Network/Journal/Growing The Embryos Civic Layer To Product
All entries
Entry 032

Growing the embryos — bridging the civic layer to the product

Date
2026-06-05
Status
Decided
Authority
Creator

Decision

A full gap analysis (Journal + Register vs. the actual codebase) surfaced a clear shape: what ships today is the civic layer — Journal, Register, Constitution, and Roadmap rendered from markdown, with auth, reactions, inline annotations, and Amplitude event tracking. The product platform — Fact graph, world-model wiki, Atlas profiles, the contribution economy — is richly specified but unbuilt.

Two decisions follow.

1. Grow the embryos, don't greenfield. Several decided surfaces already exist in primitive form in the build; the fastest, lowest-risk bridge to product is to evolve them rather than start fresh:

  • Annotations → LoreLines (Entry 010): the inline-comment layer is a document-scoped LoreLine. Extend its target from documents to entities/Facts.
  • Reactions → Fact signal (G-036, G-059): support/concern on entries becomes the low-cost quality signal on Facts.
  • profiles table → Atlas profile (Entry 027): the identity row exists; the contributor profile page — members as graph entities — does not.
  • amplitude.track → Experiment harness (Entry 031, G-059): event tracking exists; the flag/variant/holdout plumbing the A/B thesis assumes does not.

2. Build the connective tissue decided systems silently assume. The analysis found load-bearing surfaces that no single entry owns because they sit between decided systems. They become new Register gaps:

  • Unified Fact review/confirm queue (G-060) — the keystone. It is the single surface where the contribution pipeline (G-013, G-057), the consensus / anti-Goodhart confirmation mechanic (G-059), moderation, and data quality (G-026, G-012) all converge. Prioritized first among the product surfaces.
  • Notification & activity system (G-061) — silently required by annotation replies (already built, currently un-notified), grace-period expiry (Layer 10), contribution status, and the gamification impact-echo.
  • Prose↔Fact linking semantics (G-062) — the make-or-break primitive hiding inside G-057: how prose binds to Facts (markup vs. AI-detected vs. coreference).
  • Entity search relevance & ranking signals (G-063) — the mechanism under the existential discoverability bet (G-055, G-010); no entry yet defines the signals.
  • Experiment & feature-flag harness (G-064) — the build prerequisite for every A/B claim in Entry 031.
  • Payment & payout infrastructure (G-065) — absent entirely, yet silently blocking every monetization promise (G-016, G-038, Entry 005).

Priority ordering. Build the experiment harness (G-064) first — it is small and unblocks the entire experimentation thesis. The review/confirm queue (G-060) is the flywheel core. Profiles + notifications (G-061) are the identity and feedback loop that make recognition and impact-echo real.

Reasoning

Growing embryos compresses time and risk: each has working data, RLS, and UI patterns to extend rather than invent. The connective-tissue gaps are dangerous precisely because they are invisible — each decided system assumes them, so no entry owns them, and they would surface as launch-blockers. The review/confirm queue is named the keystone because one surface discharges four obligations at once (pipeline + gamification + moderation + quality), and because its core gesture is the "confirm-is-a-swipe" mechanic from G-059.

The analysis also confirmed two prior flags: data portability is registered (G-019) but lacks an actual design, and the A/B harness is assumed everywhere yet exists nowhere in code.

Open threads

  • Unified Fact review/confirm queue (G-060)
  • Notification & activity system (G-061)
  • Prose↔Fact linking semantics (G-062)
  • Entity search relevance & ranking signals (G-063)
  • Experiment & feature-flag harness (G-064)
  • Payment & payout infrastructure (G-065)
  • Data portability (G-019) — needs a design pass, not just an entry.