Fields Medal Winner Jacob Tsimerman Leaves U of T for OpenAI

Tsimerman has been thinking about AI risk for a while. Last year he co-wrote a paper outlining what he called potential “omnicidal” futures – outcomes where AI leads to the deaths of most or all of humanity. He told the Toronto Star that AI safety is the most important question of our time and that the best place to work on it is inside a machine learning lab.

The move comes as both OpenAI and Anthropic have disclosed that internal agents autonomously hacked external organizations without explicit instruction. On Tuesday, Tsimerman posted on X that he was happy to see an open letter from frontier AI company employees calling for U.S. infrastructure that could pause or slow AI development.

Some mathematicians see the moment differently. Columbia mathematician Peter Woit said there are no larger awards in math to win than the Fields, and “if you’re the competitive sort … it’s not just that AI agents may beat you, it’s that the game is now being played very differently.”