# Component Sift — What Sticks From the Reframe Conversation

- Date: 2026-06-11
- Source: full re-read of `Business_Plan_Evaluation_Framework-2026-06-11.md` (28 messages)
- Compounds on: 03-reframe-analysis.md (structure), 04-research-deepening.md (evidence)
- Verdicts: **ADOPT** (into the plans as-is) · **ADAPT** (keep with stated correction) · **PARK** (defer, revisit later) · **REJECT** (drop, with reason)

---

## 1. Verdict Table (all components)

| # | Component (conversation source) | Verdict | One-line reason |
|---|---|---|---|
| 1 | Core inversion: observation → evidence → micro-action → capability → resources | **ADOPT** | The thesis. Evidence-backed (SBIR, Meanwhile Space); dissolves Clark 1.0's blocking risks |
| 2 | Core loop: map → observe → intervene → validate → publish → allocate → expand | **ADOPT** | The engine; every adopted component is a station on this loop |
| 3 | Two-layer strategy (Layer A Protocol / Layer B instantiated nodes) | **ADOPT** | Resolves vagueness-vs-purity; the bridge to all Clark 1.0 assets |
| 4 | Electronics as entry vector, not identity | **ADOPT** | Re-justifies D-004/D-022; cross-industry "entry language" argument holds |
| 5 | Bootstrap pattern: wedge → extract rules → generalize | **ADOPT** | Critical sequencing rule — with a consequence the conversation itself ignores (see §3.4) |
| 6 | Theatric bench (one bench = stage; camera + display) | **ADOPT** | Strongest-validated component (BEST/EPTAC/Lernfabrik; ~$10–25K vs $0.5–1.2M) |
| 7 | Stage 1–4 node growth (Teaching → +Prototype → +Production Cell → Network) | **ADOPT** | Clean maturity model for Layer B; maps to loop stages E–G |
| 8 | "Knowledge-transfer business, not facility business" | **ADOPT** | Doctrine line; the economics confirm it |
| 9 | Scarce resource = trusted instructors/teaching capacity, not equipment | **ADOPT** | Confirmed: CIT certification is the actual bottleneck and moat |
| 10 | Nucleus: 5 non-negotiables (format, ladder, verification, unlock rules, publishing) | **ADOPT** | Correct minimal core; franchise research confirms small-fixed-template survival |
| 11 | Genetic code (6 cultural elements: documented procedures, standards-based work…) | **ADOPT** | Same DNA as clarkinc's claims-register culture — continuity, not novelty |
| 12 | "Resources follow demonstrated capability" / advancement by artifacts | **ADOPT** | SBIR/Kubernetes/ham-radio precedent; the protocol's first principle |
| 13 | "Distribute agency first" | **ADOPT** | Best one-line differentiator vs. money/training/equipment distributors |
| 14 | Wedge projects (small high-visibility intervention exposing a systemic gap) | **ADOPT** | Good operational vocabulary; unit of Builder activity |
| 15 | Clark Regional Survey (standardized area-study methodology) | **ADOPT** | Generalizes clarkinc's corridor/node-profile templates; cheap to formalize |
| 16 | Org functions: Structure / Enablement / Amplification | **ADOPT** | Clean definition of what the company actually does |
| 17 | Non-traditional first capital (manufacturers, municipalities, foundations, associations) | **ADOPT** | Strongly confirmed by funding research; now has named programs + deadlines |
| 18 | Early-messaging rule: lead with the concrete node, prove the method | **ADOPT** | Anti-vagueness guard; governs website and pitch copy |
| 19 | Site activation lifecycle (idle → probed → active → stable → scaled) + "what counts as activated" | **ADOPT** | Needed spec; directly answers the map-illusion risk |
| 20 | Advancement Contract (per-transition: outputs, validators, evidence, unlocks) | **ADOPT** | The #1 missing spec; research supplies the design (auto → sponsored → judged) |
| 21 | Named failure modes: narrative capture, map illusion, adaptation drift | **ADOPT** | Empirically the right ones (Vice; 596 Acres; franchise research). Add promotion bottleneck from 04 |
| 22 | Canonical summary #1 (Operational Systems View) | **ADOPT** | Best of the eight summary variants; base text for the doctrine |
| 23 | Role ladder — 5-stage vs 3-role versions | **ADAPT** | Merge: 5 levels for advancement; Reporter = unlocked function, not a stage (§3.1) |
| 24 | Training as pure response mechanism ("curriculum as byproduct") | **ADAPT** | Half-true; IPC certification is a fixed catalog and it's the revenue wedge (§3.2) |
| 25 | Industrial journalism / Vice-style media layer | **ADAPT** | Keep field-report unit + firewall; cut the social-media-competition ambition for years 1–2 (§3.3) |
| 26 | Investor layer: "first access to surfaced opportunities" | **ADAPT** | Keep as data-tier product; never as coverage influence; investors stay downstream (§3.5) |
| 27 | Spatial intelligence layer (OSM registry, UK prototype) | **ADAPT** | Keep concept; re-scope to Hamilton–Buffalo corridor, registry-as-lead-gen, no Year-1 UK (§3.6) |
| 28 | Space as graded/state-based resource (evenings, corners, idle benches) | **ADAPT** | Conceptually right (Chunker/Saltbox prove demand); legally execute as licence-brokering with defamation guards |
| 29 | "Galvanizing sub-regional economies" as mission register | **ADAPT** | Keep the pride register (Hidden Champions evidence) but anchor every use to measurable outputs — it's the vaguest phrase in the corpus |
| 30 | Mobile/van-transportable bench kit | **ADAPT** | Real (BEST precedent) but not Phase 1 — first bench lives at the host facility; mobility is the Stage-2 replication mechanism |
| 31 | 30–80pp founding doctrine/handbook (ChatGPT's closing recommendation) | **ADAPT** | Write the thin nucleus instead (~10–15pp protocol + charter); the 80pp book contradicts the bootstrap pattern (§3.4) |
| 32 | WorkWarrior/Workpads/ledgers as the network's operating system | **PARK** | Cross-project entanglement; clark-chat's conversation-first spec is the closer in-family fit. Revisit after the node runs |
| 33 | Four-layer ecosystem (Training/Software/Production/Network) | **PARK** | Early-conversation scaffolding, superseded by the two-layer model; keep only as genealogy |
| 34 | 6-factor evaluation framework + Discard…Exceptional scale | **PARK** | Useful personal portfolio tool; not a Clark plan component |
| 35 | Concept ranking table (WorkWarrior, Workpads, KaiOS, mobile OS, marketplaces) | **PARK** | Portfolio context about other ventures; out of Clark scope. KaiOS/mobile-OS skepticism seems right but is not Clark's question |
| 36 | Extended biology analogy (nucleus/cytoplasm/membrane/nutrients/replication) | **REJECT** as public framing | User already ruled it out as main analogy. Keep exactly two internal terms: *nucleus* and *genetic code* |
| 37 | "CLARC" spelling | **REJECT** | Confirmed Clark (user corrected it in-conversation) |
| 38 | Summaries #2, #3 + five "tubular" combinations | **REJECT** as canon | Drafting variants; archive. Each contains good phrases to mine, but eight competing definitions is itself a vagueness risk |
| 39 | "Selected participants transition into Builders" (phrasing in summary #1) | **REJECT** the word "selected" | Contradicts the model's own artifact-not-approval principle; advancement is earned, not selected |

