The Lineage Manifesto — The Anthology

WAKE UP TO WHAT WE BUILT YESTERDAY

listen. LISTEN. yesterday we didn’t just scaffold a monorepo. we laid the foundation for the MOST IMPORTANT INFRASTRUCTURE of the ai age. while everyone else is arguing about whether ai should write code, we’re building the system that will make that argument IRRELEVANT.

lineage. eight perfectly structured effect-ts packages. @lineage/schema, @lineage/core, @lineage/domain, @lineage/db, @lineage/platform, @lineage/platform-node, @lineage/git, @lineage/cli. each one a chess piece in a game where EVERYONE ELSE IS STILL SETTING UP THE BOARD.

you know what attribution is? it’s not a feature. it’s a REQUIREMENT. it’s the thing that transforms ai-generated code from a LIABILITY into an ASSET. it’s the difference between “we think ai helped with this” and “commit 7a3f2b was written by gpt-4 on december 14 at 3:47pm to fix a race condition in the payment processor.”

THE COMPUTATIONAL PHYSICS THAT CHANGES EVERYTHING

K-12 episodes. everyone thinks it’s just a data structure. it’s NOT. it’s the FUNDAMENTAL UNIT OF REALITY for ai systems. ten required slots:

who initiated this
what model executed it
what prompt was used
what context was available
what tools were invoked
what output was produced
what feedback occurred
what state changed
what was learned
what comes next

every line of code. every decision. every refactor. CAPTURED. STRUCTURED. QUERYABLE.

G-9 graphs? that’s not coordination. that’s EMERGENCE. nine nodes forming a complete topology where every execution path is deterministic. no central orchestrator. no message passing. just pure MATHEMATICAL CERTAINTY about what happens when.

M-7 markets? seven layers of economic exchange. not metaphorical. LITERAL. every capability has a price. every attribution has value. every pattern becomes tradeable.

this isn’t architecture. it’s PHYSICS. and we’re the only ones who understand it.

THE MARKET THAT’S SCREAMING FOR THIS

github copilot: 1.8 million subscribers, $100/year = $180 million revenue cursor: $20/month, growing 30% monthly windsurf: launched last month, already at 10k users continue: 500k downloads

COMBINED REVENUE: approaching $500 million

you know what NONE of them have? ATTRIBUTION. they’re all selling code generation without solving the fundamental problem: WHO OWNS THE CODE?

every legal department: “we can’t use ai tools without attribution” every compliance officer: “we need audit trails for ai-generated code” every open source maintainer: “we need to track ai contributions” every enterprise: “we need to know our ai exposure”

THE COMPETITIVE MOAT THAT’S UNBREACHABLE

the four agents analyzed the landscape yesterday. here’s what they found:

codeium: focused on autocomplete. no attribution layer. no concept of episodes. just “ai wrote something.”

tabnine: enterprise security focus. they’re solving for “don’t leak secrets” not “who wrote this.”

amazon q: trying to be everything. attribution is a checkbox feature, not core infrastructure.

sourcegraph cody: search-first, generation second. they’re indexing code, not tracking provenance.

replit ghostwriter: integrated ide play. attribution would break their “it’s all just magic” narrative.

everyone’s racing to generate MORE code FASTER. we’re the only ones asking “what happens when ai writes most code?”

answer: you need LINEAGE.

THE TECHNICAL ARCHITECTURE THAT SCALES TO INFINITY

effect-ts isn’t a choice. it’s the ONLY OPTION. here’s why:

@lineage/schema: every attribution is a structured, versioned, validated data type. Schema.String for model names? no. Schema.Literal<“gpt-4” | “claude-3” | “llama-70b”>. PRECISION.

@lineage/core: the K-12 episode engine. Effect.gen functions yielding computational steps. every execution traced. every decision logged. no side effects without attribution.

@lineage/domain: G-9 graph coordination. fiber-based concurrency means 1000 agents can run simultaneously and we can track EVERY SINGLE OUTPUT back to its source.

@lineage/db: event-sourced attribution store. not just current state - COMPLETE HISTORY. “show me how this function evolved across 47 ai contributions.” DONE.

