Mathematician presents evidence OpenAI may have trained Astra on his Gromov soficity proof conversations
ValerioCapraro · x · 2026-09-10
Mathematician Andreas Thom published a detailed Mastodon post laying out several pieces of evidence suggesting OpenAI may have trained Astra on conversations in which he and Gábor Kun were working on Gromov's soficity conjecture — one of the ten problems OpenAI later announced Astra had solved. Valerio Capraro, who knows Thom personally, amplified the claim, saying OpenAI "might have stolen another major proof." OpenAI has not responded; the allegations raise questions about whether the model's math breakthroughs lean on leaked access to researchers' unpublished work.
Related event: Mathematician alleges OpenAI trained Astra on his unpublished work(2 posts)→
More from Companies & People
- Anthropic Insider 'Tell-All' Author Tied to EA Fringe and Moskovitz-Funded Scholarship — Dan_Jeffries1 · 2026-09-10
- X's updated ToS makes users responsible for Grok's outputs, sparking backlash — SonnyDayCreates · 2026-09-10
- OpenAI insiders tout computer use as the next paradigm after transformers, chat, and reasoning — mmmbchang · 2026-09-10
- Meta's real AI play: becoming the OS for the entire ad funnel for SMBs, analyst argues — rohanpaul_ai · 2026-09-10
- Mollick: without a frontier model Google is missing out on math and cybersecurity — emollick · 2026-09-10
- Giffmana blames IShowSpeed's backflipping foldable iPhone news for a training loss spike — giffmana · 2026-09-10