"OpenAI posts 700 maths preprints online: mathematicians are up in arms."
https://www.nature.com/articles/d41586-026-03196-8
Very interesting, even apart from the mass #scooping of creative human mathematicians:
* Open AI has an advisory group of human mathematicians, which advised as recently as Sept 29 to release new mathematical results "as soon as possible", to minimize the time people might waste on solved problems.
https://agmai.org/general-sep29/
* At the same time, the advisory group advised against "the use of proprietary internal models by AI labs to do mathematical research" because it "risks creating a two-tier system where labs outrun the rest of the field."
* These preprints were dumped in #arXiv just before Oct 1, when arXiv started enforcing a new policy limiting contributors to two submissions per month.
https://blog.arxiv.org/2026/10/01/updated-rate-limit-policy/
* Was it a coincidence that Open AI submitted these 700 preprints just under the wire? Or did it see the deadline looming and rush to deposit these results even if it could have made them better with more time?
* BTW, the advisory group also "strongly recommend[ed] that AI labs refrain from treating the release of mathematical results as marketing vehicles to promote their models."
how are the 700 preprints actually attributed, named mathematicians or basically one OpenAI submitter? that decides whether the two submissions per month cap would even have touched them, and whose name the citation graphs end up attaching to the results.