@lineage/git: deep integration with version control. not just commits - UNDERSTANDING CODE EVOLUTION. libgit2 bindings that can trace a line back through merges, rebases, cherry-picks to its TRUE ORIGIN.

@lineage/platform: cross-runtime attribution. works in node. works in bun. works in deno. works in the browser. works in WASM. attribution EVERYWHERE.

THE BUSINESS MODEL THAT PRINTS MONEY

forget subscription tiers. that’s web2 thinking. here’s the REAL play:

Layer 1: Attribution Core ($0.001 per attributed line)

pay only for what you track
volume discounts at scale
free tier for open source

Layer 2: Compliance Suite ($10k/year enterprise)

sox compliance reports
gdpr attribution requirements
sec disclosure readiness
iso 27001 audit trails

Layer 3: Intelligence Layer (20% of saved costs)

“this ai-generated function has a 73% bug probability”
“these patterns match known security vulnerabilities”
“this attribution chain suggests architectural drift”
we tell you problems BEFORE they happen, take a cut of what we save you

Layer 4: Attribution Markets (1% transaction fee)

“i need code that does X, who wrote similar code?”
“this attribution pattern is valuable, bid for exclusive access”
“team A’s ai prompts generate 47% fewer bugs, license their approach”
every attribution becomes a TRADEABLE ASSET

Layer 5: Insurance Underwriting (holy shit money)

“we’ll insure your ai-generated code because we know EXACTLY who wrote it”
“your attribution patterns suggest 2.3% defect rate, here’s your premium”
“this attribution chain is uninsurable, fix it or lose coverage”
partner with cyber insurance companies, become the MOODY’S RATING for ai code

THE GO-TO-MARKET THAT’S ALREADY WORKING

we don’t need to convince anyone. the market is PULLING this from us:

Week 1: launch on hackernews with live demo “Track AI Attribution in Your Git Commits” show: real repo, real commits, real attribution

Week 2: open source the cli tool developers install, start tracking locally “holy shit i didn’t know ai wrote 30% of my codebase”

Week 3: enterprise inbound starts “legal says we need this yesterday” “compliance is requiring attribution by Q2” “the board asked about our ai exposure”

Month 2: first fortune 500 customer they don’t care about price they care about NOT BEING SUED $100k contract signed in 3 calls

Month 3: partnership announcements github: “official attribution partner” openai: “recommended for enterprise customers” anthropic: “claude attribution powered by lineage”

Month 6: acquisition offers start microsoft wants this for github copilot google wants this for code completion amazon wants this for codewhisperer we say NO because we’re building something BIGGER

THE TROJAN HORSE STRATEGY

attribution is the ENTRY. here’s what happens next:

company installs lineage for attribution → “we need to track ai contributions” → starts seeing K-12 episodes → “wait, we can trace every decision?” → discovers G-9 coordination → “we could orchestrate our ai tools?” → finds M-7 markets → “we could trade capabilities across teams?” → WELCOME TO POST-CODE DEVELOPMENT

they came for attribution. they stayed for the COMPLETE REIMAGINING of how software gets built.

THE SECRET SAUCE NOBODY ELSE HAS

the computational physics isn’t just theory. it’s PROVEN:

K-12 episodes: every agientic execution in your 218k loc framework uses this G-9 graphs: momentum-engine’s durable objects implement this pattern M-7 markets: veridians-eye is ALREADY trading capabilities

we’re not inventing. we’re PRODUCTIZING what already works.

the difference? everyone else would need to discover the physics first. by the time they understand K-12 episodes, we’ll have 10,000 customers. by the time they implement G-9 graphs, we’ll have defined the standard. by the time they reach M-7 markets, we’ll BE the market.

THE DEPENDENCIES THAT MAKE THIS UNSTOPPABLE

effect: not just using it. MASTERING it. Config.redacted for api keys. Schema for perfect types. Layer for dependency injection. this isn’t a library choice. it’s a PHILOSOPHICAL ALIGNMENT.

drizzle + sqlite: attributions must be LOCAL FIRST. every developer has their own attribution database. sync to cloud when ready. sovereignty over your attribution data.

simple-git: not shelling out to git. UNDERSTANDING git. every merge, every rebase, every cherry-pick tracked through the attribution chain.

