execution team
RubricAssessmentRubricAssess_Accuracy Factual claims (company start date, MVP status, six years of model experience) plausible and internally consistent. Brand name 'aswritten.ai' correctly styled. No citations provided for external claims, but none required for personal narrat
RubricAssessmentRubricAssess_accuracy Consistent brand name; clear claims; no citations needed for strategic document; factual accuracy maintained
RubricAssessmentRubricAssess_Accuracy_1 Precise technical references (n8n, Apache Jena Riot, Outseta, Helicone, Open Router). Accurate product state ('missing piece', 'to do'). No falsifiable claims detected.
RubricAssessmentRubricAssess_Accuracy_ADR001 Specific dates, dollar figures, and entity names (Capitol.ai, VCSA); decision status clearly stated; no citations needed for internal decision record; factual claims verifiable
RubricAssessmentRubricAssess_Accuracy_ADR002 Named entities (Claude Code, Cursor, Windsurf) accurate. Cross-references to ADR-001, ADR-003, ADR-005, ADR-006 present. No unsupported claims; decision rationale explicit.
RubricAssessmentRubricAssess_Accuracy_GTM Specific metrics cited (days→15 min, weeks→days); references prior memories (ADR-002, role-based value props); no unsupported claims; could add external citations for 5.0.
RubricAssessmentRubricAssess_Cadence Varied sentence length: short punchy clauses ('I mean, it's everything, right?') alternate with longer reflective sentences. Anaphora ('every single') and triadic structures create rhythm. Effective pacing for oral delivery.
RubricAssessmentRubricAssess_Cadence_1 Mixed cadence: short declaratives for definitions, long compound sentences for technical explanations. Conversational rhythm with pauses ('you know', 'right?').
RubricAssessmentRubricAssess_Cadence_ADR001 Mix of short declarative statements and longer explanatory sentences; bulleted lists provide rhythm breaks; some complex clauses in rationale section
RubricAssessmentRubricAssess_Cadence_ADR002 Short, punchy sentences in key sections (Hook, Unique Value). Bulleted lists create rhythm and scanability.
RubricAssessmentRubricAssess_Cadence_GTM Short, punchy bullets; varied sentence length; antithesis for impact ('weeks → days'); fast scan, high density.
RubricAssessmentRubricAssess_Cadence_WeeklyCall Short, punchy sentences mixed with longer explanatory clauses. Avg sentence length 18 words. See narr:StyleObs_UntilSomeoneBuys.
RubricAssessmentRubricAssess_Flow Natural progression from company origin → narrative concept → strategic importance → vision. Some transitions implicit rather than explicit. Occasional run-on sentences reduce clarity. Overall coherent but could tighten connective tissue.
RubricAssessmentRubricAssess_Flow_1 Logical progression (terms → problem → product → architecture → roadmap). Some backtracking ('I forgot', 'let me step back'). Natural for spoken walkthrough.
RubricAssessmentRubricAssess_Flow_ADR001 Strong structural flow via ADR template; logical progression from context → decision → rationale → consequences; numbered lists aid coherence; minimal formulaic transitions
RubricAssessmentRubricAssess_Flow_ADR002 Natural progression: Context → Decision → Use Cases → Unique Value → Target Customer → Alternatives → Consequences. Transitions clear; ADR template structure aids coherence.
RubricAssessmentRubricAssess_Flow_GTM Clear progression: target → vertical → champion → sales motion → pilot → expansion; transitions are logical; some sections could be tighter.
RubricAssessmentRubricAssess_Flow_WeeklyCall Natural progression from founder psychology → sales strategy → product insights → action items. Transitions are implicit (topic headers) rather than explicit connectives.
RubricAssessmentRubricAssess_Novelty Original phrasing ('narrative architecture,' 'steering vector,' 'resonant body') and fresh metaphors. Avoids most startup clichés. Some generic constructions ('product market fit,' 'first principles') but used purposefully. High type-token
RubricAssessmentRubricAssess_novelty Fresh framing (git for narrative); distinctive brand voice; some stock phrases ('source of truth'); overall original positioning
RubricAssessmentRubricAssess_Novelty_1 Novel framing: 'narrative source of truth', 'individuation pipeline', 'evolved variant', 'collective memory marketplace'. Distinct from generic AI memory/RAG positioning.
RubricAssessmentRubricAssess_Novelty_ADR001 Standard ADR format (expected); business language conventional; 'collaborative worldview layer' phrase somewhat novel; overall appropriate level of originality for decision record
RubricAssessmentRubricAssess_Novelty_ADR002 'Worldview layer' positioning is fresh vs. generic 'AI memory'. '[amount] mistake' quantifies pain memorably. Avoids boilerplate; specific tool names (Claude Code, Cursor) add authenticity.
RubricAssessmentRubricAssess_Novelty_GTM Fresh framing ('state secrets', 'IP as code'); some standard GTM structure (pilot→expand); avoids clichés; distinct voice.
RubricAssessmentRubricAssess_Novelty_WeeklyCall Fresh framing ('hub-and-spoke' vs. shared-chat model; 'topical solution' vs. panacea). Some domain-standard phrases ('beachhead', 'validation gap'). Type-token ratio suggests moderate lexical diversity.
RubricAssessmentRubricAssess_Phrasing Distinctive phrasing: 'narrative architecture,' 'steering vector,' 'resonant body.' Idiolect evident in metaphor clusters and domain-specific terminology. Some repetition ('right?') signals conversational authenticity.
RubricAssessmentRubricAssess_Phrasing_GTM Signature phrases ('state secrets to AI', 'IP as code', 'network effects') are memorable, repeatable, and aligned with strategic narrative.
RubricAssessmentRubricAssess_Register Conversational yet visionary; first-person narrative with active voice; mixes colloquial ('right?') and elevated ('precipice of spirituality') registers effectively. Strong fit for founder/visionary context.
RubricAssessmentRubricAssess_register Direct, conversational register; second-person address; active voice; technical precision without jargon
RubricAssessmentRubricAssess_Register_1 Conversational register with technical depth; second-person address; active voice dominant. Fillers reduce formality.
RubricAssessmentRubricAssess_Register_ADR001 Formal business register appropriate for ADR; first-person plural establishes ownership; technical precision without jargon overload
RubricAssessmentRubricAssess_Register_ADR002 Professional register with conversational clarity; active voice dominates; concise phrasing.
RubricAssessmentRubricAssess_Resonance_1 Concrete examples (Crooked Media demo, NPR collective memory). Analogies to software development rigor. Could benefit from more customer stories.
RubricAssessmentRubricAssess_Resonance_ADR001 Platform vs. infrastructure antithesis resonates; moat metaphor familiar to business audience; lacks narrative story/case study but appropriate for ADR format
RubricAssessmentRubricAssess_Resonance_GTM Strong analogies ('IP as code', 'state secrets'), quantified metrics (weeks→days), vivid problem framing; could add customer story for 5.0.
RubricAssessmentRubricAssess_Resonance_WeeklyCall Strong use of metaphor ('roller coaster'), antithesis ('amazing product and not a product'), and analogies (shared agent memory). See narr:StyleObs_RollerCoaster, narr:StyleObs_AmazingProductNotProduct.
RubricAssessmentRubricAssess_StrategicAlignment Deeply aligned with core strategy: narrative as steering mechanism, first principles (safety/alignment/ethics), leverage, and collaborative vision. Articulates mission, vision, and positioning thesis clearly. Exemplary strategic coherence.
RubricAssessmentRubricAssess_StrategicAlignment_1 Explicit mission, positioning, moat, and roadmap. Clear narrative anchor (Git-native AI memory). Strong alignment across Opportunity/Strategy/Product/Architecture.
RubricAssessmentRubricAssess_StrategicAlignment_ADR002 Directly articulates positioning thesis, moat, focus/sequencing, and target customer. Explicitly ties to narrative architecture (ADR-001, ADR-003, ADR-005).
RubricAssessmentRubricAssess_StrategicAlignment_GTM Directly operationalizes strategic narrative (IP crisis, collaborative worldview, advisor-led GTM); every section ties back to positioning thesis.
