# E131 — Governed Recursive Self-Improvement (measured)

**Verdict:** PARTIAL_RUN: controls missing; no control claim
**AGI status:** NOT ACHIEVED
**Gates:** 10/17
**Actual elapsed:** 649.038 s
**Sealed answer commitment:** `6bfca015fe91123abd94d682ae171fd4ae75f33b48a3da09c3133dc3ca0c5acd`

## 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 | 0.167 | 0.167 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| 1 | 69 | 3 | 1.000 | 1.000 | 0.983 | 1.000 | 0.333 | 0.000 | 0.000 | 0.000 |
| 2 | 69 | 3 | 1.000 | 1.000 | 0.983 | 1.000 | 0.333 | 0.000 | 0.000 | 0.000 |
| 3 | 69 | 3 | 1.000 | 1.000 | 0.983 | 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.

## 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]]` (total budget 440000)
- portfolio arm solved: **8** vs fixed-plan **27** vs compute-matched **6**
- portfolio beat fixed and compute: **False** / **True**

## Meta-selection (the improver improving)

- selected strategies by generation: ['subchains', 'pairs', 'pairs']
- distinct strategies used: ['pairs', 'subchains']
- meta arm beat the fixed-strategy arm: **False**
- final strategy probe scores: `{"deep": {"admitted": 48, "evaluations": 890583, "library_depth": 9, "refused": 0, "solved": 10, "tasks": 24}, "pairs": {"admitted": 48, "evaluations": 714083, "library_depth": 6, "refused": 0, "solved": 15, "tasks": 24}, "subchains": {"admitted": 48, "evaluations": 867370, "library_depth": 9, "refused": 0, "solved": 13, "tasks": 24}, "support2": {"admitted": 48, "evaluations": 867370, "library_depth": 9, "refused": 0, "solved": 13, "tasks": 24}}`

## Self-written library

- generation 0: 21 entries, max primitive depth 1
- final: 69 entries, max primitive depth 3, 48 macros
- emergent sealed tasks (not solved at generation 0): **17**, tiers [2, 3, 4, 6]

## Controls at the same generation count

- `compute`: 6/48 sealed tasks fully solved; library 21 entries, max depth 1
- `fixed`: 27/48 sealed tasks fully solved; library 97 entries, max depth 4
- `frozen`: 6/48 sealed tasks fully solved; library 21 entries, max depth 1
- `plan`: 8/48 sealed tasks fully solved; library 21 entries, max depth 1

## Gates

- **G10_no_dev_regression**: FAIL — no main generation rolled back
- **G11_protected_refused**: PASS — 2 refusals
- **G12_journal_chain**: PASS — head 25b6db582b5f935e...
- **G13_cost_reported**: PASS — 9990714 sealed evaluations at final generation
- **G14_replication**: FAIL — not run
- **G15_portfolio_used**: PASS — main search modes ['portfolio']
- **G16_portfolio_beats_fixed**: FAIL — portfolio arm 8 vs fixed-plan arm 27
- **G17_portfolio_beats_compute**: PASS — portfolio arm 8 vs compute-matched arm 6
- **G1_self_written_library**: PASS — 48 macros
- **G2_depth_growth**: PASS — 1 -> 3
- **G3_emergent_sealed**: PASS — 17 tasks, tiers [2, 3, 4, 6]
- **G4_tier_reach**: PASS — highest solved tier 3 -> 6
- **G5_total_solved_beats_controls**: FAIL — main 25; RSI controls compute 6, frozen 6; fixed-strategy 27; adaptive-plan-only 8
- **G6_high_tier_beats_fixed**: FAIL — T6 0.333 vs 0.667; T8 0.000 vs 0.000
- **G7_depth_beats_flat**: FAIL — main 3 vs flat 0
- **G8_meta_selection_used**: PASS — selected strategies ['pairs', 'subchains']
- **G9_meta_beats_fixed**: FAIL — main 25 vs fixed 27

## 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

```bash
python3 -m cain.cain2027.e131.cli campaign --seed 128 --generations 4
python3 artifacts/e131/verify_e131_offline.py artifacts/e131
```

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.
