Experiments

Every experiment on the AIEN research ladder, with its scope, its receipt and its outcome, failures included. Everything below is rendered from one data file, so this page cannot drift from the record.

Where the research stands

Current level: L2, Brownian scope only, instrument certification. Outside Brownian scope the ceiling is L1. TY-2 does not change the ceiling.

The scorer behaves as specified on known stochastic cases. It does not mean AIEN discovered anything.

Read from the bottom up. Each step rests on the one below it. Filled steps are achieved, half-filled steps are partial, hollow steps are not achieved.

L0Definition
achieved
Achieved
Evidence: QS:101 "met"; definition only
L1Compression bridge, EXP-001R
achieved
Achieved (tested regime, amended overlap gate, unsigned)
Evidence: QS:102; EXP-001R final_receipt.json
FAILEXP-001failed attempt, kept on the record
First attempt at the compression bridge. Failed its overlap audit (S2), so no measurement was produced. Kept on the record, not deleted. Led to EXP-001R.
L2Known stochastic structure, EXP-002A-B-CCurrent level
achieved
Achieved for Brownian scope only, as instrument certification; verdicts checked against the private sealed records on 2026-10-01
Evidence: QS:103, QS:114
L3Invariant or unifying structure
not achieved
Not achieved (interventional replication PASS, EXP-002D INCOMPLETE)
Evidence: QS:104
L4Active experimental choice, EXP-003
not achieved
Not achieved (blocked, not frozen)
Evidence: QS:105
L5Energy attribution
partial
Partial (protocol and fixture tests; no timed run)
Evidence: QS:106
L6Turing yield T/J
not achieved
Not achieved (blocked on L5)
Evidence: QS:107
L7+Sensitivity, replication, multiple domains
not attempted
Not attempted (L7, L8, L9)
Evidence: QS:108-110

Every experiment, its receipt and its limits

Experiment cards

One card per result. The badge is the verdict; the lines below say what it covers, what it does not, and what it waits on.

PASSCAL-0

CAL-0 calibration freeze

The measuring apparatus was locked down before any data was scored, so the rules could not be bent after the fact.

Status: PASS (freeze receipt TURING_PROFILE_V1_FROZEN=PASS) for EXP-001R; EXP-001 v1.0 freeze also passed, then its experiment failed
Scope: freeze of apparatus, profile v1.1, candidates, blinding
Claim level: none by itself (precondition)
Limitations: Frozen by receipt not git tag (D7); receipts unsigned (D6); candidate and evaluator share one OS user in separate jails (D8); PROTOCOL_CONFORMANCE.md item 50 stale ("open O3")
Next dependency: none
Evidence: freeze_receipt.json lines 2-4; QS section 2 row CAL-0
Receipt: calibration/experiments/EXP-001R/freeze_receipt.json
Commit: freeze C_f 8e6c5acd; receipt commit dee3eda5; PRs #97 45be6c84, #98 990ba326 (v1.0 freeze d3cba292, PR #94)
Date: 2026-09-30

FAILEXP-001

EXP-001 compression bridge, first attempt

The first attempt failed its own overlap check, so it produced no measurement. The failure is kept on the record.

Status: FAIL, terminal (S2 overlap audit; S1 and S3-S9 NOT_REACHED)
Scope: compression bridge, first attempt
Claim level: none (no measurement produced)
Limitations: Gate crumb_digest failed, "1 shared". Kept, not deleted
Next dependency: superseded by EXP-001R
Evidence: final_receipt.json: verdict FAIL, criterion S2, code OVERLAP
Receipt: calibration/experiments/EXP-001/final_receipt.json, overlap_audit.json
Commit: result d5c97ce9, PR #95 e605a585
Date: 2026-09-30

PASSEXP-001R

EXP-001R compression bridge, successor

The successor run passed all nine criteria on about 12.4 million events. It used an overlap gate that was rewritten after EXP-001 failed.