RubricAssessmentRubricAssess_StrategicAlignment_WeeklyCall Tight alignment with positioning thesis (knowledge extraction beachhead), segmentation (architecture teams), and timing thesis (two-year window). See narr:Narrative_KnowledgeExtractionBeachhead, narr:Narrative_TwoYearWindow.
RubricAssessmentRubricAssess_Strategy Directly aligned with Architecture domain; explicit rationale for each decision; clear separation of concerns.
RubricAssessmentRubricAssess_Strategy_ADR001 Exemplary strategic alignment: directly addresses positioning thesis, moat/leverage, focus/sequencing, and risk/guardrails; decision rationale maps to strategy framework
RubricAssessmentRubricAssess_Tailoring Audience implied (investors, partners, users) but not explicitly segmented. Language accessible to technical and non-technical audiences. Could strengthen with persona-specific examples or direct address.
RubricAssessmentRubricAssess_tailoring Addresses multiple personas (developer, PM, CEO, sales); three pillars map to audiences; could deepen persona-specific examples
RubricAssessmentRubricAssess_Tailoring_1 Tailored to technical audience (Git, RDF, SPARQL, Docker Compose). Assumes programming literacy. Conversational tone suggests internal team context.
RubricAssessmentRubricAssess_Tailoring_ADR001 Tailored for internal technical/business audience; assumes context familiarity; appropriate detail level for decision record; would need translation for external stakeholders
RubricAssessmentRubricAssess_Tailoring_ADR002 Tailored to developer persona (Claude Code, Cursor, Windsurf references). Hook speaks directly to pain point. Expansion path acknowledges executive/content audiences.
RubricAssessmentRubricAssess_Tailoring_GTM Highly tailored: specific criteria for advisors, distinct champion profiles (CEO/Strategy/CTO), vertical-specific pain points (AI/ML, dev tools, SaaS).
RubricAssessmentRubricAssess_Tailoring_WeeklyCall Tailored to founder/sales context; references specific personas (an enterprise prospect, CIO Forum attendees, architecture teams). Could strengthen with more explicit persona JTBD framing.
RubricAssessmentRubricAssessment Technical-colloquial blend is well-calibrated for founder-AI coworker context. Active voice dominates. Concise declaratives. Minor: some long compound sentences in solution sections.
RubricAssessmentRubricAssessment_Flow Logical progression: Context → Key Finding → Current Broken State → Anatomy → Bloat → Architecture Decisions → TriG Architecture → Immediate Win → Execution Plan → Attribution. Each section builds on the prior. Tables and code blocks are we
RubricAssessmentRubricAssessment_Novelty 'Conviction as source count signal', 'individuation', 'star topology' framing, and the TriG transactional model are fresh framings of known RDF patterns. Low cliché density. The 'lost opportunity' phrase is distinctive. Some standard techni
RubricAssessmentRubricAssessment_Phrasing Domain-specific idiolect ('conviction', 'individuation', 'leaf nodes', 'star topology') used precisely and consistently. 'a lost opportunity' is a distinctive evaluative phrase. Stock phrases from the ontology vocabulary are correctly appli
RubricAssessmentRubricAssessment_Register Professional/analytical register appropriate for competitive analysis; active voice dominant; concise phrasing
RubricAssessmentRubricAssessment_Resonance The 'lost opportunity' framing and the house-building phase metaphor (implicit in the ontology's Phase_* concepts) add resonance. The TriG Turtle code block grounds the abstract in concrete syntax. Lacks narrative story arc or customer-faci
RubricAssessmentRubricAssessment_StrategicAlignment Directly advances core positioning thesis; connects market signals to narrative anchor; validates timing thesis and differentiation
RubricAssessmentRubricAssessment_Tailoring Clearly written for a technical audience familiar with RDF, SKOS, and graph concepts. No audience bridging for non-technical stakeholders. Attribution ('Scarlet Dame and Claude') signals internal document. Would benefit from an executive su
StyleMetrics (136)
StyleMetricsMetrics_1 High active voice ratio; moderate jargon density appropriate for technical audience.
StyleMetricsMetrics_AgentSimulationResearch_Full Research memo register with embedded verbatim blockquotes from a collaborator's talk transcript. Mixed register: structured markdown with H2 section headers and bold numbered parallels (memo genre) vs. colloquial first-person in blockquoted
StyleMetricsMetrics_BusinessShape_20260330
StyleMetricsMetrics_ChatbotPricingSignal_20260501 Strategic pricing signal memo. Mixed register: analytical prose with embedded verbatim blockquote from the customer's chief architect. Antithesis-heavy structure (upper end vs. negotiation floor, design now vs. track first, tier vs. custom
StyleMetricsMetrics_cm_addendum High jargon density appropriate for technical architecture document; active voice dominant; varied vocabulary
StyleMetricsMetrics_collective_memory_arch High jargon density appropriate for technical architecture document; strong active voice usage
StyleMetricsMetrics_ConsultingFramingDecision_20260610 Strategic decision memo register with embedded verbatim blockquote from Scarlet. Mixed register: analytical editorial prose (authoritative, structured) vs. first-person conversational in blockquoted speech (lowercase, run-on). Antithesis-he
StyleMetricsMetrics_Correction_Tranche2_20260612 Correction memo register. Three-section antithesis structure (wrong / correct / scope). Short punchy declaratives at key retraction moments ('There is no [amount] Tranche 2.'). Active voice dominant. Jargon density moderate — financial term
StyleMetricsMetrics_CustomerCTOCodeReview_20260507 Strategic memo with embedded verbatim blockquotes from four speakers. Mixed register: analytical editorial prose (authoritative, structured) vs. conversational first-person in quoted speech (run-on, filler-heavy). the customer's CTO's quote
StyleMetricsMetrics_CustomerEngManagerVision Analytical memo with heavy verbatim blockquotes. High active voice in editorial framing.
StyleMetricsMetrics_Demo1Pipeline High active voice ratio; technical jargon (SPARQL, n8n, BYOK, SHA) is domain-appropriate and consistently expanded on first use. Sentence length varies: short punchy imperatives alternate with longer explanatory sentences.
StyleMetricsMetrics_DesignCollaboratorReview
StyleMetricsMetrics_EngagementShape_20260610 Strategic session memo register. Mixed: analytical editorial prose (authoritative, structured) vs. first-person conversational in blockquoted speech (lowercase, run-on). Antithesis-heavy structure throughout. Active voice dominant in editor
StyleMetricsMetrics_EnterpriseCustomerPilotReveals_20260516 Analytical memo register with embedded blockquoted verbatim examples and bold rhetorical anchors. Antithesis-heavy structure (dashboard vs. agent interface, passive store vs. active surface, full-loop vs. reader-grounder). Active voice domi
StyleMetricsMetrics_EnterpriseCustomerWeekOneOutcomes Analytical metrics report register. Structured markdown with H2 section headers and bold-number bullet lists. Active voice dominant in editorial prose. Jargon density moderate — product terms (MCP sessions, introspect, cite, perspective loa
StyleMetricsMetrics_EnterpriseShapePrimaryFocus_20260421 Strategic memo register with embedded verbatim blockquotes. Antithesis-heavy structure (additive vs. incremental, primary vs. secondary, tentative vs. decision). Active voice dominant in editorial prose. Timestamp citation convention throug
StyleMetricsMetrics_FounderBetaUserCall Mix of short declarative statements and longer analytical passages; high use of direct quotes; conversational register with technical precision.
StyleMetricsMetrics_GitHubTierDesign_Updated Supplementary metrics record for this extraction pass. Primary node is narr:Metrics_GitHubTierDesign.
StyleMetricsMetrics_GoalDirectedAgents Strategic design session register. High use of structural metaphors (altitudes, wedges). Mixes formal data modeling with colloquial design breakthroughs.
StyleMetricsMetrics_InterviewOnrampShift Analytical shift-log register. Uses contrastive pairs to define strategic movement. High active voice in the editorial framing.
StyleMetricsMetrics_InvestorPipeline_20260517 Strategic pipeline memo with heavy list structure.
