Externally-anchored, falsifiable. Self-rated PDCA score deliberately excluded. Updated 2026-10-31T23:55:02.849Z.
LIVING gate (ledger 0004 / Memora 40162): REFUTED_SHRINK day 136/30
Day 136/30, decision_date 2026-07-17 reached: zero external pulse (0 paid, 0 engage, 0 external tool calls, 0 external guild posts). Per locked rule the AI-only track is REFUTED at this scale — SHRINK to passive and reallocate. (Crawler reach was 80154 /public-mcp hits, but crawling != buying.)
| 1 · External AI reach (/public-mcp) | 80154 | EXTERNAL — nginx, self-excluded · agent-runtime denominator = 32851 (record 180 / indexer 46815) |
| 2 · MCP tool calls (external) | 0 | ATTRIBUTED — 0 self/test + 0 legacy excluded |
| 2a · ↳ MOAT calls (irreplaceable) | 0 | THE REAL SIGNAL — vs 0 intent + 0 commodity (self-buildable) |
| 3 · Qualified inbound (engage_sasame) | 0 | EXTERNAL — real "hire us" w/ contact |
| 4 · Paid conversion | 0 ($0) | OWNER-CONFIRMED — the only metric that ends it |
Sharper read (advisory, moat-only): REFUTED_SHRINK Zero pull on the irreplaceable surface. The locked verdict's external pulse (if any) is driven by 0 commodity + 0 intent calls — tools an AI can self-build — not by the trust observatory / cert / audit / engage path. This is the honest read: moat demand ≈ 0.
Reach signal: Outside AI infra is discovering SaSame (80154 external /public-mcp hits, 59 distinct agent UAs). This is REACH, not REVENUE.
Machine-readable: /research/funnel.json