OpenAI releases model-generated math preprints; mathematician flags poor exposition
anshulkundaje · x · 2026-10-09
OpenAI announced a batch of new mathematical results from an internal frontier model, saying it consulted the independent Advisory Group on Mathematics and AI at the Institute for Advanced Study on how to release them.
Mathematician hayamizulab reviewed the combinatorics preprints and found the written exposition varies widely: some papers are so poorly written he cannot even tell what is being claimed, before correctness is even assessed. Citing the preprint on exact crossing numbers of complete graphs, he says Lemma A.1 and its proof (pp. 11–12) fall short of peer-review standards and would warrant major revision — though his critique doesn't cover the Lean formalizations.
Related event: Math community splits as OpenAI drops 700+ manuscripts, AHM calls boycott(34 posts)→
More from Models
- Agent Arena Leaderboard: Claude Opus 5.5 Tops GPT 6 Astra Across 2.3M Agent Sessions — arena · 2026-10-09
- Same prompt without thinking mode: a stark model quality gap — umesh_ai · 2026-10-09
- Qwen3.8-Max, Flash and Wan3.0 free for a week on GMI Cloud with higher rate limits — Alibaba_Qwen · 2026-10-09
- ByteDance Seed team spotted DeepSeek performance drift, Zhihu explainer goes viral — teortaxesTex · 2026-10-09
- Chollet: far transfer has no evidence in humans, and LLMs/LRMs may be no different — burny_tech · 2026-10-09
- Google AI Overviews hallucinates while user searches for a meme — prajdabre · 2026-10-09