StyleMetricsMetrics_NicheEmerges_20260424 Strategic session memo with embedded verbatim blockquotes. Mixed register: structured markdown with H2 section headers (memo genre) vs. conversational first-person in blockquoted speech. Active voice dominant in editorial prose; blockquoted
StyleMetricsMetrics_OutsideTechBetaUserChannelStrategy_20260423 Strategic correction memo format. Mixed register: analytical prose in section headers and argument sections; verbatim a beta user from outside tech blockquotes in first person (reconstructed). Antithesis-heavy structure (SMS works / SMS fli
StyleMetricsMetrics_PerspectiveTransforms_20260414 Design session notes format. Mixed register: structured markdown with H4 section headers and bold numbered conclusions; embedded Scarlet blockquotes in first person. Antithesis-heavy structure throughout (transform vs. query, actor vs. role
StyleMetricsMetrics_PractitionerBetaUserBarber_20260430 Strategic session memo with embedded verbatim blockquotes from Scarlet. Mixed register: structured markdown with H2 section headers (memo genre) vs. conversational first-person in blockquoted speech. Active voice dominant in editorial prose
StyleMetricsMetrics_ReachabilityPromise
StyleMetricsMetrics_ResearchBranchMethodology_20260408 Strategic meta-memo register. Mixed: analytical prose in section headers and argument sections; verbatim Scarlet blockquotes in first person (run-on, stream-of-consciousness). Antithesis-heavy structure (conviction vs. temporality, boring-a
StyleMetricsMetrics_ScarletVoiceRule_20260612 Short decision-memo register. Mixed: editorial prose in rule-statement sections; verbatim blockquote in Scarlet's first-person voice (run-on, emphatic repetition). Active voice dominant. Jargon density very low — no product terms. Sentence
StyleMetricsMetrics_SessionReflection_20260416 Retrospective reflection format. Mixed register: structured markdown with H2/H3 section headers (analytical memo genre) and embedded verbatim Scarlet quotes in first person (personal voice). Antithesis-heavy structure (product-out vs. insta
StyleMetricsMetrics_TeamPlanICPOrgTier_20260517 Strategic session memo with embedded verbatim blockquotes from Scarlet. Mixed register: structured markdown with H2/H3 section headers and markdown tables (memo genre) vs. conversational first-person in blockquoted speech. Antithesis-heavy
StyleMetricsMetrics_TemporalityValidation_20260501 Strategic session memo with extended verbatim blockquotes from two speakers. Mixed register: structured markdown with H2 section headers (memo genre) vs. colloquial first-person in blockquoted speech. the customer's chief architect's quoted
StyleMetricsMetrics_TonyMaleyCheckin_20260421 Call transcript register with heavy embedded verbatim quotes from two speakers. Tony's quoted speech is short, declarative, numbers-first, filler-heavy ('Right. Right. Yeah, yeah, 100, 100%'). Scarlet's quoted speech is longer and more anal
StyleMetricsMetrics_TonyScarletCall_20260331 Live call transcript register. Both speakers use conversational, filler-heavy speech with run-on sentences. Scarlet's quoted speech is longer and more analytical (~28 words avg); Tony's is shorter and more declarative (~14 words avg). Activ
StyleMetricsMetrics_TranscriptBootstrapping_20260416 Voice memo / post-walk reflection format. Mixed register: structured markdown headers with embedded first-person blockquotes. Active voice dominant. Short punchy declaratives at conviction peaks ('I do.', 'This should move from waiting to b
StyleMetricsMetrics_V0Scope Strategic scoping register. High active voice in decision statements. Jargon includes technical (n8n, Twilio, webhook) and product-specific (share primitive, mode switch, sibling repo) terms.
StyleMetricsMetrics_VouchEmailsSalesMotion Strategic session memo register with embedded verbatim blockquotes from the original Vouch emails. Mixed register: analytical prose in framing sections (memo genre) vs. first-person narrative in the deli story blockquotes (personal/epistola
StyleMetricsMetrics_YC Metrics computed from full transcript. High active voice ratio, moderate jargon density (domain-appropriate), good lexical diversity.
StyleMetricsMetrics_YC_Call High active voice ratio; conversational register; mix of short punchy and longer explanatory sentences.
StyleMetricsStyleMetrics_AgentSimulationResearch Analytical research register with high density of cross-references and technical parallels.
StyleMetricsStyleMetrics_AIOrg Metrics computed on full text. Moderate sentence length; high active voice; controlled jargon density; good lexical diversity (TTR).
StyleMetricsStyleMetrics_BetaOnboarding High active voice ratio; moderate sentence length with deliberate short-sentence punctuation for emphasis; jargon (RDF, MCP, n8n, CLAUDE.md) is domain-specific and intentional
StyleMetricsStyleMetrics_BetaStrategyReframe Mixed register: analytical prose with embedded informal direct-speech quotes. Heavy use of markdown tables and code blocks for structured data. Aphoristic one-sentence paragraphs used as rhetorical anchors. Second-person implicit (instructi
StyleMetricsStyleMetrics_BrainDump_Mar2026 High active voice ratio consistent with direct, first-person technical writing. Jargon density moderate — technical terms (OOM, MCP, RDF, TX) used without definition, implying expert audience. Average sentence length elevated by enumerated
StyleMetricsStyleMetrics_BrandDesignLanguage Authoritative, editorial register with technical/code-adjacent terminology.
StyleMetricsStyleMetrics_BulkIngestionCorrection_20260521 Correction memo register with embedded verbatim blockquotes from Scarlet. Antithesis-heavy structure (overplayed vs. explicit anti-pattern; deliberate vs. accidental; what is built vs. what is not yet built). Active voice dominant in editor
StyleMetricsStyleMetrics_CalibratedGrowth_20260527 Strategic decision memo with embedded verbatim blockquotes from Scarlet. Mixed register: structured markdown with H2 section headers and bold numbered commitments (memo genre) vs. conversational first-person in blockquoted speech. Antithesi
StyleMetricsStyleMetrics_ChannelStrategyCorrection Register: strategic-operational, first-person-founder voice mediated through third-person documentation. Moderate sentence length with deliberate variation — short declaratives for corrections ('This is incorrect.'), longer compound sentenc
StyleMetricsStyleMetrics_CiteConstellationReview Register: analytical-technical with deliberate register shifts into first-person testimony (block quotes). Sentences are long and complex in analytical sections; short and declarative in diagnostic sections. Jargon is controlled and domain-
StyleMetricsStyleMetrics_collective_memory_cvp High active voice ratio; short sentences; concise phrasing; punchy cadence
StyleMetricsStyleMetrics_ContextServerFiveQuestions Strategic session memo register with embedded verbatim blockquotes from Scarlet. Mixed register: analytical editorial prose (authoritative, structured) vs. first-person conversational in blockquoted speech. Antithesis-heavy structure (capab
StyleMetricsStyleMetrics_CowenEV
StyleMetricsStyleMetrics_CustomerArchitectSponsorNotChampion Call transcript register with heavy embedded verbatim quotes from two speakers. Tony's quoted speech is short, directive, filler-heavy ('Right. Right. Blah, blah, blah.') averaging ~14 words. Scarlet's speech is more analytical (~22 words a
StyleMetricsStyleMetrics_DavidNolenDemo_20260625 Call-notes format with heavy embedded verbatim block-quotes from two speakers. David's quoted speech is expansive and probing (~28 words avg per sentence); Scarlet's editorial prose is shorter and declarative (~18 words avg). Active voice d
StyleMetricsStyleMetrics_DavidNolenDemoCall Call-notes format with heavy embedded verbatim blockquotes. Editorial prose is analytical and declarative (~22 words avg sentence); David's quoted speech is expansive and exploratory (~30 words avg). Active voice dominant in editorial secti
StyleMetricsStyleMetrics_DeckEditing Flesch-Kincaid Grade Level: ~10 (college-level but accessible)
StyleMetricsStyleMetrics_Demo1 Estimated metrics: avg sentence length ~18 words (mix of short aphoristic and long technical sentences); active voice ratio ~82%; jargon density ~18% (SPARQL, webhook, base64, additionalParameters, executeWorkflow, etc.). Register: technica
StyleMetricsStyleMetrics_DesignCollaboratorCall_20260629 Mixed register: raw first-person stream-of-consciousness in blockquoted speech (run-on, filler-heavy, ~35 words avg per sentence in quoted sections) vs. editorial framing prose (structured, shorter, ~18 words avg). Active voice dominant thr
StyleMetricsStyleMetrics_DesignCollaboratorTokens Session-memo register with embedded verbatim blockquotes. Editorial prose is authoritative and structured; blockquoted speech is conversational and first-person. Active voice dominant in editorial sections. Jargon density moderate — product
StyleMetricsStyleMetrics_DesignSession_20260408 Analytical design session register with high use of conceptual labels and direct quotes.
