The Slow Flywheel: Spec First, Crowd Later

What this means for you

GRACE will not launch with a hero video. It launches with a spec. The flywheel is six stages long, and each stage is blocked by the previous one shipping in full. The order is not negotiable, because category-defining work loses if it skips ahead.

The pitch

Publish the spec, then post Show HN with a design journal entry on the same day. Earn the first hundred stars on the strength of the work. File the arXiv preprint sixty days later. Wait for the first external implementation before claiming a category exists.

Who it’s for

The advisor or peer asking why there is no launch date pinned yet.

Proof points

• Six-stage flywheel is written down in architecture/MARKETING-KPI-TRACKER.md, each stage explicitly blocked by the prior one
• Year-one targets, dated: spec v1.0 by 2026-Q2, 500 GitHub stars by 2026-Q4, arXiv preprint by 2026-Q3
• Show HN target: top five for six hours, with 100+ upvotes, on the same day as the spec publishes
• Year-one design journal target: 10 entries with 5,000+ reads, monthly cadence

timeline
  title Six-stage GRACE flywheel
  Stage 1 : Publish the spec : SPEC-01..05 + CLI
  Stage 2 : Show HN + journal : same-day post
  Stage 3 : First 100 stars : compounding signal
  Stage 4 : arXiv preprint : 60 days post-launch
  Stage 5 : First external tool : built by someone else
  Stage 6 : Category named : "GRACE-compatible" used by others