THROUGHLINE — a generative venture framework

A holistic, generative framework for building a venture from raw motivation down to constructed artifice — and keeping every layer coherent as decisions change. Instantiated first on OPEN PANEL (framework/INSTANCE_OPENPANEL.md); the core is venture-agnostic. Status: DRAFT v0.1. Companion: LAYERS.md, LENSES.md. This framework is itself generated by this framework (its name is a Layer-4 decision).

0. The problem it solves

Business-development and brand work is usually produced as disconnected artifacts — a deck, a name, a legal structure, a token, a pitch — authored at different times by different hands. They drift out of coherence: the pitch says one thing, the cap table another, the brand a third. Nobody can answer “if we change this, what else is now wrong?”

THROUGHLINE treats a venture as one generative stack: a small set of layers generated top-down from motivation, grounded bottom-up by research, and held coherent by a machine-checkable dependency spine. The novelty is not any single layer — it is the throughline: explicit lineage so a change at the source sweeps cleanly to every dependent artifact (the slime-mold / critpath model, §4).

1. The stack (seven layers + a spine)

Generated top-down (each layer is an input to the next); validated bottom-up (each layer is checked against the research lenses, §3). Detailed in LAYERS.md.

 0  SUBSTRATE   founder motivation · values · risk appetite · constraints (the brief)
 1  THESIS      the falsifiable belief about the world the venture exploits
 2  VISION      the end-state if the thesis wins
 3  IDENTITY    who the venture IS — character · archetype · values (≠ how it looks)
 4  BRAND       the EXPRESSED identity — name · voice · visual · narrative
 5  ARTIFICE    constructed instruments — entity · capital · token/DAO · governance
 6  MECHANISM   how humans actually engage — incentives · defaults · habit loops · funnels
 ═  SPINE       lineage graph + critical-path sweep keeping all layers coherent

Two axes. Vertical = generation (motivation → artifice) and grounding (research characterizes upward). Horizontal = coherence (the SPINE threads every layer; any canonical change propagates along it). A venture is “done” for now when the stack is internally consistent and each layer is legible under its lenses — its properties along every dimension explicit and owned by the founder, favorable or not.

Why these seven. They are the minimal set where each answers a distinct question and none collapses into another: why (0), what-belief (1), toward-what (2), who (3), how-presented (4), built-how (5), engaged-how (6). Skipping a layer is the usual failure mode — e.g. jumping from THESIS straight to ARTIFICE (a clever legal structure with no IDENTITY) is exactly the trap OPEN PANEL’s first 501(c)(3) draft fell into.

Selection criteria — and the one the framework REFUSES

The framework optimizes for exactly two things, and explicitly refuses a third:

Coherence — the layers do not contradict each other or themselves.
Fidelity — the stack faithfully expresses the Layer-0 SUBSTRATE, including the founder’s own definition of success (which may deliberately reject conventional optimization).
NOT desirability — but success still counts. Separate two things that get conflated. Success — does the venture achieve the founder’s own goals (which can absolutely include profit, durability, scale, impact, longevity)? — is legitimate and worth measuring. Desirability — the external frame that a venture ought to maximize margin, defensibility, scale, fundability, and exit-readiness — is not a yardstick the framework imposes. It surfaces those properties; it never marks a venture down for declining to maximize them.

Canonical case — Arizona Iced Tea: held at 99¢ for 30+ years, refuses price hikes, refuses to sell, refuses to maximize margin. By every desirability heuristic it’s “irrational” — and it is also a durable, hugely successful, fiercely coherent venture that privileges a value (its promise to the customer) over desirability. A desirability-privileging tool would tell Arizona to raise the price. THROUGHLINE must generate and honor Arizona as-is, not “improve” it toward attractiveness. The SUBSTRATE is sovereign.

The deeper hazard — desirability is often imposed, not chosen. It rides in through the ARTIFICE layer (capital & ownership structure). Public markets and VC carry a growth/ margin-maximization mandate that can override the venture’s own success definition: video-game studios that shipped great, even profitable games have been gutted or shuttered under shareholder pressure for failing to be maximally desirable (recurring monetization, predictable growth) — desirability actively destroying success, craft, and identity. So a Layer-5 structure choice — stay private · steward-ownership / mission-lock · go public — is also a choice of how much external desirability pressure to admit. OPEN PANEL’s PBC / steward-trust path (business/ENTITY_STRUCTURE_REVIEW.md) is exactly this defense.

This is what makes the core universal rather than a desirability-optimizer: it measures a venture against the founder’s own success definition and tells the truth about every property — it does not substitute the market’s preferences for the founder’s.

2. Identity vs Brand vs Artifice (the three most-confused layers)

IDENTITY (3) is who you are — the venture’s character, values, and archetype. Stable. Answers “what would we never do?” Lives upstream of expression.
BRAND (4) is how identity is expressed — name, voice, visual system, narrative. The swappable surface. (OPEN PANEL’s naming/clearance work — brand/NAMING_CLEARANCE.md — is Layer 4.)
ARTIFICE (5) is the deliberately-constructed machinery — entity form, capital structure, token/DAO, governance. The instruments that make identity operable in the world. (“Artifice” is meant positively: designed mechanism, not deceit.)

Keeping them distinct is what lets you re-skin the BRAND without touching IDENTITY, or swap the ARTIFICE (501c3 → PBC) without re-deciding who you are.

