The supplied snippets do not substantiate the case’s claim about an OpenAI model proving that non-sofic groups exist. Instead, they describe a separate OpenAI claim: an unreleased general-purpose reasoning model produced a counterexample to an Erdős conjecture in discrete geometry, with mathematicians reportedly extracting and rewriting the result and external experts reviewing it. No snippet establishes Sébastien Bubeck’s role, the validity or novelty of a non-sofic-groups proof, or even that such a proof was produced; those points remain pending expert assessment.
No intersection with Scott’s own wikis or existing radar pages was found. Moreover, the supplied grounding does not substantiate the central non-sofic-groups claim, so there is not yet a grounded development to assess against Scott’s work or positions.
queries asked of Scott's wikis
- AI-generated proofs and expert validation
- formal verification of model-discovered mathematics
- autonomous theorem proving versus human proof extraction
- novelty and training-data provenance in AI discoveries
- reasoning-model harnesses for open mathematical problems
- AI mathematics claims and evaluation standards
2026-08-07T18:32:08Z
No proof, expert verdict, or first-party technical account has emerged after several days, leaving the original report both uncorroborated and misgrounded; absent new primary evidence, this episode has faded.
2026-08-01T18:23:17Z
The added activity still supplies no inspectable expert verdict or independent corroboration, so the case remains a misgrounded, unverified report rather than evidence of a mathematical breakthrough. Further amplification without a proof, review, or first-party technical account does not advance it.
2026-08-01T17:23:39Z
grounded: novel/none — No intersection with Scott’s own wikis or existing radar pages was found. Moreover, the supplied grounding does not substantiate the central non-sofic-groups cl
2026-08-01T17:23:04Z
The purported external assessment provides no inspectable conclusion and does not independently corroborate the non-sofic-groups claim. With the existing grounding still pointing to a different mathematics result, this remains an unverified report rather than evidence of a breakthrough.
2026-08-01T17:21:38Z
evidence attached: hn.story.49136236 — An external assessment of OpenAI's reported mathematics results materially bears on whether its claimed proof is valid and novel.
2026-08-01T10:22:21Z
grounded: novel/none — No intersection found in Scott’s wikis or the radar. Moreover, the supplied grounding does not substantiate the claimed non-sofic-groups proof or Bubeck’s role,
2026-08-01T10:21:49Z
case created — A claimed model-generated resolution of a major open mathematics problem is consequential and independently resolvable through expert review.