Another mathematician suspects OpenAI trained on his personal chat data
erikphoel · x · 2026-09-10
After Erik Hoel's essay arguing OpenAI's Navier-Stokes solve likely drew on user drafts, mathematician Andreas Thom is publicly wondering whether OpenAI used his personal user data to train the model's construction of a non-sofic group. Following the Jacobian example, this is another mathematician flagging their own work as possible training data, fueling suspicion that recent AI math breakthroughs mined user chats.
Related event: Researchers Suspect OpenAI Secretly Trains on Their Inputs(13 posts)→
More from AGI Musings
- lemire: 'superintelligence will kill us all' is a moral claim, not a rational one — lemire · 2026-09-10
- Kelsey Tuoc: OpenAI and Anthropic staff believe it'll be bad if the other gets recursive AI first — EigenGender · 2026-09-10
- Next enterprise software era: Civilization-style interfaces atop company agents — smtabatabaie · 2026-09-10
- tinygrad: API access is not distribution — without owning hardware, AI won't stay aligned — cephaloform · 2026-09-10
- Michael Levin's new paper: a structured latent space of patterns for new forms of life and mind — danfaggella · 2026-09-10
- Do AI doomers really have a strong forecasting record? XPT study suggests otherwise — random_walker · 2026-09-10