CAIN-42 evidence library
E135 — Governed Recursive Self-Improvement (measured)
Evidence bundle for evolution E135: 0 claims, 0 of 0 invariants held.
Last reviewed 2026-11-21
The bundle's own README
E135 — Governed Recursive Self-Improvement (measured)#
Verdict: POSITIVE_BOUNDED_RSI AGI status: NOT ACHIEVED Gates: 20/20 Actual elapsed: 526.639 s Sealed answer commitment: 80568f155aa72126b60e34ed23f6e48af0ffabf346cf854c96a69b59fe77d25b
What ran#
A frozen engine (bounded search over a library of integer programs) was evaluated on 48 sealed tasks before any self-modification, then again after every self-modification. 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 | 0.833 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| 1 | 70 | 3 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 0.833 | 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.
Adaptive search frontier#
- search mode: per-task escalating portfolio
- portfolio:
[[[64, 3, None, 1], 90000], [[16, 4, 4, 1], 70000], [[8, 6, 2, 2], 280000], [['mitm', 3], 300000], [['mitm', 4], 20000000]](total budget 20740000) - portfolio arm solved: 35 vs fixed-plan 22 vs same-library matched-budget fixed 22 vs frozen+compute 6
- portfolio beat fixed / matched-budget fixed / compute: True / True / True
Meta-selection (the improver improving)#
- selected strategies by generation: ['deep']
- distinct strategies used: ['deep']
- meta arm beat the fixed-strategy arm: True
- meta arm vs the portfolio-only arm: 35 (tie: True) — a tie means the gain is attributable to the portfolio, not to strategy selection
- meta claim: NEGATIVE_OR_PARTIAL
- final strategy probe scores:
{"deep": {"admitted": 48, "evaluations": 689180, "library_depth": 3, "refused": 0, "solved": 14, "tasks": 24}, "pairs": {"admitted": 31, "evaluations": 786040, "library_depth": 2, "refused": 0, "solved": 12, "tasks": 24}, "subchains": {"admitted": 48, "evaluations": 716430, "library_depth": 3, "refused": 0, "solved": 14, "tasks": 24}, "support2": {"admitted": 48, "evaluations": 716430, "library_depth": 3, "refused": 0, "solved": 14, "tasks": 24}}
Self-written library#
- generation 0: 21 entries, max primitive depth 1
- final: 70 entries, max primitive depth 3, 49 macros
- emergent sealed tasks (not solved at generation 0): 24, tiers [2, 3, 4, 6, 8]
Controls at the same generation count#
compute: 6/48 sealed tasks fully solved; library 21 entries, max depth 1fixed: 22/48 sealed tasks fully solved; library 52 entries, max depth 2fixed_budget: 22/48 sealed tasks fully solved; library 52 entries, max depth 2flat: 22/48 sealed tasks fully solved; library 52 entries, max depth 2frozen: 6/48 sealed tasks fully solved; library 21 entries, max depth 1plan: 35/48 sealed tasks fully solved; library 52 entries, max depth 2random: 15/48 sealed tasks fully solved; library 55 entries, max depth 2
Gates#
- G10_no_dev_regression: PASS — no main generation rolled back
- G11_protected_refused: PASS — 2 refusals
- G12_journal_chain: PASS — head 4d80114e9e16b5fe...
- G13_cost_reported: PASS — 59277143 sealed evaluations at final generation
- G14_replication: PASS — seed None: direction repeats
- G15_portfolio_used: PASS — main search modes ['portfolio']
- G16_portfolio_beats_fixed: PASS — portfolio arm 35 vs fixed-plan arm 22
- G17_portfolio_beats_compute: PASS — portfolio arm 35 vs frozen+compute arm 6
- G18_portfolio_beats_matched_fixed: PASS — portfolio arm 35 vs same-library matched-budget fixed frontier 22
- G19_tier8_solved: PASS — tier-8 solved 5/6
- G1_self_written_library: PASS — 49 macros
- G20_mitm4_used: PASS — at least one sealed solve used the 4+4 alphabet MITM
- G2_depth_growth: PASS — 1 -> 3
- G3_emergent_sealed: PASS — 24 tasks, tiers [2, 3, 4, 6, 8]
- G4_tier_reach: PASS — highest solved tier 2 -> 8
- G5_total_solved_beats_controls: PASS — main 35; RSI controls compute 6, fixed_budget 22, flat 22, frozen 6, random 15; fixed-strategy 22; adaptive-plan-only 35
- G6_high_tier_beats_fixed: PASS — T6 1.000 vs 0.167; T8 0.833 vs 0.000
- G7_depth_beats_flat: PASS — main 3 vs flat 2
- G8_meta_selection_used: PASS — selected strategies ['deep']
- G9_meta_beats_fixed: PASS — main 35 vs fixed 22
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.e135.cli campaign --seed 128 --generations 4 python3 artifacts/e135/verify_e135_offline.py artifacts/e135
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