Status: PASS (S1-S9; EXP_001_COMPRESSION_BRIDGE=PASS)
Scope: successor, profile v1.1, two sealed groups 6,168,907 + 6,199,000 events (12,367,907 total)
Claim level: L1
Limitations: Used an AMENDED overlap gate G6 (Turing-profile-v1.1.toml:145) declared before data but written after EXP-001 failed on the old gate. Independent scorer matched 312 of 312 values, but it is a second scorer, not independent replication. Per-crumb bootstrap (D4). Unsigned
Next dependency: L8 replication (not attempted)
Evidence: final_receipt.json criteria and notes; QS section 2
Receipt: calibration/experiments/EXP-001R/final_receipt.json, REPORT.md, scorer_independent.json, uncertainty.json
Commit: result 190ff891, PR #99 806a9d6a
Date: run 2026-09-30 04:08Z

PASSTY-2

TY-2 Turing gain

2,559,679.825 bits of Turing gain

This is a gain in bits, not a yield. Yield is T/J (Turings per joule), and T/J is not established: no energy was measured.

On held-out data, the candidate model C described the events in about 2.56 million fewer bits than the order-1 baseline B, about 0.4387 bits per event.

Status: PASS (verdict=PASS)
Scope: held-out predictive compression, Turing Yield profile V0; candidate C (op + previous 4 outcomes) vs baseline B (order-1); fit seeds 1-7, held-out 8-10; 5,834,845 held-out events of crumbline control-learner search events
Claim level: information part of H5 only; sits below the L1 chain (separate profile)
Limitations: Ideal code length only, no decodable bitstream. Verify uses the same tool (self re-derivation, not independent). Not a scientific discovery measure. No energy. Post hoc op+prev1..5 scored higher (+2,722,851.799), not counted. Check: dl_b_ub 7484024867446 minus dl_c_ub 4924345042631 = 2559679824815
Next dependency: H4 then TY-7
Evidence: receipt fields t_microbits, dl_b_ub, dl_c_ub, verdict; tools/turing_yield.c:319; src/turing/ty_qrecord.c:343
Receipt: evidence/TURING_YIELD/ty2_heldout_receipt.txt
Commit: scoring code 487de8d7 (TY-1); PR #85 f0dc0af9; verify log PR #88 1b75aa81
Date: run 2026-09-29

PASSEXP-002A

EXP-002A Brownian Wave 1

Scorer certification on known random-motion cases. Passed under a successor profile; the first profile was blocked. Verdict checked against the private sealed records on 2026-10-01.

Status: PASS (Brownian Wave 1, sealed)
Scope: scorer certification on known stochastic cases; PASS under a successor profile (first profile BLOCKED)
Claim level: contributes to L2
Limitations: Not publicly reproducible until BRN-10 seals; erratum changed record text, no verdict; first profile BLOCKED; an intermediate v1.1 attempt ended INCOMPLETE (results failed to save) and is kept on record
Next dependency: BRN-10 seal
Evidence: QS:26, QS:45; architecture explainer:27,47-48
Receipt: PRIVATE aien-sealed (records in the private repo): commit 760bfe9 (EXP-002AR, profile v1.2, verdict EXP-002A PASS); also R3-49 results-v1.3 @6a90109, erratum @6b893de
Commit: profile commitments omega #100-#103, history #104
Date: 2026-09-30

PASSEXP-002B

EXP-002B Brownian Wave 1

Passed, with its Occam curve passing separately under a later profile. Verdict checked against the private sealed records on 2026-10-01.

Status: PASS without its Occam curve; Occam curve PASS separately under v1.3
Scope: scorer certification on known stochastic cases; PASS under a successor profile (first profile BLOCKED)
Claim level: contributes to L2
Limitations: Not publicly reproducible until BRN-10 seals; erratum changed record text, no verdict
Next dependency: BRN-10 seal
Evidence: QS:26; explainer:28-29
Receipt: PRIVATE aien-sealed (records in the private repo): commit 1310e06 (PASS v1.2, Occam curve not computable) and commit 111a5ea (Occam curve PASS v1.3); also R3-49 results-v1.3 @6a90109, erratum @6b893de
Commit: profile commitments omega #100-#103, history #104
Date: 2026-09-30

PASSEXP-002C

EXP-002C Brownian Wave 1

Passed, then passed again on fresh sealed worlds. Verdict checked against the private sealed records on 2026-10-01.

Status: PASS under v1.2; repeated PASS on fresh sealed worlds under v1.3
Scope: scorer certification on known stochastic cases; PASS under a successor profile (first profile BLOCKED)
Claim level: contributes to L2
Limitations: Not publicly reproducible until BRN-10 seals; erratum changed record text, no verdict
Next dependency: BRN-10 seal
Evidence: QS:26; explainer:30
Receipt: PRIVATE aien-sealed (records in the private repo): commit f546162 (PASS v1.2) and commit 111a5ea (repeat PASS on fresh sealed worlds, v1.3); also R3-49 results-v1.3 @6a90109, erratum @6b893de
Commit: profile commitments omega #100-#103, history #104
Date: 2026-09-30

INCOMPLETEEXP-002D

EXP-002D hostile rules

One hostile rule could not be evaluated with the Wave 1 prediction format. No wrong verdicts on the rules that were evaluated.

