According to Dynamic Beating monitoring, the newly awarded Cole Prize winner Jacob Tsimerman has announced a shift to AI safety and will be joining OpenAI. Several OpenAI research leads have publicly welcomed his appointment.
His concerns about AI are not a recent development. In 2025, he co-authored a paper with Andrew Critch, systematically outlining the potential pathways for AI to cause nearly all of humanity's death. The paper did not claim that extinction was inevitable but rather aimed to drive preventive measures.
He has also stopped admitting new math Ph.D. students. Tsimerman believes that AI may be better at mathematics than mathematicians within two years. He does not want students to prepare for a career that may disappear.
Interestingly, one of the undergraduates he mentored at Harvard, Levent Alpöge, has joined Anthropic. Alpöge recently used Claude Fable 5 to find a counterexample, overturning the 87-year-old suspended Jacobian Conjecture.
While the teacher goes to OpenAI to research AI control, the student at Anthropic uses AI to overturn mathematical conjectures. The future that Tsimerman worries about is already unfolding around him.