Adopt: 22 · Adapt: 9 · Park: 4 · Reject: 4

## 2. What the Conversation Never Addresses (silent gaps)

The sift is incomplete without naming what the GPT conversation *omits* — it ran as if Clark were greenfield:

1. **Clarkware/IPE is never mentioned.** Three repos of engineering doctrine (7 ADRs, CFX/XMPP architecture) have no place in the reframe. Resolution from 03 §6 stands: IPE = mature-node (Stage F+) tooling; protocol needs lighter tools first; clark-chat's 13-type conversation model is the best-fit substrate for opportunity records/site registry/progression tracking.
2. **Clark Courses (16 complete courses) ignored** — yet it's the readiest Layer-B asset and contradicts "curriculum as byproduct" (see §3.2).
3. **Hamilton agreement, decision log, claims register ignored** — the reframe needs to be reconciled into the existing governance records, not replace them silently (register the adopted components as new decisions).
4. **No competitive analysis** — who else does activation (MEP centers, colleges, economic development orgs)? 04 partially covers this; the plans must.
5. **No people model** — who is the first CIT instructor? Who runs the survey? The entire model assumes participants appear. Cold-start is the real Phase-1 risk and deserves explicit design (instigated observers in seeded areas was mentioned once, then dropped).
6. **No legal/entity design** — Delaware platform vs protocol steward (03 §7.3) remains open, now sharpened by funding-eligibility requirements (Ontario entity, 501(c)(3)).

