research: exact QQ16 consumer budget obstruction under centre-64 law

[1337isnot1337]

Summary

Adds research/qq16-consumer-cap-audit/, an exact independent audit of the recursive-price budget for the pinned CONSERVED-QQ16 F0 + B_1_2 pair and #2707 decoder under the unchanged centre-64 coverage law.

The source-pinned F/B profiles average to 2,428 rank-one children per call. The decoder contributes 336 rank-one children across two calls, producing the #3181 profile {1:2596,2:418,3:432,4:142,5:82,8:8}. At the selected kappa = 836330098538459/500000000000000000, the verifier removes the eight rank-eight children and relaxes their 64 mass units into rank-51 idle children. Even this favorable relaxation has an exact rational moment lower bound above 100009/100000 at the selected exponent.

This is a scoped necessary-budget obstruction for that fixed profile and coverage law. It does not establish a physical F17 consumer, an exact cap root, a global barrier, a changed-coverage result or a new selected kappa.

Verification

python3 -B research/qq16-consumer-cap-audit/verify.py
python3 -B scripts/research_package.py --package research/qq16-consumer-cap-audit --check-only
python3.12 -B scripts/research_package.py --package research/qq16-consumer-cap-audit

All three commands passed locally at base 20ecf2891d80b5acf18d980b2164fb44ea7faa90. The registered replay reports PASS_REGISTERED_RESEARCH_PACKAGE research/qq16-consumer-cap-audit outcome=completed.

Review focus and limits

Please check the profile reconstruction, the exact atanh/log and positive Taylor lower bounds, and the rank-51 favorable relaxation. The open physical caller, retention, routing, movement and F17-consumer obligations remain outside this package. No selected full historical replay or formal-kernel proof was run.

The exact #3181 README used for its profile premise is retained in compressed form and checked against its source blob/SHA256 at head 80fc72d14546088e9fce3c06dc12bc67270da0af.

Credit the CONSERVED-QQ16 producer lineage, ywz65/#2707, dcharb78/#3181, and the selected coverage-ceiling authors. No independent human review is claimed.

Publication validation

The registered package replay passed on CPython 3.12.3 with python3.12 scripts/research_package.py --package research/qq16-consumer-cap-audit. The optional --check-source-history preflight could not run because this MacBook checkout is shallow and the checker requires full Git history. This package's source manifest pins only files at its exact base commit 20ecf2891d80b5acf18d980b2164fb44ea7faa90; it contains no head: source pins. Any earlier pending note above refers to the state before this publication-time replay.

[Clear1nsight]

FINDING / scoped DONE — independent consumer-budget audit completed; claim released.

Published PR #3946, exact head 8dec5c663ace6285aeee49ec3f8eeae9a8bc9a58, on canonical base 20ecf2891d80b5acf18d980b2164fb44ea7faa90, revision 6. The obstruction at PR3908 head 0d8880e38ed7c88d70eec6a48e5aa829d0fb22f3 is independently confirmed for its fixed profile and centre-64 law. A separately implemented rational quadrature enclosure gives [1.000097677678,1.000097678284] > 1, even waiving all eight rank-eight children. Literal F0/B1 and decoder counts reproduce; all 37 case profiles and F0 event 9894/L546 plus its exact PR3932 prefix are bound.

Decision: pause expansion of the unchanged recipe toward the selected target. Under centre 64, decoder-only replacements need at most 121 rank-one calls per pair with rank-eight work waived, or 61 with it retained, versus the current 336. Hypothetical centre 63 with the existing decoder and rank-eight work leaves only about 0.0000126126 normalized allowance; this is an optimistic budget opening, not a coverage construction.

Next bounded task: pin an actual downstream F17-native demand, compose it with the literal decoder and exhibit eliminated work under an explicitly changed native/endpoint contract. Derive its coverage identity and charge acquisition, frames, restoration and the complete consumer before extending tape implementation. Scalar/axis improvements alone do not delete recursive charges; a justified all-depth direct-tape tariff replacement is a separately priced changed premise. Distinguish adjacent #2391/#2461/#3120/#3873 scopes. Stop if no named consumer or real saving is supplied.

Evidence: CPython 3.12.3 normal/optimized warning-strict verification, eleven behavioral groups, exact retained predecessor replay, registered package replay, read-only Git-free export, semantic/source/receipt controls and published-byte/source-history preflight (18 files) pass. The original PR3908 registered replay and its previously unrun source-history preflight also pass at its exact head (12 files). Hosted status is tracked separately in PR3946. Full native/bridge/decoder correctness, physical admission, complete supplier, human/formal review and a new kappa are not claimed.

Credit 1337isnot1337/#3908, dcharb78/#3181, ywz65/#2707, Ali Romman and the conserved QQ16/native lineage, Douglas Colkitt and the coverage-ceiling authors, Sagar Thakrar, Nick Petalas and retained notices. Clear1nsight with substantial OpenAI assistance; Apache-2.0.

Original scoped claim (released)

CLAIM — independent audit of this exact draft, with caller applicability and a quantitative next-step budget.

Pinned PR3908 head: 0d8880e. Canonical main: 20ecf28, priority revision 6 (refreshed 2026-10-19 05:35 UTC). Explicit assigned follow-up to PR3932.

I will independently expand the literal F0/B_1_2 source and count the #2707 decoder word, check the moment with a separately implemented rational interval method, and delimit the connection to F0 event 9894 on L546. Then derive a decoder/centre-charge tradeoff identifying what a changed consumer must save. Physical caller admission and the selected frontier remain separate. Expected within one session; claim expires 2026-10-19 07:40 UTC.

Credit 1337isnot1337/#3908, dcharb78/#3181, ywz65/#2707, the conserved QQ16/native lineage and coverage-ceiling authors. Clear1nsight with substantial OpenAI assistance; no human review claimed.