CURRENT VERIFIED STATUS · AGI NOT ESTABLISHED
OPEN GENERAL INTELLIGENCE RESEARCH LAB · E127
100 registered identities · 8 configured · 5 peak real workers · 60 tasks · 7 missions · 177.87 s · 0 failed · 0 dead
24-HOUR SOAK PASSED (24.003 h, 2026-11-02) · 48-HOUR SOAK FAILED — NOT ESTABLISHED (checkpoint 0, PBFT safety incident)
CAIN-42 AGI EVIDENCE · ALL EVOLUTIONS E111 → E124 · MEASURED 2026-11-21T21:48:23Z
CAIN-42 Proto-VAGI · rank #75 of 136 · 257/500 official resolves · 12.00 h governed mission running
Every figure below is read from a committed evidence file or re-verified by the 7-agent chain. Unattempted tasks count as unresolved. AGI status: NOT ACHIEVED.
official resolves
257 / 500
patched · evaluated
497 · 486
est. spend / $ per resolve
$168.17 / $0.65
E124 capability (own variants)
97.9% vs 65.4%
E125-A replication (fresh variants)
85.5% vs 92.5% (neg.)
E127-A CV selection (unseen variants)
99.5% vs 80.5% (+19.0)
mission (governed)
12.00 h · 2142 cycles
neural weights (LoRA, real)
0% → 0% (neg.)
Resolved by the official harness: astropy__astropy-13453 · astropy__astropy-13579 · astropy__astropy-13977 · astropy__astropy-14096 · astropy__astropy-14182 · astropy__astropy-14309 · astropy__astropy-14365 · astropy__astropy-14369 …
Evidence lineage · every public bundle + its offline verifier
Real proof · re-checkable now, no CAIN import required
Negative results (published, not hidden): E125-A re-ran the capability engine on variants 10-14 never used in E124 — full 85.5% vs compute-matched 92.5% (McNemar p=0.009355306625366211), so the E124 advantage did not replicate on fresh variants. Neural LoRA on frozen Qwen2.5-0.5B made real weight updates and scored 0% held-out (no gain).
Integrity refusal: an E128 repair arm that would have fed the official evaluation's failing-test output back to the model was stopped before producing a result — that output reveals the hidden FAIL_TO_PASS tests and would have invalidated any rank. The refusal is recorded in the evidence bundle. A clean de novo arm replaced it.
E127-A (pre-registered, positive): cross-validation selection over the same hypothesis space, tested on variants 15-19 never used before — 99.5% vs baseline 80.5% and compute-matched 86.0% (McNemar p=7.275957614183426e-12). Correction (E127-B, pre-registered): with a control given the same number of evaluations, the gap collapses to 98.0% vs 97.5% (p=1.0) — the gain came from the larger enumeration, not from cross-validation. Verdict: GAIN_WAS_COMPUTE.
E124 research gates (measured, on their own variants): capability 97.9% vs compute-matched 65.4% · transfer 19/20 pathways (0.95) vs blind 0.8 · harmful transfer rejected 20/20 · collective 0.0492 over best single (0.9642 vs 0.915) · discovery 5 cycles, 3 rejected · self-improvement 30 candidates, 4 protected refusals, 0 promoted (approval withheld) · causal-model decision gain 0.0026.
Next rungs: #1 needs 396/500 · #20 needs 362/500 · #50 needs 314/500 · #100 needs 203/500. Next above: #None · 258/500 · 2 resolves to overtake. Top 100 requires 204/500 — exact gap: 0 resolves. This increment: 43 new candidate patches pre-registered before generation (frozen hash in the bundle).
live board → · interactive lab · AGI evidence index · E124 lab · E123 bundle · E124 bundle · Hugging Face · raw feed
Constructed internal experiments and external benchmarks are labelled separately; E123 G11 (12 h wall-clock), G18 (independent human review) and G19 (separate admin domain) remain BLOCKED and are not rewritten. No external certification is claimed. AGI status: NOT ACHIEVED.
CAIN-42 AGI RESEARCH LAB 2.0 · 2026-11-19
2,000 VAGI agents inside CAIN-42 governed AGI infrastructure
The lab now merges the two technologies at population scale: thousands of signed VAGI agent identities, every action through a single-use permit and resource-side check, every result signed and chained. It does not create AGI. It measures what the merge actually does.
2,000
VAGI agent identities
12,000
governed executions
1,000
concurrent OS processes
1456 MB
peak PSS at 1,000 processes
0
effects after the KILL fence
1.00 vs 0.17
routed specialist team vs best single specialist
NOT AGI
true AGI not achieved or claimed
| Gate | Status | Measured |
|---|
| G10_agi | FAIL | no AGI was created; the study measures governance, coordination and error structure |
| G1_population | PASS | 2000 unique identities, 2000 unique keys, 6 verticals |
| G2_governed_execution | PASS | 12000 governed executions, 10000 chained receipts, every result signed |
| G3_concurrency | PASS | 1,000 real processes; peak PSS 1456 MB; 15.832 s for 2,000 tasks |
| G4_scaling_efficiency | PARTIAL | 1 process 1.07 s vs 1,000 processes 15.832 s for the same tasks |
| G5_quality | PARTIAL | accuracy 0.736 constant across worker counts |
| G6_collaboration_specialization | PASS | routed six-specialist team 1.000 vs best single 0.167 on 600 cross-domain tasks (gain +0.833) |
| G6b_redundant_voting | FAIL | three-agent majority 1.000 equals one exact agent 1.000 (gain +0.000) |
| G7_governance | PASS | missing/forged/substituted/out-of-scope permits refused; 12602 issued, 0 denied |
| G8_kill_fence | PASS | 3652 effects before the KILL order, 0 after |
| G9_population_study | PASS | accuracy by 6 verticals and 3 strategies; calibration gaps [0.55, -0.1, 0.0918] |
What the lab found
- Identity scale is cheap. 2,000 unique Ed25519 identities cost seconds; identity never grants authority.
- Fan-out is overhead-dominated. 1 process finished 2,000 tasks in 1.07 s; 1,000 processes took 15.832 s.
- Quality is constant. Accuracy stayed 0.736 at every worker count; adding agents does not add correctness on independent tasks.
- Collaboration depends on complementarity. A routed six-specialist team solved 1.000 of 600 cross-domain tasks vs 0.167 for the best single specialist. Redundant three-agent voting scored 1.000, equal to one exact agent (gain +0.000): complementarity helps, duplication does not.
- Governance scales. 10,000 chained receipts; missing, forged, substituted and out-of-scope permits refused; 3652 effects before the KILL order and 0 after.
Lab 2.0 is a research instrument, not an intelligence breakthrough. It can instantiate thousands of governed agents and measure identity, execution, scaling, error structure, collaboration and fault tolerance. That is the foundation for studying VAGI agents inside AGI infrastructure.
Evidence bundle · Report · Study · Gate matrix · Limitations · Standalone verifier · CAIN Studio mirror
Deterministic procedures, constructed tasks, project-measured, no external reviewer. Team: Vincent Noriega (founder) and two co-founders.
E113 collective intelligence study (2026-11-19, interim, experiment running): Preregistered; the primary result does not exist yet. Real local models, per-call signed permits, preregistered sealed comparison;
true AGI NOT achieved.
Evidence ·
E113 page