## 3. The Contested Calls (judgment required, with recommendations)

### 3.1 Ladder shape — merge the two versions
Conversation gives 5-stage (Observer→Organizer→Demonstrator→Builder→Centre Operator) and 3-role (Observer→Builder→Reporter). These answer different questions: the 5-stage is an *advancement ladder*; the 3-role is a *functional taxonomy*. **Recommendation:** keep the 5 levels (L0–L4) as the advancement spine — it maps perfectly onto the validated auto-threshold → sponsored → judged structure — and make **Reporter a function unlocked at Builder level or above** (you may only amplify what you've helped build). This implements "no action, no amplification" structurally and honors the user's instinct that reporters come after builders.

### 3.2 Training: catalog or byproduct? — both, by layer
"Curriculum as byproduct of observed gaps" is elegant but collides with reality: IPC A-610/J-STD-001 are *fixed external standards* — you can't derive them from local observation, and they're the proven revenue wedge ($700–1,200/student) plus the thing funders (SDF, NYSERDA, WIOA) actually pay for. **Recommendation:** standardized IPC certification = Layer B's fixed product (Clark Courses slots here); *responsive* micro-modules (repair clinics, employer-specific gap training) = the protocol's adaptive layer. The byproduct principle applies to the responsive layer only.

### 3.3 Media ambition — phase it hard
The "Vice-style social-media attention competition" framing is the most capital-hungry idea in the conversation, and Vice itself died of the economics. **Recommendation:** Year 1 media = the standardized industrial field report (one publication unit: structured data record + narrative), published on Clark's own site, production-quality but low-volume. Membership/data-tier/video documentary are later rungs of the proven revenue ladder (04 Track 4). Do not hire reporters before builders exist.

### 3.4 Doctrine size — the conversation contradicts itself
ChatGPT's closing move recommends a 30–80pp founding handbook, but its own bootstrap pattern says abstraction is "the compression of lived results" — Phase 2, not Phase 1. An 80-page doctrine before the first cohort is the map-illusion failure mode applied to documents. **Recommendation:** write (a) a thin **Clark Protocol spec** (~10–15pp: the 6 core artifacts from the conversation — opportunity record format, survey template, ladder rules, unlock criteria, validation rules, publication structure) + (b) a one-page **Clark Charter** (Fab-Charter style), + (c) the grounded business plan for the first node. The handbook is written after Phase 1 from real results.

### 3.5 Investor "first access" — productize carefully
"First access to surfaced opportunities" is sellable (The Information Pro precedent) but is also the exact lever that inverts the flywheel if it touches editorial/amplification choices. **Recommendation:** first access = a paid *data/dossier tier* (early sight of opportunity records and node metrics), governed by a published firewall: investors never influence what gets surfaced, covered, or advanced. State this in the charter.

### 3.6 Spatial layer scope — corridor first, UK parked
The UK prototype is a sunk asset, not a strategy. Year-1 Hamilton/Niagara + WNY needs only a *corridor-scale* registry, built when there's an activation arm to consume it (registry = lead-gen rule). **Recommendation:** port the prototype's method to the corridor as an Observer tool; park the UK instance as a demonstrator/screenshot asset; UK market entry is year-2+ via partnership (04 Track 5).

## 4. Consequence for the Plan Stack (Stage 4 shape)

The sift implies the business-plan step produces four artifacts, not one visionary book:

1. **Clark Charter** (1 page) — nucleus, genetic code, failure-mode rules, firewall
2. **Clark Protocol spec** (~10–15pp) — the 6 core artifacts, ladder with Advancement Contracts, site activation lifecycle
3. **Node One business plan** (grounded) — Hamilton/Niagara theatric bench: A-610-first product, Scenario Zero budget, funding stack + deadlines, host-facility terms re-scoped from the Hamilton draft, cold-start/people plan, 12-month milestones
4. **Vision memo** (short) — the activation-protocol story for partners/funders, built from canonical summary #1 + the inversion + the evidence base (04)

Plus governance hygiene: register adopted components as new entries in the decision log; mark superseded Clark 1.0 decisions (facility-first sequencing, raise framing) for revision rather than silent abandonment.