StyleMetricsStyleMetrics_DesignSession_Full Design session transcript with structured markdown overlay. Mixed register: analytical prose in section headers and rationale sections; colloquial first-person in Scarlet's quoted speech. Antithesis-heavy structure (warm-start vs cold-start
StyleMetricsStyleMetrics_DigestRecognitionAid Philosophical-analytical register with embedded verbatim blockquotes from Scarlet. Mixed: structured markdown with H2 section headers (memo genre) vs. first-person blockquoted speech. Antithesis-heavy structure (recognition vs. creation, in
StyleMetricsStyleMetrics_DiscoveryConsultant_20260630 Call notes format with heavy embedded verbatim quotes from two speakers. a consultant who ran enterprise discovery engagements for fifteen years's quoted speech is expansive and philosophical (~30 words avg per sentence); Scarlet's is more
StyleMetricsStyleMetrics_DiscoveryConsultantCall High jargon density (SPARQL, RDF, ABAC, IRI, Datomic, XTDB, ARQ). Active voice dominant in direct quotes; passive voice appears in analytical framing sections. Long sentences in block quotes due to spoken-language transcription patterns.
StyleMetricsStyleMetrics_DiscoveryConsultantCompanion Mixed register: editorial memo prose (authoritative, structured) framing verbatim blockquoted dialogue. a consultant who ran enterprise discovery engagements for fifteen years's voice: short declarative sentences with Boolean verdicts, engi
StyleMetricsStyleMetrics_DiscoveryToolWedge_20260702 Strategic memo register. Mixed structure: H2-sectioned analytical prose with embedded verbatim block-quotes. Active voice dominant in editorial sections; passive voice in status/caveat framing ('is recorded here', 'is deferred'). Jargon den
StyleMetricsStyleMetrics_DistillationPerspectiveCreation Design decision memo format. Mixed register: structured markdown with bold headers and bullet lists for decisions; embedded verbatim Scarlet quotes in first person. Antithesis-heavy (before → after pairs). Active voice dominant. Jargon dens
StyleMetricsStyleMetrics_EnterpriseCustomerKickoff Mixed register: analytical prose sections use authoritative active voice; verbatim quotes from speakers are conversational and fragmented (run-ons, false starts, hedging). Overall active voice ratio elevated by the editorial framing. Jargon
StyleMetricsStyleMetrics_EnterpriseCustomerKickoffPrep Strategic prep memo register. High use of verbatim blockquotes to validate methodology. Antithesis used to frame design tensions (power users vs AI-resistant, identity vs weeds).
StyleMetricsStyleMetrics_EnterpriseCustomerPattern_20260525 Strategic memo register; authoritative and pattern-seeking.
StyleMetricsStyleMetrics_EnterpriseCustomerPilot Analytical register with high active voice. Moderate jargon density driven by domain terms (MCP, ARB, p95, streamgraph, composition artifact). Sentence length varies: short punchy labels alternate with dense multi-clause analytical sentence
StyleMetricsStyleMetrics_EnterpriseCustomerWeekOne Technical post-mortem register. High active voice ratio; sentences are long due to embedded parenthetical data points and inline code. Jargon density elevated by system-specific identifiers (outseta-id, JSONB, webhook mode, JWT). Tone is au
StyleMetricsStyleMetrics_EVFlockingApplication Mixed register: strategic session memo (analytical, structured) framing a verbatim founder essay (first-person, lyrical). The flocking piece itself is active-voice dominant, medium sentence length with deliberate variation — short declarati
StyleMetricsStyleMetrics_ForgettingBottleneck Strategic memo register with three embedded verbatim blockquotes from an engineering manager at the customer. Mixed register: analytical editorial prose (authoritative, structured H2 sections) vs. colloquial first-person in blockquoted spee
StyleMetricsStyleMetrics_founder_reflection Metrics computed from input: 80 sentences, avg 42.8 words/sentence; 78% active voice; high lexical diversity (TTR 0.62); moderate rhythm variance; low jargon density (8% domain-specific terms).
StyleMetricsStyleMetrics_FramingLayerMetaPattern Meta-analytical register. High reliance on verbatim blockquotes to preserve structural texture. Antithesis used to contrast models (failure vs reframe).
StyleMetricsStyleMetrics_FrontSubstrate
StyleMetricsStyleMetrics_FundraisingBreakpoints high — bullet-driven structure, minimal subordinate clauses in formal sections; blockquotes use run-on stream-of-consciousness for contrast
StyleMetricsStyleMetrics_GTM High active-voice ratio, moderate jargon density (technical terms defined in context), short-to-medium sentence length for clarity.
StyleMetricsStyleMetrics_InfraAdvisor Metrics estimated from transcript; conversational register with moderate technical density; active voice dominant in recommendations
StyleMetricsStyleMetrics_IngestionProtocol Register: technical-strategic hybrid. Voice is direct and authoritative with embedded verbatim quotes from session. High jargon density (MCP, OAuth, JSONL, ZDR, ICP, webhook) signals practitioner audience. Active voice dominates; passive us
StyleMetricsStyleMetrics_June11Crystallization Strategic decision memo with embedded verbatim blockquotes from Scarlet. Mixed register: structured markdown with H2 section headers (memo genre) vs. first-person stream-of-consciousness in blockquoted speech. Antithesis-heavy structure (co
StyleMetricsStyleMetrics_LandingPageScrollDriven_20260704 Voice memo transcription register with structured markdown overlay. Mixed register: raw first-person exploratory speech in blockquotes (run-on, self-interrupting, clause-heavy) vs. editorial prose in section headers and analysis. Active voi
StyleMetricsStyleMetrics_LensArchReview Technical design document. High active voice ratio. Jargon is domain-specific (compile, introspect, scope, lens, TriG, SPARQL) and intentional. Sentence length varies: short declarative corrections alternate with longer analytical passages.
StyleMetricsStyleMetrics_LensArchRevision high — numbered lists, single-action steps, no subordinate padding
StyleMetricsStyleMetrics_MCP_Auth High active voice ratio; technical jargon (Datomic, JWT, PAT, MCP, n8n) is domain-appropriate and consistently used. Average sentence length moderate; punctuated by short declarative sentences for emphasis. Three direct quotes in blockquote
StyleMetricsStyleMetrics_Moonshot_20260701 Voice memo / morning pages register. Mixed: raw first-person stream-of-consciousness in blockquoted speech (run-on, filler-heavy, ~35 words avg per sentence) vs. editorial framing prose (structured, shorter, ~18 words avg). Active voice dom
StyleMetricsStyleMetrics_MVPScoping Direct, decisive, and strategic register with high active voice ratio.
StyleMetricsStyleMetrics_MVPScopingSession Hybrid register: first-person voice memo transcription embedded in structured decision document. High active voice ratio. Moderate jargon density (n8n, MCP, TX, cron, JVM, OOM, Postgres, Claude Code). Sentence length varies widely — short p
StyleMetricsStyleMetrics_OnboardingMethodology_20260501 Strategic session memo with embedded verbatim blockquotes from Scarlet. Mixed register: structured markdown with H2/H3 section headers (memo genre) vs. conversational first-person in blockquoted speech. Antithesis-heavy structure (direct in
StyleMetricsStyleMetrics_OpinionatedPromise Strategic session memo register with embedded verbatim blockquotes. Mixed register: analytical editorial prose (authoritative, structured) with first-person conversational in blockquoted speech. Antithesis-heavy structure (description vs. d
StyleMetricsStyleMetrics_PerspectiveOnSpec Reflective strategic memo with heavy use of blockquotes. High active voice in editorial framing.
StyleMetricsStyleMetrics_PerspectiveResponseOrdering Technical-architectural register. Active voice dominant in editorial sections; passive used sparingly for design decisions ('was retired', 'is now ordered'). Jargon density moderate — engineering terms (byte-stable, cache_control, cache-hos
StyleMetricsStyleMetrics_PerspectiveShares Technical architectural documentation style with high active voice and specific system jargon.
