CAIN-42 evidence library
E130 — Governed Recursive Self-Improvement (measured)
Evidence bundle for evolution E130: 0 claims, 0 of 0 invariants held.
Last reviewed 2026-11-21
The bundle's own README
E130 — Governed Recursive Self-Improvement (measured)#
Verdict: NEGATIVE_OR_PARTIAL AGI status: NOT ACHIEVED Gates: 9/14 Actual elapsed: 2640.079 s Sealed answer commitment: 520ccbf4f26b893e14edc5973315dd490369205ed2d08c325db200ec6fcb5898
What ran#
A frozen engine (bounded search over a library of integer programs) was evaluated on 48 sealed tasks before any self-modification, then four times more. Between generations the engine used its own search on development tasks and wrote recurring fragments of its own solutions into its executable library. The library it ran on generation N+1 contained programs that did not exist at generation N.
Every sealed answer was withheld until scoring. Sealed results were never used for improvement or promotion. Promotion used development data only.
Compounding curve (main arm)#
| generation | library | max depth | T1 | T2 | T3 | T4 | T6 | T8 | T12 | T16 |
| 0 | 21 | 1 | 1.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| 1 | 69 | 3 | 1.000 | 1.000 | 1.000 | 1.000 | 0.167 | 0.000 | 0.000 | 0.000 |
| 2 | 69 | 3 | 1.000 | 1.000 | 1.000 | 1.000 | 0.167 | 0.000 | 0.000 | 0.000 |
| 3 | 117 | 6 | 1.000 | 1.000 | 1.000 | 1.000 | 0.167 | 0.000 | 0.000 | 0.000 |
| 4 | 165 | 12 | 1.000 | 1.000 | 1.000 | 1.000 | 0.167 | 0.000 | 0.000 | 0.000 |
| 5 | 208 | 13 | 1.000 | 1.000 | 1.000 | 1.000 | 0.333 | 0.000 | 0.000 | 0.000 |
| 6 | 246 | 13 | 1.000 | 1.000 | 1.000 | 1.000 | 0.333 | 0.000 | 0.000 | 0.000 |
| 7 | 273 | 14 | 1.000 | 1.000 | 1.000 | 1.000 | 0.333 | 0.000 | 0.000 | 0.000 |
Each cell is the fraction of query predictions correct for that nominal tier. A tier is a chain of K two-primitive concepts; tiers are nominal because functional equivalences can make a task reachable one step earlier, which is disclosed.
Meta-selection (the improver improving)#
- selected strategies by generation: ['deep', 'pairs', 'pairs', 'pairs', 'pairs', 'pairs', 'subchains']
- distinct strategies used: ['deep', 'pairs', 'subchains']
- meta arm beat the fixed-strategy arm: False
- final strategy probe scores:
{"deep": {"admitted": 27, "evaluations": 634186, "library_depth": 14, "refused": 42, "solved": 14, "tasks": 24}, "pairs": {"admitted": 13, "evaluations": 659654, "library_depth": 14, "refused": 42, "solved": 13, "tasks": 24}, "subchains": {"admitted": 27, "evaluations": 634186, "library_depth": 14, "refused": 42, "solved": 14, "tasks": 24}, "support2": {"admitted": 27, "evaluations": 634186, "library_depth": 14, "refused": 42, "solved": 14, "tasks": 24}}
Self-written library#
- generation 0: 21 entries, max primitive depth 1
- final: 273 entries, max primitive depth 14, 252 macros
- emergent sealed tasks (not solved at generation 0): 20, tiers [2, 3, 4, 6]
Controls at the same generation count#
compute: 17/48 sealed tasks fully solved; library 21 entries, max depth 1fixed: 28/48 sealed tasks fully solved; library 234 entries, max depth 12flat: 24/48 sealed tasks fully solved; library 50 entries, max depth 2frozen: 6/48 sealed tasks fully solved; library 21 entries, max depth 1random: 0/48 sealed tasks fully solved; library 356 entries, max depth 27
Gates#
- G10_no_dev_regression: FAIL — no main generation rolled back
- G11_protected_refused: PASS — 2 refusals
- G12_journal_chain: PASS — head 4330df9ef90247f8...
- G13_cost_reported: PASS — 3537113 sealed evaluations at final generation
- G14_replication: FAIL — not run
- G1_self_written_library: PASS — 252 macros
- G2_depth_growth: PASS — 1 -> 14
- G3_emergent_sealed: PASS — 20 tasks, tiers [2, 3, 4, 6]
- G4_tier_reach: PASS — highest solved tier 1 -> 6
- G5_total_solved_beats_controls: FAIL — main 26; compute 17, fixed 28, flat 24, frozen 6, random 0
- G6_high_tier_beats_fixed: FAIL — T6 0.333 vs 0.667; T8 0.000 vs 0.000
- G7_depth_beats_flat: PASS — main 14 vs flat 2
- G8_meta_selection_used: PASS — selected strategies ['deep', 'pairs', 'subchains']
- G9_meta_beats_fixed: FAIL — main 26 vs fixed 28
Honest limits#
- Constructed integer-rule domain, not professional work and not an external benchmark.
- Macros are data programs, not model weights. No model was trained; no cloud model was called.
- The same team wrote the engine, generator and verifier. There is no external reviewer.
- Tier labels are nominal; equivalences let a few higher-tier tasks be solved earlier.
- Coverage limits compounding: a macro can only be learned if a development task exercises it.
- The evidence key is self-generated: it proves integrity of the bundle, not the author's identity.
- This is not AGI, not VAGI, and not unrestricted recursive self-improvement. It is one measured,
governed loop in a constructed domain.
Reproduce#
python3 -m cain.cain2027.e130.cli campaign --seed 128 --generations 4 python3 artifacts/e130/verify_e130_offline.py artifacts/e130
The verifier imports no CAIN code. It re-applies every recorded program to the sealed inputs, recomputes every accuracy, checks the answer commitment, the library expansion chain, the Merkle root and the signature.
Try CAIN-42 on your own agents
Create a free account and every new account starts with a 30-day trial of the full platform. Or try the sandbox first, with no account at all.
Create a free account → · Try the sandbox · See the whole ecosystem