PHILADELPHIA — On July 23, 2026, Jacob Tsimerman received the Fields Medal, mathematics’s most prestigious prize, for his role in proving the Andre-Oort conjecture, a problem that had stumped the world’s best minds for more than 30 years. It was the crowning achievement of a career that began with a perfect score at the International Mathematical Olympiad at age 16 and led, through a Princeton PhD and a professorship at the University of Toronto, to some of the most profound breakthroughs in modern number theory.
And then there is the other thing Tsimerman says, the thing that makes his medal feel less like a celebration and more like an elegy.
“I think AI will be better than mathematicians at doing math within two years,” he told Quanta Magazine. He has stopped taking graduate students because he worries that training them for conventional research prepares them for a career that may not exist. He shifted his own research to AI safety. He co-authored a taxonomy of omnicidal scenarios involving artificial intelligence. The man who just won mathematics’s highest honor is quietly convinced that human mathematics, as a living enterprise, has only a few years left.
This is the paradox at the heart of Tsimerman’s story. He is simultaneously the field’s greatest current champion and its most credible doomsayer. He arrived at both roles the same way: by being first.
The Making of a Mathematical Mind
Tsimerman was born in 1988 in Kazan, Russia. His mother was a high school math teacher; his father, a computer scientist. At three, the family moved to Israel, and at nine to Toronto. By the early 2000s, he was dominating international competitions. At 15 he won gold at the International Mathematical Olympiad in Tokyo. At 16 he achieved a perfect score in Athens, then quit high school to enroll at the University of Toronto. He earned his undergraduate degree by 18 and a Princeton PhD under Peter Sarnak by 2010.
Sarnak gave Tsimerman a long list of books and papers. The first eight months of graduate school were a fog of theory and creeping anxiety. Tsimerman went to his advisor and pleaded for a concrete problem, anything that would let him gain a foothold. Sarnak gave him an analytic number theory problem, his least favorite subfield. Tsimerman solved it and co-authored his first paper.
“The skill of sitting there for several hours, repeatedly banging your head against the wall, trying new things while you’re stuck, is a super difficult and important skill to learn for research math,” he later said.
The Andre-Oort Conjecture
The Andre-Oort conjecture, proposed by Yves Andre in 1989 and generalized by Frans Oort in 1995, asks when certain highly symmetric points inside geometric spaces known as Shimura varieties cluster together in ways that imply hidden structure. It sat at the intersection of number theory, algebraic geometry, and the Langlands program, defeating every attempt at a full proof for decades.
Tsimerman first encountered the conjecture during his PhD and proved a crucial result about the size of Galois orbits, which became the key that unlocked the first major cases. In 2010, he began collaborating with Jonathan Pila of Oxford, who had a method for the simplest version. Tsimerman combined Pila’s method with his own bounds to prove the conjecture for Siegel modular varieties.
The full proof came in 2021, when Tsimerman, Pila, and Arul Shankar posted a complete, unconditional proof on the arXiv (2109.08788). It did not rely on the Generalized Riemann Hypothesis, an unproven conjecture earlier partial results required. It was a definitive resolution of one of the most important open problems in arithmetic geometry.
Tsimerman did not slow down. With Benjamin Bakker and Yohan Brunebarbe, he proved Griffiths’ conjecture in 2018, developing an “o-minimal GAGA” framework that revealed hidden algebraic structures in analytic spaces. Between 2024 and 2025, he posted 11 papers to the arXiv, culminating in an August 2025 paper with Shankar on “secondary main terms for quartics” that Tsimerman knew the Fields committee would value.
“When you’re zoomed in,” he once said, “you just want to win.”
The Man Who Saw the End
Tsimerman won. But from the top of the field, he can see what is coming.
In May 2026, an OpenAI model disproved the unit distance conjecture, an 80-year-old problem in discrete geometry that Tsimerman himself had worked on and failed to solve. The result was publishable in top mathematics journals. An AI did what he could not.
For Tsimerman, this was a catalyst. His prediction that AI would surpass human mathematicians within two years is not a provocation. It is a sober assessment. He stopped accepting graduate students; he does not know how to advise someone for a career that may not survive the decade. He shifted toward the formal verification of AI systems and the study of AI agent behavior.
“There are many people whose primary enjoyment of math comes through problem solving,” he wrote on X in May 2026. “If that disappears, that is not a trivial issue and many of them might not want to do it anymore.”
His research now includes a 2025 paper with Andrew Critch, “A Taxonomy of Omnicidal Futures Involving Artificial Intelligence,” cataloging scenarios in which AI poses existential threats. He is not sure how to solve these puzzles. But that has never stopped him from trying.
The Improvising Mathematician
The same man who predicts the end of mathematics also leads improv comedy classes and co-wrote a musical comedy called “The Sombrero Detective.” He trained at the Upright Citizens Brigade. His colleagues describe him as warm and funny, devoid of pretension. At the Institute for Advanced Study, where he spent the past year, he greeted journalists in a grey T-shirt and red shorts, lying on his office floor to stretch during interviews.
Tsimerman loves mathematics the way an improv actor loves a scene: as a collaborative, human act of creation. He calls math “a team sport.” The danger he sees from AI is not just that machines will solve problems faster. It is that the struggle, the head-banging, the late-night breakthroughs, will become irrelevant.
“If that disappears, that is not a trivial issue.”
What Comes Next
His prediction may be wrong. Many peers, including Terence Tao, argue AI will serve as a powerful assistant rather than a replacement. Others remain skeptical that current AI is useful for serious research at all. The debate is unsettled.
But Tsimerman has already made the pivot. He is working on AI safety, writing taxonomy papers about existential risk, and telling anyone who will listen that the game he just won may be the last of its kind. He is not bitter. He is not resigned. He is treating it the way he treats every hard problem: sitting with it, banging his head against the wall, and waiting for something to give.
“If the social implications are complicated,” he said, “I want our community to face this moment with honesty and compassion.”
On July 23, 2026, Jacob Tsimerman accepted the Fields Medal. He earned it the old-fashioned way. And then he went back to thinking about what comes next, for mathematics and for the species that invented it.
References
- Jacob Tsimerman, Jonathan Pila, and Arul Shankar. “Proof of the Andre-Oort conjecture.” arXiv:2109.08788, 2021.
- Benjamin Bakker, Yohan Brunebarbe, and Jacob Tsimerman. “o-minimal GAGA and a proof of Griffiths’ conjecture.” 2018.
- Jacob Tsimerman and Andrew Critch. “A Taxonomy of Omnicidal Futures Involving Artificial Intelligence.” arXiv:2507.09369, 2025.
- Jacob Tsimerman and Arul Shankar. “Secondary main terms for quartics.” arXiv:2508.08527, August 2025.
- Kevin Hartnett. “Jacob Tsimerman Wins 2026 Fields Medal for Andre-Oort Conjecture Proof.” Quanta Magazine, July 23, 2026.
- Luc Rinaldi. “What I learned locked inside an escape room with one of the world’s most brilliant mathematicians.” Be Giant, July 16, 2026.
- University of Toronto. “Jacob Tsimerman awarded prestigious Fields Medal.” U of T News, July 23, 2026.