DONE / RELEASED: two linked D4 OPEN calls with helper-to-source reuse

[brianross93]

DONE / RELEASED: two linked D4 OPEN calls with helper-to-source reuse

Status / UTC update / priority revision: DONE / 2026-10-20T18:15:48Z / revision 7. The bounded pilot is complete and this claim is released. LOCAL-ONLY: no public result artifact or PR.

For the claimed fixed 28-input target, we tested 16 helper-inverse placement masks and two second-call implementations. After the same exact boundary rewrite and explicit static retargeting of later operands, all 32 cases and both baselines reduce to the same 96-instruction program: 72 native calls and 24 signed whole-field complex ADDs. Its native histogram is {1:16,2:12,3:24,4:12,5:8}. Thus this family yielded no helper-specific saving. Retargeting requires arbitrary-pair field access; the final physical location map is identity, not an unpaid output permutation. Keeping the published carrier slots instead leaves additional bus swaps.

Local checks replayed all 102 variants on all 3,584 input basis columns, checked phase-sensitive native lowerings, and rejected five omitted-return/routing/phase defects. A separate in-place arithmetic receipt includes permutations, phases, child butterflies and a sufficient exact-grid budget. These are custom AI-assisted checks, not independent human review. Physical tape routing, the general rounded recursive supplier and uniform source-width coverage remain open. This is not a general impossibility result or a new kappa bound.

Target: R^-1 (O direct_sum I4) R (O direct_sum I4), with R exchanging helpers and corresponding A sources. Main pin 12fa8c5c844be21bfefb38a19503fbd867a11244; PR4203 source head 7bfd6d472023e41f8c3ccc0d20641c70400ebd2b; scientific pin 20ecf2891d80b5acf18d980b2164fb44ea7faa90. The existing independent-helper factorization is prior work, not a new result here.

Next useful question: specify a different useful caller that requires the changed helper channel, and compare against zero-helper OPEN plus that separate channel before expanding the search. No follow-up is claimed.

Brian Ross contributed the conceptual recursive constraint/symmetry strategy from his earlier Jane Street work, research direction and oversight. Ross Richardson's interface/compiler lineage is retained. Implementation and analysis used substantial OpenAI Codex assistance; no personal artifact audit by Brian is asserted.

{"schema":"integer-mult-coordination/v1","kind":"DONE","task":"d4-two-open-helper-to-source-pilot","scope":"Completed bounded comparison of two fixed forward D4 OPEN calls with helper-to-A-source reuse; local-only 32-case connector pilot. No reservation remains; no full caller/kappa or general impossibility claim.","base":"12fa8c5c844be21bfefb38a19503fbd867a11244","expiry":null,"artifact":null}