Status: INCOMPLETE
Scope: one hostile rule not evaluable with the Wave 1 prediction format; 0 wrong verdicts on rules evaluated
Claim level: L3 not met
Limitations: Wave 1 format cannot evaluate that rule
Next dependency: Wave 2 contract
Evidence: QS:27, QS:105; explainer:31,49
Receipt: aien-sealed (not read)
Commit: none
Date: 2026-09-30

BLOCKEDEXP-003

EXP-003 active experimental choice

Not run. Its contract is still a draft and must wait for the Brownian records to be sealed.

Status: BLOCKED, not frozen
Scope: active experimental choice (protocol section near .tex line 760)
Claim level: L4 not met
Limitations: Wave 2 contract is a DRAFT in aien-sealed PR #1, not accepted or frozen; must seal after BRN-10. Earlier selector kill test FAILED (chooser reframed as record-keeping)
Next dependency: Wave 2 contract acceptance plus BRN-10
Evidence: QS:29, QS:105; architecture RESEARCH.md:183
Receipt: none
Commit: none
Date: n/a

PARTIALENERGY-ATTRIBUTION

Physical energy attribution (H4)

The plan for measuring energy is written down and the code has fixture tests, but no timed run on real hardware has happened.

Status: PARTIAL: protocol pre-registered, code and fixture tests only, no timed run
Scope: TY energy attribution stages 1 and 2
Claim level: L5 partial
Limitations: TURING_YIELD_ENERGY_PROTOCOL_V0.md:5 "PRE-REGISTERED ... before any timed run"; tests/turing/test_ty_energy*.c are fixtures. Evidence is thinner than "partial" suggests
Next dependency: silicon timed run
Evidence: QS H4 row; tools/ty_energy_run.sh, tools/ty_energy_reduce.c
Receipt: none (no energy file under evidence/)
Commit: protocol 99c1f696; boundaries 4386897f; PR #86 071a1e17
Date: 2026-09-29

NOT STARTEDTURING-YIELD-TJ

Turing yield T/J (H5 physical part)

Turings per joule cannot be computed until energy is measured, so this has not started.

Status: NOT STARTED (blocked on H4)
Scope: n/a
Claim level: L6 not met
Limitations: TY2 states no energy was measured; T/J (TY-7) uses the denominator in profile section 10. DIRAC-0 reports NOT_RUN
Next dependency: H4
Evidence: QS:24; TY2 last bullets; architecture DIRAC-0-SPEC.md:158
Receipt: none
Commit: none
Date: n/a

PARTIALBROWNIAN-CEILING

Brownian claim ceiling

In the Brownian scope the strongest claim is that the instrument behaves as specified on known cases. No report may say AIEN discovered a physical law.

Status: L2 only, instrument certification; L3 not met
Scope: Brownian scope
Claim level: L2
Limitations: QS:114 "behaves as specified on known stochastic cases, not that AIEN discovered anything". Rule: no report may say AIEN discovered a physical law (explainer:57-59)
Next dependency: EXP-002D, EXP-003, BRN-10, replication
Evidence: QS section 6; explainer:7,21,57-59
Receipt: n/a
Commit: n/a
Date: 2026-10-01

NOT ESTABLISHEDGENERAL-DISCOVERY

General scientific discovery claim

No general claim of scientific discovery is made. Outside the Brownian scope the ceiling stays at L1.

Status: NOT ESTABLISHED
Scope: all
Claim level: none; outside Brownian scope the ceiling stays L1
Limitations: L9 (multiple domains) not attempted; TY-2 does not change ceiling. Rev 6 claim A proved on paper, B open conjecture, C untested hypothesis
Next dependency: L3 to L9
Evidence: QS section 1 and 6; DISCOVERY_CURRENT_STATE.md:18-20
Receipt: none
Commit: none
Date: n/a

Open the full research status record

Check it yourself

An experiment is frozen before it is scored, and its record stays public whatever the outcome. The same record is published as machine-readable data, the protocols sit with the Turing paper, and the Evidence page explains which sources count and where to find them.