Axiom Math formalizes classification of finite simple groups in Lean
Ken Ono, a number theorist known for his work on Ramanujan and integer partitions, writes in an essay published by The Guardian that he nearly left the field last year before deciding to remain. He is a professor of mathematics at the University of Virginia, currently on leave, and now works as founding mathematician at Axiom Math. He is a former vice president of the American Mathematical Society.
The essay opens with the result that he says has unsettled his discipline. Ono writes that on 20 November, Levent Alpöge announced that, working with Claude, he had proved the Riemann hypothesis, a 167-year-old problem about the distribution of prime numbers. The Lean and Mathlib communities have since certified the formal proof, Ono writes, before adding: “We are only beginning to understand the argument.”
He poses the question he says he cannot answer for anyone else: “What kind of future awaits someone beginning a life in mathematics?”
Ono’s own reckoning with the technology dates to May 2025, when he led the number theory team for Epoch AI’s FrontierMath project, which wrote difficult problems to test large language models. He describes watching an unreleased ChatGPT model compute all the moments of Tamagawa numbers of elliptic curves, drawing on work by Manjul Bhargava and Roger Heath-Brown and on papers his father wrote in the 1960s. “I was stunned,” he writes.
His father, a number theorist, died in January. Ono writes that his father had taught him as a boy about Ramanujan, the Indian mathematician whose notebooks have occupied mathematicians for generations. When Ono considered working on AI, he feared his father would see it as abandoning mathematics. Instead, he writes, his father told him AI reminded him of Ramanujan — that discoveries from an unexpected source might leave much to understand. Ono says he does not take the comparison literally, describing Ramanujan as a human being with an extraordinary imagination and life. “My father’s response brought me to tears,” he writes. “It gave me permission to keep being a mathematician in a way I had not anticipated.”
At Axiom Math, Ono writes, the work is to formalize, verify and extend mathematical proofs in Lean, a language in which a small checking program verifies each step. He says the approach can expose gaps and make proofs inspectable and reusable, but that it does not remove the human role. “Human judgment remains essential: we must check that the formal statement says what we intended,” he writes.
Ono says his team released the first complete Lean formalization of the classification of finite simple groups, among the largest proofs in mathematics. The classification was announced in 1983 and required decades of work by many mathematicians, with a major gap filled in 2004. Ono writes that AI helped produce 43.1 million lines of Lean code and that, in the process, the team found a problem in a published route to a key theorem and supplied an alternative argument. Lean’s kernel checks the resulting formal proof. The mathematics, he adds, belongs to the generations who developed it.
“After 24 days of work, I watched the final milestones land on my laptop. I cried,” Ono writes. “I thought about the people who had made the classification possible, and about the father I could no longer tell.”
He writes that he is 58, has advised 35 PhD students and 17 postdocs, along with hundreds of undergraduates, and has had opportunities young mathematicians cannot assume. That asymmetry shapes his view of what he can ask of them. “I cannot ask a graduate student to regard uncertainty as an adventure because I have found rewarding work within it,” he writes. “Nothing I write can reassure someone just starting a career. I can explain why I stayed.”
Ono describes a Berkeley colloquium held two days after OpenAI announced its Navier–Stokes solution, at which a scheduled lecture became a town hall lasting almost two and a half hours. Students asked about credit, transparency, hiring, and how mathematicians whose work came first were treated. Some asked what he would do to help them, he writes, and campus security was locking the building as they left.
He says he shares the concerns set out in a declaration titled “A Severe Misalignment of AI in Mathematics,” signed by Terry Tao and other Fields Medalists, which warns about rushed announcements, inadequate attribution and the loss of mathematical understanding. At a National Academies meeting, he says, he put his position this way: “hope is not a plan, and nostalgia is not mentoring.”
Ono writes that universities, journals, employers and funding agencies must decide what they support and reward, and argues that students and younger colleagues should have a voice while those decisions are still open, as should mathematicians who choose to work without AI. He says he will argue for that participation wherever he has a voice.
In his own advising, he writes, he will ask what a young mathematician understood, originated or made possible, adding: “Access to a powerful system is no substitute for those contributions.” For work he helps publish, he writes that readers should know what has been proved, what remains assumed and who contributed the ideas.
He closes by saying he wants his students and their students to develop ideas that are genuinely their own, with colleagues who listen and the material support to pursue a mathematical life. His position at Axiom, he writes, is one part of how he is trying to contribute and does not replace his obligations as a teacher and mentor. “I may be wrong about how this will unfold,” he writes. “I am still learning what it means to be a mathematician.”
The essay carries a note stating that the views expressed do not necessarily reflect those of Axiom Math, the University of Virginia, the National Academies or any other institution with which Ono is affiliated.