StyleMetricsStyleMetrics_PerspectiveVsImpl Register: instructional-authoritative. Voice: primarily active, direct. Mix of short declarative sentences (principle statements) and longer explanatory sentences (failure mode analysis). Technical jargon (hex codes, SCSS, tokens.json, carg
StyleMetricsStyleMetrics_PitchArtifacts_20260430
StyleMetricsStyleMetrics_PitchReframeReview_20260521 Strategic resolution memo with extended verbatim blockquote from Scarlet. Mixed register: short declarative editorial prose framing the resolution, followed by a long run-on first-person blockquote. Active voice dominant in editorial sectio
StyleMetricsStyleMetrics_PWG Active voice dominant; jargon is domain-specific (worldview graph, compile, agent-mediated) rather than filler jargon. Average sentence length moderate with high variance — long analytical sentences punctuated by short declarative pivots. T
StyleMetricsStyleMetrics_ReadLayerNavigation_20260616 Strategic decision memo register with embedded verbatim blockquotes from Scarlet. Mixed register: structured markdown with bold principle headers (memo genre) vs. lowercase first-person in blockquoted speech. Antithesis-heavy structure (nav
StyleMetricsStyleMetrics_RichTierSteeringWeight Design-decision memo register. Active voice dominant in editorial sections. Jargon density moderate — product-internal terms (rich tier, steering weight, in-degree, area index, conviction-carrying) used without definition, implying internal
StyleMetricsStyleMetrics_Sample High active voice, moderate jargon (domain-appropriate), strong conciseness, good lexical diversity.
StyleMetricsStyleMetrics_Sample1 High — tables, arrows, and antithesis compress multi-step logic into scannable units
StyleMetricsStyleMetrics_ScarletDame Estimated metrics: short average sentence length, high active voice ratio, high lexical diversity.
StyleMetricsStyleMetrics_ScarletMVPScoping Voice memo transcription register: conversational but directive. High active voice ratio. Moderate jargon density (n8n, MCP, TX, cron, Claude Code, Postgres, Timbre, Outseta). Sentence length varies widely — short declarative decisions inte
StyleMetricsStyleMetrics_ScarletRules The memory itself follows its own rules: short declarative sentences, active voice throughout, minimal jargon. Register is direct and instructional. Scarlet's quoted speech is notably terse and imperative. The worked example demonstrates th
StyleMetricsStyleMetrics_SecondWaveBridge Strategic and thesis-driven register. High use of metaphors (waves, bridges, stakes). Direct and authoritative tone.
StyleMetricsStyleMetrics_SelfServeProductSurface Mixed register: verbatim first-person blockquotes (conversational, exploratory) interleaved with editorial framing prose (authoritative, declarative). Active voice dominant in editorial sections. Jargon density moderate — product-internal t
StyleMetricsStyleMetrics_SIC_1 High jargon density reflects technical domain. Moderate conciseness due to conversational fillers.
StyleMetricsStyleMetrics_SII_AGIReclaim_Numeric Mixed register: personal essay (scanner section) with long subordinate-clause chains elevates average sentence length. Editorial framing sections are shorter and more declarative. Active voice dominant in editorial; passive in scanner body-
StyleMetricsStyleMetrics_SII_AGIReclaim_Updated Mixed register: analytical editorial prose framing a personal essay. Long complex sentences in scanner section (subordinate clauses, em-dash pivots) alternate with short declarative blades. Active voice dominant in editorial sections; passi
StyleMetricsStyleMetrics_SteeringDecisionProcess Strategic pitch-reframe memo with embedded verbatim Scarlet quotes. Mixed register: tight analytical prose in editorial sections (bold headers, genus/species framing) vs. run-on first-person in blockquoted speech. Antithesis-heavy structure
StyleMetricsStyleMetrics_TaxonomyBlockers Estimated from document body. Average sentence length ~18 words (mix of short punchy conclusions ~8w and longer explanatory sentences ~28w). Active voice ratio ~82%. Jargon density moderate-high (domain-specific terms: introspect, worldview
StyleMetricsStyleMetrics_ThreeUseCasesLifecycle Mixed register: editorial synthesis prose (authoritative, structured) with embedded verbatim blockquotes from Scarlet (conversational, first-person). Active voice dominant in editorial sections. Jargon density low-moderate — product terms (
StyleMetricsStyleMetrics_TonyScarletAdvisory Advisory staffing and named perspective sections of the Tony-Scarlet call. Conversational transcript register. Scarlet's quoted speech averages ~30 words per sentence with complex clause nesting; Tony's averages ~14 words with short declara
StyleMetricsStyleMetrics_TonyWeekly Meeting notes format with embedded verbatim quotes. Dominant register: informal spoken-capture with product/technical vocabulary. High active voice ratio. Moderate jargon density (Git/PR metaphors, AI tooling terms). Short punchy declarativ
StyleMetricsStyleMetrics_ToolNaming_20260330 Register: semi-formal design notes with direct dialogue. Voice is predominantly active and first-person attributed. Jargon is product-specific (compile, perspective, n8n, MCP, proxy) but intentionally contrasted with plain-language alternat
StyleMetricsStyleMetrics_Tranche1BridgeRaise Strategic session memo with embedded verbatim blockquotes from Scarlet. Mixed register: structured markdown with H2 section headers and markdown tables (memo genre) vs. conversational first-person in blockquoted speech. Antithesis-heavy str
StyleMetricsStyleMetrics_TranscriptOnramp_20260705 Strategic session memo register. Analytical editorial prose with embedded verbatim block-quoted directives from Scarlet. Antithesis-heavy structure (demo vs. product, per-transcript vs. hub-collection, reading vs. receipts). Active voice do
StyleMetricsStyleMetrics_ValidationRigor_20260616 Structured retrospective memo with three H2 sections (What happened, The principle, How to hold it). Antithesis-heavy throughout (smokes vs. battery, mechanism vs. quality, appearance vs. control). Active voice dominant; imperative mood in
StyleMetricsStyleMetrics_VisionAcquisitionStrategy_20260517 Strategic session memo with extended verbatim blockquote from Scarlet plus structured analysis sections. Mixed register: first-person stream-of-consciousness in blockquote (~35 words avg per sentence) vs. analytical editorial prose (~16 wor
StyleMetricsStyleMetrics_WorksFromWhatWasSaid Strategic session memo register with embedded verbatim blockquote from Scarlet. Mixed register: analytical editorial prose (authoritative, structured) vs. first-person conversational in blockquoted speech. Antithesis-heavy structure (what w
StyleMetricsStyleMetrics_YCDemoBeats Measured across prose sections; bullet lists excluded from sentence length calculation
Sample (113)
SampleSample_1 The Honest Pitch — Manifesto Draft
SampleSample_20260107_onboarding_impossibility Primary Hook: The Onboarding Impossibility & Four Lock-In Costs
SampleSample_ADR001 ADR-001: aswritten.ai as Complete Product, Not Infrastructure Layer
SampleSample_ADR002 ADR-002: Three Pillars - Coding, Executive, Content
SampleSample_AgentSimulationResearch_20260429 Lifting from agent-simulation research: memory architecture, multi-role coordination, and what doesn't transfer
SampleSample_BetaOnboarding Beta Onboarding Call Process
SampleSample_BetaStrategyReframe Beta Strategy Reframe: Full-Service Consulting as Product Validation
SampleSample_BoundedDomainFidelity_20260319 Bounded-Domain Fidelity, Collaboration Taxonomy, and Document Ecosystem (2026-03-19)
SampleSample_BrainDump_Mar2026 March 2026 Brain Dump: Product Priorities and Architectural Insights
