ZRW lab series · public zero-run path
ChainPipe → IndexMind → CrewHive: ship verifiable zeros first. Flagship seat = int32 zeros×10k → 8 B via live process rail (beats gzip-9 73 · brotli-11 13). Teaching sketch = pure-zero byte RLE (not the 8 B int seat).
POST /api/process · corpus zeros · n=10000
idle — click Run live
Public residual sketch · zrw_zero_run_public.py
idle — click Run sketch
Gate 2 PASS · frozen bench ZRW_ZEROS_10K_BENCH_LATEST.json · 8 B < brotli-11 13 · gzip-9 73 · suite first 2 GB/month then 8¢/GB on /pps.
Public verify harness (no install — hits live rail): /zrw_verify.py · corpus /zeros10k.bin · ledger ZRW_VERIFY_HARNESS_LATEST.json · python3 zrw_verify.py zeros10k.bin → zrw=8 gz=73 · GRADE PASS. Next gate: float · text · delta · high_entropy · general under lattice.
Domain rankings
Lab rank within each field (1 = best among measured set)
Active records
Metric scorecard
| Metric | Lab | Lab value | Field | Rank | Status | As of |
|---|
Industry field (top 12 roles)
Context — not a claim we beat all of these on every corpus.
| # | Codec / class | Role | Strength |
|---|