• Sources: primary, discussion, discussion
  • Summary: The history log in the openai/math repository records three manuscripts withdrawn and 14 revised since the 2026-10-06 release, with the entries dated 2026-10-07. The log states that a sign error in "Algebraicity of Weil classes on split abelian eightfolds" invalidates a stabilization-trace cancellation argument and the construction that argument supports. Two dependent manuscripts that used the construction were withdrawn with it.
  • Why it matters: The release this archive covered yesterday in AI now carries a vendor-published defect log, and the first retraction traces to a sign error in an argument that two dependent manuscripts built on, which is the failure mode a passing formal check does not cover.
  • Follow-up: Further withdrawals or revisions recorded in the openai/math history log.