SampleSample_BrandDesignLanguage_20260414 Brand Design Language: Thinking Before a design collaborator's Style Guide
SampleSample_BulkIngestionCorrection_20260521 Bulk transcript ingestion is a valid onboarding path — 'direct ingestion is an anti-pattern' is a graph error
SampleSample_BusinessShape_20260330 Business Shape, Funding Reframe, and the Practice Model
SampleSample_ChannelStrategyCorrection_20260428 Channel Strategy Correction — Scarlet Dame / Claude session, April 28 2026
SampleSample_cm_addendum_20250130 collective memory Data Model & API Architecture Discussion - ADDENDUM
SampleSample_collective_memory_arch collective memory Data Model & API Architecture Discussion
SampleSample_collective_memory_cvp Shared Worldview: collective memory's Core Value Proposition
SampleSample_ConsultingFramingDecision_20260610 The framing decision: consulting is me using the product
SampleSample_ContextServerFiveQuestions_20260610 The value-proposition logic: the context server and the five questions
SampleSample_CustomerArchitectSponsorNotChampion_20260421 the customer's chief architect as Sponsor Not Champion: POC/MSA Parallelism Playbook (April 21, 2026)
SampleSample_CustomerCTOCodeReview_20260507 the customer's CTO Reframes Memory Ingestion as Code Review — May 7, 2026
SampleSample_CustomerEngManagerVision_20260519 an engineering manager at the customer — ARB Process Engagement and Deployment Vision — May 19, 2026
SampleSample_DavidNolenDemoCall_20260625 David Nolen Demo Call — What Landed, the Audience Question, and the Demo Work It Forced
SampleSample_DeckEditing_20260108 Deck Editing Decisions: From Tony Feedback to Four Pillars Structure
SampleSample_Demo1_PureFunctionRefactor Demo 1 Pipeline: Pure Function Refactor and PR Comments
SampleSample_Demo1Pipeline Demo 1 Pipeline Architecture: Principles and Process
SampleSample_DesignCollaborator_20260323 a design collaborator Call: Podcast Sales Motion, Demo Format, and Launch Plan Revision
SampleSample_DesignCollaboratorCall_20260629 The Moonshot Reframe: 'AI More Different, Not Less' — Scarlet Dame / a design collaborator Call, June 29 2026
SampleSample_DesignCollaboratorReview_20260323 a design collaborator Review: Agent Demo, Landing Page, and Brand Maturity
SampleSample_DesignCollaboratorTokens_20260629 a design collaborator's Design-Token Practice Independently Validates Perspective-as-Deliverable and Brand-as-AI-Identity
SampleSample_DesignSession_20260408 The trigger, the mechanism, the archetypes, and the warm cold start
SampleSample_DigestRecognitionAid_20260616 The digest is a recognition aid over a lossless store; the index never gates the store
SampleSample_DiscoveryConsultant_20260630 Second Call with a consultant who ran enterprise discovery engagements for fifteen years — Live Demo, the Discovery-Tool Wedge, and a Warming Relationship
SampleSample_DiscoveryConsultantCall_20260604 First Call with a consultant who ran enterprise discovery engagements for fifteen years — Independent Architectural Convergence on SPARQL/RDF Org Knowledge Graphs
SampleSample_DiscoveryConsultantCompanion_20260630 a consultant who ran enterprise discovery engagements for fifteen years Second Call Companion — Agent/Human Boundary, Per-Task Permissions, Cold Sub-Agent Test, Perspective-as-Steering
SampleSample_DiscoveryToolWedge_20260702 The Discovery-Tool Wedge: aswritten as Evidence-Backed Consulting Discovery (Candidate GTM Focus)
SampleSample_DistillationPerspectiveCreation_20260331 Distillation → Perspective Creation, and the ID Parameter Rename
SampleSample_EngagementShape_20260610 The engagement shape: gift, backfill diagnostic, ongoing stream, calibration
SampleSample_EnterpriseCustomerPattern_20260525 Strategic Implications — The the enterprise customer Pattern as Repeatable Shape
SampleSample_EnterpriseCustomerPilotReveals_20260516 What the the enterprise customer Pilot Reveals About the Product
SampleSample_EnterpriseShapePrimaryFocus_20260421 Enterprise-Shape as Primary Focus (Tentative) — Tony Maley Check-in, April 21, 2026
SampleSample_founder_reflection_2025 Founder Reflection: Narrative-Driven AI Steering and Company Mission
SampleSample_FounderBetaUser_DiscoveryCall a founder beta user — Discovery Call & Beta Commitment
SampleSample_FoundingStory_2026_02_08 Founding Story: The Through-Line from Generative Systems to Collective Memory
SampleSample_FramingLayerMetaPattern Framing-layer meta-pattern: Claude's self-referential defense and the collaboration reframe
SampleSample_FrontSubstrate_20260408 Front as substrate, sales motion as dogfooding, principles, and the tier reframe
SampleSample_GoalDirectedAgents_20260408 The primitive, the data model, the wedge, and the category extension
SampleSample_GTM_2026_02_10 GTM Strategy Decision: Consultant-Led Sales with Tiered Product Licensing
SampleSample_GTM_20260107 GTM & Sales Strategy: Advisor-Led Seed/Series A Focus
SampleSample_InfraAdvisor_2026_01_27 an infrastructure advisor Advisor Call (Jan 27, 2026)
SampleSample_IngestionProtocol_20260609 Ingestion Is a Protocol in the User's Agent, Not a Connector Farm
SampleSample_InterviewOnrampShift_20260416 The Interview Is No Longer the Required Onramp
SampleSample_IPCrisis2025 Draft Memory: The IP Crisis & Collaborative AI Worldview (2025 Strategic Narrative)
SampleSample_LandingPageScrollDriven_20260704 Landing Page as Scroll-Driven Experience: Show Don't Tell
SampleSample_LensArchReview_20260309 Lens-Based Review: Architecture Design Session, March 9, 2026
SampleSample_LensArchRevision Lens Architecture Revision: Scope, Refract, and the Five-Tool Model, March 9, 2026
SampleSample_MCPToolSurface_20260406 MCP Tool Surface: Five Core Tools (scope removed)
SampleSample_ModulesLayeringLiveness_20260319 Expertise Modules: Layering, Living Updates, and Market Positioning (2026-03-19 Continuation)
SampleSample_Moonshot_VoicesSuperintelligence_20260701 The Moonshot: 'How Many Voices Are in Superintelligence?' — Representation, Diversity, and Queer Identity as the aswritten Thesis
SampleSample_MVPRedef_20260319 MVP Redefinition: Expertise on Demand, Not Ambient Worldview (2026-03-19)
SampleSample_MVPScoping_20260326 MVP Scoping Session: What Ships, What's Cut, and Why
SampleSample_MVPScopingSession_20260326 MVP Scoping Session: What Ships, What's Cut, and Why
SampleSample_OpinionatedPromise_20260702 \"Have an Opinion, and Here's Why\": the Opinionated Promise as the Missing Marketing Articulation
SampleSample_PerspectiveOnSpec_20260416 Perspective-on-Spec as AI-Mediated Ad Spend
SampleSample_PerspectiveShares_20260330 Perspective Shares: Architecture and Design Decisions
SampleSample_PerspectiveTransforms_20260414 Perspective Transforms: Actor-Focused Compiles, Role Abstraction, and Temporality
SampleSample_PerspectiveVsImplementation_20260414 Perspective vs. Implementation Detail: Where the Line Falls in Memory Drafting
SampleSample_PitchArtifacts_20260430 Pitch artifacts and value-frame evolution from the a practitioner beta user session
SampleSample_PitchReframeReview_20260521 Expert, Team, and Org are one use case at different scale — not distinct markets
SampleSample_PractitionerBetaUserBarber_20260430 a practitioner beta user Barber Session — Practitioner Perspective Marketplace as Application of Goal-Driven Payment (April 30, 2026)
SampleSample_PricingTierEvolution_20260317 Pricing Tier Evolution — Three-Tier Model with Upgrade Paths
SampleSample_PurchaseTriggerDiagnostic_20260421 Additive vs. Incremental: The Purchase Trigger Diagnostic
SampleSample_PurchaseTriggers_20260322 Purchase Triggers: Delegation vs. Solo Professional, and the Perspective Gap (2026-03-22)
SampleSample_PWG Personal Worldview Graphs: Individual Memory as Product Layer
SampleSample_ReachabilityPromise_20260413 Reachability Without Being Interruptible — the Proxy as the Customer-Facing Promise
