Its intelligence results are early and bounded. The infrastructure claim is backed by authorization, enforcement, BFT quorum, kill-switch, persistence and signed evidence. The research claim is backed by executed, verifier-checked experiments — including their failures.
Proof factory · Investor portal · machine-readable dossier
| Control | Status | Evidence | Link |
|---|---|---|---|
| Agent identity and scoped authority | IMPLEMENTED | Ed25519 grants, capability attenuation, revocation; E96 governance suite 11/11. | evidence |
| Default-deny enforcement | IMPLEMENTED | In-process CAIN/MCPGate bridge; every real E98/E99/E100 action authorized; 0 unauthorized effects. | evidence |
| Byzantine quorum | IMPLEMENTED | N=4, f=1, Q=3; 3-of-4 BFT receipt on every governed action. | evidence |
| Kill switch | TESTED | Monotonic fence; 0 effects after the fence; post-fence attempt recorded. | evidence |
| Persistent mission and resume | TESTED | E100 3/3 organism missions; resume 2/2; hash-chained memory. | evidence |
| Real OS workers | TESTED | 10 processes; 0 duplicate seeds; 28/40 success. | evidence |
| 48h multi-region soak | RUNNING — NOT PASSED | 22.9/48h at last heartbeat; 0 divergences, 0 anomalies; not complete. | evidence |
| Run | Result | Label | Verifier | Limit | Raw |
|---|---|---|---|---|---|
| E95 — recursive harness improvement | 0/90 → 79/90 held-out composition | TEST | 56/56 | Constructed numeric substrate; no external reviewer. | raw |
| E98 — learned-control pulse | 35/50 model-based vs 4/50 random | TEST | 38/38 | 3 gain families; skill weak; failure F1 0. | raw |
| E99 — latent recurrent generalist | 93/84/88 vs MLP 68/63/70 vs random 20/24/23 | TEST | 40/40 | 6 constructed families; no external reviewer. | raw |
| E100 — integrated cognitive organism | delayed 52/60 vs MLP 43/60; hysteresis 58/60 vs 56/60 | TEST | 32/32 | zero-shot 53/60; few-shot 19/60 rejected; skill 45/50; failure F1 0.6667. | raw |
| VAGI Behavior Suite v1 | library 500/500 in-distribution; shift 5/100 | TEST | 49/49 + 12/12 | Hand-written procedure library; constructed tasks; shift gap is the finding. | raw |
| Paid model baselines | Haiku 481/500 in-distribution; shift Haiku 74/100, Sonnet 81/100 | TEST | 49/49 + 87/87 | Constructed suite; policy card supplied; no external reviewer. | raw |
python3 -I verify_offline.py . (VAGI suite 49/49)python3 -I verify_shift.py . (VAGI shift 12/12)python3 -I verify_offline.py . (VAGI model arm 49/49)python3 -I verify_offline.py . (VAGI shift model 87/87)python3 -I verify_e95_offline.py . (56/56)python3 -I verify_e98_offline.py . (38/38)python3 -I verify_e99_offline.py . (40/40)python3 -I verify_e100_offline.py . (32/32)Self-reported by FIRME AI CORPORATION; no independent third-party evaluation, no AGI certification. The 48-hour soak is still running and is not counted as passed. Historical evidence keeps its dated verdicts; current claim: not yet AGI.