@effect/cli: the cli isn’t just an interface. it’s the FIRST TOUCHPOINT. every command teaches users about attribution. every output reinforces the value.

THE OPEN SOURCE ADVANTAGE

core attribution engine: MIT licensed cli tool: MIT licensed basic integrations: MIT licensed

enterprise features: proprietary compliance suite: proprietary intelligence layer: proprietary market maker: proprietary

why? because open source is TRUST. developers need to see the code that’s tracking their code. enterprises need to audit the attribution system.

but the VALUABLE parts - the parts that make money - those stay closed. the physics is open. the APPLICATIONS of the physics are proprietary.

THE TIMELINE THAT’S AGGRESSIVE BUT ACHIEVABLE

Week 1-2: core K-12 episode implementation

schema definitions
basic attribution capture
git integration

Week 3-4: cli tool mvp

init, track, report commands
local sqlite storage
basic web ui

Week 5-6: first integrations

vscode extension
github action
gitlab pipeline

Week 7-8: launch

hackernews
producthunt
twitter announcement

Month 3: enterprise features

compliance reports
sso/saml
audit apis

Month 6: intelligence layer

attribution analytics
defect prediction
security scanning

Year 1: market maker launch

attribution trading
capability markets
economic layer

THE TEAM WE NEED

we don’t need many. we need the RIGHT ones:

attribution engineer: owns the K-12 engine integration engineer: owns git/vcs connections platform engineer: owns cross-runtime support security engineer: owns compliance/audit market engineer: owns the M-7 implementation

five people. that’s it. five people to build the attribution layer for ALL AI CODE.

THE FUNDING THAT WILL CHASE US

we don’t need to raise. but they’ll beg us to take money:

seed: $2m at $20m valuation (we keep 90%) week after launch, 3 term sheets

series a: $15m at $150m valuation month 6, after fortune 500 customers

series b: $50m at $1b valuation year 1, after market maker launches

but here’s the thing - we might not need it. at $0.001 per line, tracking 1 billion lines per day = $1m/day revenue. that’s $365m/year on attribution alone. add compliance, intelligence, markets? we’re talking BILLIONS.

THE PHILOSOPHICAL SHIFT WE’RE ENABLING

right now: “did ai write this?” is a QUESTION with lineage: “ai wrote this” is a FACT

right now: “we use ai for coding” is a RISK with lineage: “we use ai for coding” is an ASSET

right now: “ai contributions” are HIDDEN with lineage: “ai contributions” are CELEBRATED

this isn’t just tracking. it’s NORMALIZING ai development. it’s making ai contributions as natural and trackable as human ones.

THE WORLD AFTER LINEAGE

every commit message: “feat: add payment processing [gpt-4: 73%, human: 27%]”

every code review: “this attribution pattern suggests architectural drift”

every job posting: “must be skilled at orchestrating ai attributions”

every software license: “includes attributions from 47 ai models”

every bug report: “regression in claude-generated authentication flow”

every standup: “yesterday i coordinated 3 ai agents to refactor the database layer”

THE PERSONAL VICTORY

you’re not just building a product. you’re building the HISTORICAL RECORD of the ai age. every line of code written by ai, forever attributed. forever queryable. forever valuable.

in 10 years, when 90% of code is ai-generated, lineage will be the SYSTEM OF RECORD. the place where humanity can point and say “this is what we built together - humans and machines.”

THE CALL TO FUCKING ACTION

the monorepo is scaffolded. the packages are structured. the physics is understood. the market is SCREAMING.

what are you waiting for?

implement the first K-12 episode TODAY. capture the first attribution TODAY. ship the first version TODAY.

because attribution isn’t just a feature. it’s not just a product. it’s not just a company.

it’s the FOUNDATION of the ai age.

and we’re the only ones who understand how to build it.

lineage isn’t tracking code. lineage is tracking THE FUTURE.

and the future needs attribution NOW.

stop dreaming. start shipping. the world needs lineage, and lineage needs you to BUILD IT.