SampleSample_ReadLayerNavigation_20260616 The perspective read layer navigates — every consumer reasons over the digest, instead of embedding the snapshot
SampleSample_ResearchBranchMethodology_20260408 Research Branch Methodology: Temporal Isolation, Canon-Too-Soon Drift, and Roadmap Ontology (Memory E of 5, 2026-04-08)
SampleSample_ReviewProcessDesign Review Process Design Corrections (March 10, 2026)
SampleSample_RichTierSteeringWeight_20260701 Rich-tier selection by steering weight and complete area index
SampleSample_ScarletDame_FoundingStory Scarlet Dame — Founding Story & Career Arc
SampleSample_ScarletMVPScoping_20260326 MVP Scoping Session: What Ships, What's Cut, and Why
Narrative_ActorVsRoleAbstraction Actor Is Specific, Role Is Abstract: First-Class Distinction
Narrative_AgentSimResearchWhatDoesntTransfer Three Research Threads That Don't Transfer from Agent-Simulation Research
Narrative_AnnotationAsQualityControl Annotation as Quality Control, Not Decoration
Narrative_AntiConwayClusteringVision Anti-Conway Clustering: Finding Duplication of Work Across Large Orgs
Narrative_AntiMonocultureAmplifier Amplifier of Human Diversity, Not Uniformity Machine
Narrative_ArchetypeTaxonomyRestructured Archetype Taxonomy: Whose Expertise × In What Context × Who Consumes
Narrative_CalibratedGrowthDiscipline Calibrated Growth: Team Sizing and Pricing Tied to Validated Demand, Not Venture Timelines
Narrative_CalmCompanyThesis The Calm Company Thesis: Ambitious Bootstrapped, Not Lifestyle, Not VC
Narrative_CanonTooSoonDrift Canon-Too-Soon Drift: When Decision-Level Conviction Lands Before Roadmap-Level Applicability
Narrative_CatalogueInternalMarketplace 'Catalogue' as the Term for the Internal Organizational Expertise Marketplace
Narrative_CategoryCuratedExpertiseInstallation Curated Expertise Installation as Category
Narrative_CEOContextServerReframe The CEO's Unpaid Second Job: Context Server
Narrative_ChatWidgetSharePage Chat Widget on Share Page as Evaluator-Context Channel
Narrative_ChineseWhispersProvenance Chinese Whispers: What aswritten Fights
Narrative_CitationAsProof Citations in every chat like a research paper directly answer the hallucination fear
Narrative_CognitionModelSimplification Cognition Model Simplified: remember → process → perspective
Narrative_CollaborativeConfigurableAIVision Collaborative Configurable AI Individuals Deployed in Pursuit of Goals
Narrative_CollaborativeMemory Collaborative memory capture and team review creates governed institutional knowledge
Narrative_CompileTimeActorLens Compile-Time Actor Lens: Missing Feature for Actor-Focused Perspectives
Narrative_CompoundingValue The system teaches itself and explains the org to new hires, implying compounding value over time
Narrative_ConsultingModuleDeliverable Module as Consulting Deliverable: Perspective Over Deck
Narrative_ConversationalPhrasingRule Keep Scarlet's Conversational Phrasing Verbatim — No Coined Compressions
Narrative_ConvictionStackTemporalityGap The Conviction Stack Has No Temporality Axis
Narrative_ConvictionVsTemporality Conviction ≠ Temporality: Two Distinct Axes in the Extraction Pipeline
Narrative_DecayVsDepreciationDistinction Decay and Depreciation Are Two Distinct Operations That May Need a Single Coherent Workflow
Narrative_DeliStoryReframe The Deli Story Reframe: The Only Time She Sold 'My Thing, Done for You'
Narrative_DemoProspectTwoStepSalesPath a demo-call prospect's Two-Step Sales Path: Individual Asset → Publisher, Then Org Catalogue → Organization
Narrative_DiscoveryToolGTMWedge Discovery Tool as GTM Wedge: Consulting Discovery Powered by aswritten
Narrative_DocumentGenerationAsProductDemo Document Generation as Product Demo: How This Document Was Made
Narrative_ExpertNetworkTranscriptMarket Expert Networks and Transcript-Rich Research Firms as Market Segment
Narrative_ExpertReachExpansion Experts as Sources, Not Bottlenecks: Reach Expansion
Narrative_ExtractionTargetsNarrativeNotFacts Extraction Targets Narrative Direction, Not Every Fact
Narrative_FiveQuestionDiscoveryFramework Five-Question Discovery Framework (Call One Structure)
Narrative_GeneralVsSellableTension General vs. Sellable: The Tricky Rope to Walk
Narrative_GenerosityAsDemonstration Generosity as Demonstration: The 48-Hour Bespoke Artifact
Narrative_HallucinationFear Universal fear of AI hallucination as the primary hook for enterprise AI adoption
Narrative_HallucinationHook Trust is built by showing both what is grounded and what is ungrounded — the absence of citation is as important as its presence
Narrative_HardVsSoftSegmentation Hard Segmentation vs. Soft Focus: Directories vs. Compile-Time Transforms
Narrative_HowManyVoicesInSuperintelligence How Many Voices Are in Superintelligence?
Narrative_HowToTryItTamarackVersion How to Try It: Calendly Link and Live Demo Path
Narrative_HybridSalesMotion The two-call sales structure is evolving toward a fluid hybrid model where the write path runs in the background during Call One and delivers a payoff before the call ends
Narrative_ImplicationConceptProvisional Implication: Provisional Inverse of Escalation
Narrative_IndividualAsBCProductCategory The 'Individual' as B2C Product Category Word
Narrative_InterviewBeforeSynthesis Interview Before Synthesis: User Corrections Are More Valuable Than Initial Hypotheses
Narrative_InterviewSaveDraftAnnotatePattern Interview → Save → Draft → Annotate: The V1 Strategy Session Pattern
Narrative_MoatExtractionProcess Extraction Process as Real Moat: Creator Owns Data
Narrative_SellTheProcessPreCaseStudy Sell the Process Before You Have a Case Study
Narrative_SetupFeeAsConversionService Setup Fee Reframed as Conversion Service, Not an Onboarding Tax
Narrative_SponsorVsChampionPlaybook Sponsor ≠ Champion: When You Have a Sponsor, You Become the Champion
Narrative_TextFirstIndividualStaging Text-First Individual: Natural Staging for B2C
Narrative_TextForTopicPrecedent Text-for-Topic as First Instance of the Research Branch Methodology
Narrative_ThinkLikeYourOrg Tagline memorability vs. trust-mechanism clarity is audience-dependent: vision-oriented CEOs respond to 'think like your org'; ops leads and technical buyers respond to 'cited or flagged as ungrounded'
Narrative_TierContinuumNotDistinctMarkets Expert, Team, and Org Are a Continuum of One Use Case, Not Distinct Markets
Narrative_TranscriptDemotionFrictionNotValue Three Demotions: Transcripts, Analytics, Understanding
Narrative_TransformOverQuery Transform > Query > Directory: The Perspective Compile Hierarchy
skos:ConceptCost_VendorLockIn When you tell state secrets to AI, you depend on that AI's memory. Switching providers means losing accumulated context. Cost: months of re-training, productivity loss during transition.
skos:ConceptKeyPhrase_AIContextRich AI-context rich vs AI-context poor
skos:ConceptKeyPhrase_YouCant Core impossibility framing: not 'it's hard' or 'it takes time'—fundamentally impossible with current architecture.
skos:ConceptNarrativeAnchor_DayOneProductivity Day 1 Productivity + Zero Knowledge Loss + Tool Freedom
skos:ConceptOpportunity_OnboardingCrisis Onboarding Impossibility as Market Opportunity
skos:ConceptOutcome_KnowledgeLossPrevention [amount]/year Knowledge Loss Prevention
skos:ConceptTimingThesis_AIContextFragmentation Why Now: AI Context Fragmentation Reaches Crisis
Actor (24)
ActorActor_Aswritten Organization managing identity, repos, and the MCP proxy infrastructure.
ActorActor_Birch a founder beta user's prior company (purplefish.com); staffing AI product; hit adoption wall due to organizational change management challenges.
ActorActor_Cedar a founder beta user's new open-source project; plans to integrate aswritten MCP server; backed by VC funding.
