A second mathematician asks OpenAI for proof about its training data

A new party has entered the Navier-Stokes dispute. In a series of posts on Mastodon, mathematician Andreas Thom said interactions he and his colleagues had with ChatGPT may have contributed to OpenAI's mathematical results. One of the ten results the company announced last month concerned non-sofic groups, Thom's own field; OpenAI acknowledged the result built on earlier work by Thom and Gábor Kun and quietly amended its writeup after criticism. Thom said he emailed researchers Sébastien Bubeck and Mark Sellke, and that the answer addressed only whether his conversations could be accessed directly, not whether they had entered the training data pools. The dispute began days earlier with questions from Tristan Buckmaster of New York University about his use of Codex. OpenAI did not immediately respond to The Verge's request for comment.