3. The lenses (research disciplines as operational levers)

Four cross-cutting disciplines characterize each layer — they diagnose, they do not approve (per §1’s refusal of desirability). Operationalized as checklists and decision-levers in LENSES.md — not literature reviews.

Lens	Surfaces (describes, does not grade)	Applied hardest at layers
IO-psych	founder/team motivation structure, org identity, role & selection, culture	0, 3
Behavioralism	choice architecture, incentives, the habit loops a MECHANISM assumes, funnels, ethics-of-nudge	6, 4
Economic research	market structure, defensibility (or its absence), unit economics, commons/public-goods, mechanism design	1, 5
Identity / semiotics	archetype, narrative, naming, distinctiveness, trademark exposure	3, 4

What counts as a defect — and what doesn’t. A defect is incoherence or self-deception: a layer that contradicts another, or claims a property it lacks — a MECHANISM assuming a behavior the behavioral lens says won’t occur (gamble unowned); a THESIS claiming a moat it doesn’t have. It is not a defect to decline the desirability levers: refusing to raise prices (Arizona’s 99¢), refusing to scale, choosing no moat, foreclosing an exit, staying small or weird — owned strategic properties. Boundary: genuine un-viability on the founder’s own terms (losing money per unit while intending to profit) is a real, surfaceable finding — but that’s a success/fidelity concern, not a desirability judgment. Success ≠ desirability; the lens forces the property into the open, then lets the founder own it — it never rejects a venture for declining to maximize.

4. The spine (coherence as machinery)

Each artifact carries a lineage block (tools/critpath/LINEAGE.md): what it conforms_to, depends_on, defines, consumes. critpath builds the DAG, ranks the trunk by downstream flux, finds the critical path (longest conformance chain), and — on any canonical change — runs a sweep that classifies every artifact as SOURCE / REVIEW / HIDDEN-DEP / UNTRACKED. The gap between declared dependency and literal usage is the actionable finding.

This is the slime mold: artifacts are nodes, conformance edges are tubes, a decision change is a nutrient pulse from the source; starved nodes are stale. The spine is what makes the framework “generative” rather than a one-shot template — you can regenerate any layer and mechanically find everything downstream that must follow.

5. The generative process (six moves)

Detailed in LAYERS.md (per-layer prompts) and below. Each move maps to machinery we already have.

ELICIT — extract the SUBSTRATE as a structured brief (motivation, constraints, goal) via forced-choice interview. (Machinery: AskUserQuestion-style decision capture.)
GENERATE — produce each layer as a lineage-annotated artifact; record every fork as a canonical decision in the source-of-truth doc. (Machinery: a CANON.md-pattern.)
GROUND — run each artifact through its lens levers (LENSES.md); fix lens defects.
CLEAR — clear the ARTIFICE & BRAND against reality: names, namespaces, entities, tickers. (Machinery: tools/landgrab.)
COHERE — run critpath sweep after every canonical change; resolve REVIEW/ HIDDEN-DEP/UNTRACKED. (Machinery: tools/critpath.)
ITERATE — amendments feed back to the source via an amendment back-edge (the way ENTITY_STRUCTURE_REVIEW.md fed CANON). The stack is a loop, not a waterfall.

6. Universalizable core vs instance

The core (this repo’s framework/ + tools/): the layer model, the lens lever-sheets, the six-move process, and the machinery (lineage/critpath, landgrab, decision-capture, source-of-truth pattern). Venture-agnostic.
An instance: a populated stack — a source-of-truth doc + per-layer artifacts + cleared artifice + a green spine. OPEN PANEL is instance #1 (INSTANCE_OPENPANEL.md).

To instantiate for a new venture: copy the framework/ + tools/, run ELICIT, and generate down the stack. The tooling is the same; only the content changes.

7. The name — THROUGHLINE, and when the others win

The framework’s own Layer-4 decision. Working name THROUGHLINE; the alternatives carry genuinely different value props (use this as the template for naming any instance):

Name	Value prop it pitches	Pick it when…	Watch-out
THROUGHLINE ★	coherence — the line through every layer; names the real innovation (the spine). BD/narrative vernacular (“find the throughline”), verb-able.	the pitch centers on keeping a venture coherent as it changes.	common word → weak trademark distinctiveness.
FOUNDRY	generation — where ventures are forged from raw motivation. Concrete, industrial.	the pitch centers on producing ventures fast (a generator/studio).	heavily crowded in tech (Palantir Foundry, many) — clearance pain.
TAPROOT	grounding — fundamentals-first; the deep root everything grows from. Organic, on the network/slime-mold motif.	the pitch centers on motivation-first rigor over surface branding.	reads as “depth,” not “coherence/generation.”
PHYSARUM	distinctiveness — the slime-mold organism literal; the coherence engine as a creature. Maximally ownable.	you want a strong, unique trademark and a memorable origin story, and will explain it.	esoteric; hard to spell/say; needs a gloss every time.

Recommendation: THROUGHLINE for internal/operator use — it names what’s actually novel. If this ever ships as a product, reconsider PHYSARUM (ownable mark + on-motif story) or FOUNDRY (clearest outsider pitch of the generative promise). Run the chosen name through tools/landgrab exactly like any BRAND-layer decision.

DRAFT v0.1. The core is venture-agnostic; OPEN PANEL is the worked instance. Not legal/ financial advice — instance ARTIFICE still requires counsel.