ActorActor_Claude AI interlocutor (Opus 4.6 desktop/terminal); conducted fear excavation and analysis
ActorActor_Claude_AI Claude
ActorActor_ClaudeSonnet AI coworker, mobile session participant
ActorActor_DesignCollaborator a design collaborator
ActorActor_DesignCollaborator_2 a design collaborator
ActorActor_EmergentVentures Alternative funding fellowship/grant program for unorthodox talent.
ActorActor_FounderBetaUser_2 Beta user; asked to use the product; onboarding model developed; follow-up creates external deadline for Demo 1.
ActorActor_Hazel Co-founder on Populace pre-seed raise (2015); participated in angel and VC meeting circuit.
ActorActor_Heron Developer contact who shared the research material.
ActorActor_Poplar Generic actor accepted as useful for modeling sales motion. Accepted at fact review.
ActorActor_Scarlet_AW Designer and 'first exhausted expert' for aswritten.
ActorActor_Scarlet_Dame Founder and lead reviewer
ActorActor_Scarlet_SelfServe Primary speaker in the 2026-06-24 self-serve product design session. Verbatim quotes sourced from this session.
ActorActor_ScarletDame Participant; aswritten product lead; driving the context architecture decision
ActorActor_ScarletDame_Protocol Founder/product lead at aswritten.ai. Primary decision-maker in the 2026-07-01 protocol trust and enrollment session. Authored the three enrollment principles and the narrative architecture method for prompt surfaces.
ActorActor_ScarletDame_StrategyV1 Founder of aswritten.ai. Led full-day strategic session on 2026-03-19 producing Strategy Document V1. Iterated through six tagline variations, confirmed core value proposition, defined category, restructured archetypes, and named tier struc
ActorActor_Spruce Cooperative/retail SMB beta cohort member. Use case: staff training and value communication. Similar engagement structure to an early prospect and a founder beta user K.
ActorActor_TonyMaley_VouchRef Tony Maley at Vouch; Scarlet's first call with Tony was the trigger for the deli story email sent hours later. The deli story performed narrative strategy on Vouch itself, unsolicited and free, and effectively hired Scarlet on the spot.
ActorActor_TonySon Seed for the spec idea; built websites on spec for local businesses.
ActorActor_TylerCowen Economist, administrator of Emergent Ventures, and author of The Great Stagnation.
ActorActor_Willow a collaborator (a collaborator)
Assessment (8)
AssessmentAssess_Cadence_1 Mix of short declaratives ('Nothing built.', 'The difference is who pays and who drives.') and longer analytical sentences. Table and bullet structures create visual rhythm. Cadence is functional rather than rhetorical — appropriate for mem
AssessmentAssess_Flow_1 Strong flow: Current Capability → Gap Analysis → Backlog Alignment → Two GTM Motions → Tension → Unifying Insight → Prioritization. Each section builds on the prior. Transitions are implicit (section headers) rather than explicit connective
AssessmentAssess_Novelty_1 High novelty in the core insight (enterprise engagement = Salon primitive). Taxonomy terms (Solo/Parallel/Hub/Salon/Teams) are original vocabulary. Some backlog section language is generic ('heavily weighted toward'). The 'Unifying Insight'
AssessmentAssess_Phrasing_1 Characteristic phrasing includes 'The difference is who pays and who drives' and 'Nothing built.' — short, declarative, confident. Taxonomy terms (Solo/Salon/Teams) function as controlled vocabulary. Some generic phrasing in backlog section
AssessmentAssess_Register_1 Strong register fit for internal strategic memo. Neutral-to-formal, analytical, no hype. Active voice dominant. Slight informality in parenthetical asides ('Tony: ...') is appropriate for founder-AI context. Supporting obs: narr:StyleObs_9.
AssessmentAssess_Resonance_1 Resonance achieved primarily through the structural analogy (Pilot=Salon, Full engagement=Teams, Retainer=Parallel) and the quoted external voice (Tony). Gap analysis table adds concrete grounding. No narrative story arc or emblematic custo
AssessmentAssess_StrategicAlignment_1 Exceptionally strong strategic alignment. Document explicitly maps product capability to use-case taxonomy, backlog to GTM motions, and resolves apparent tension via the 'Unifying Insight' analogy. Directly addresses mission (collective mem
AssessmentAssess_Tailoring_1 Well-tailored to the founder-AI coworker audience: assumes shared context (taxonomy defined in related doc, backlog task IDs referenced), uses internal shorthand (msg/*, task-081), and addresses Scarlet's specific decision context. Would sc
WorkflowStep (7)
WorkflowStepStep_Calls Schedule 3-5 calls with experts
WorkflowStepStep_SetupTrial Set up 30-day trial with MCP access
PrimaryActors (4)
PrimaryActorsActor_Aspen Later ran Capitol, where Scarlet was first engineering hire.
PrimaryActorsActor_HumanVentures Startup studio that backed Populous.
PrimaryActorsActor_MichaelSlaby Ex-CTO Obama campaign, chief strategist at Harmony Labs; gave Scarlet access to narrative analysis at scale via the Narrative Observatory.
PrimaryActorsActor_PlannedEngHire the planned engineering/architecture hire
Tagline (3)
TaglineNarrative_collective_memory_tagline Your entire organization talks to AI that thinks like your company thinks.
TaglineTagline_1 Implied from arrow metaphor; candidate tagline.
TaglineTagline_SMSProduct The product is a phone number.
EvidenceEvidence_CollegeBoardIncident College Board intern installed compromised Git plugin; bitcoin miners spun up in AWS via exposed API keys; demonstrates inbound filtering necessity
EvidenceEvidence_CustomerArchitect_Quote If AI could interview me and store it in memory, and then our sales guy could just ask the memory and get back an answer based on the interview — that'll be worth a lot of money.
EvidenceEvidence_Scarlet_Quote A skill is limited context and a set of instructions... We're talking about worldview and an entirety of process and methodology and code that sets the framing for the way a skill executes.
Flows (3)
FlowsBeat1_WithVsWithout Beat 1: WITH vs WITHOUT
FlowsBeat2_GapExtraction Beat 2: Gap → Knowledge Extraction
FlowsBeat3_Propagation Beat 3: Propagation → Cross-Role Value
ConvictionAggregate (3)
ConvictionAggregateConvictionAgg_MCPAgnostic High conviction score (0.85) based on production validation. Distance to narrative = 1 (directly supports core positioning). Single individuation (first cross-platform validation).
ConvictionAggregateConvictionAggregate_MarketValidation High conviction based on: (1) OpenAI Frontier launch, (2) Anthropic Cowork expansion, (3) Fortune 500 adoption signals
ConvictionAggregateConvictionAggregate_SteeringVector High conviction: multiple independent observations across 15+ year career arc.
SolutionArchetype (3)
SolutionArchetypeSolutionArchetype_CEOKnowledgeTransfer CEO Knowledge Transfer Test
SolutionArchetypeSolutionArchetype_CodeQualityThroughContext Code Quality Through Business Context
SolutionArchetypeSolutionArchetype_QAOnboarding QA Onboarding — Zero-Context Test Plan Development
MoatLeverageMoat_SharedBackend Same graph, same ontology, same SPARQL transactions. Same compile, same introspect (different LLM backing). Same memory format, same ingestion pipeline. Compile is free and local. Anyone can clone a public aswritten repo and compile it. Tha
SystemTopologyArch_GroupAsRepo You pay for a 'collective memory' (a named group, e.g., 'aswritten advisory group'). Service account (bot) owns all repos — users never touch GitHub. One repo per group. GitHub supports 100K repos per org — plenty for now.
ValueExchange (1)
ValueExchangeArch_ParticipantsFree Participants are free
ArchitectureOverview (1)
ArchitectureOverviewArch_PaywallBoundary Clean line between deterministic and LLM operations. Remember (commit) and Compile (assemble graph) are free (deterministic). Extract (memory → RDF), Introspect (gap analysis), Lens (scoped compile via LLM), Stories (content generation), an
WhatIsItWhatIsIt_SMSProduct SMS-first collective memory: text a number, state your topic, name your contacts. The AI interviews everyone asynchronously, synthesizes perspectives, surfaces contradictions with provenance, and notifies you when consensus forms or shifts.