Jean-Pierre Serre Turns 100: The Youngest-Ever Fields Medalist Is Still Lecturing in Paris

On the afternoon of September 15, in the Hermite amphitheater of the Institut Henri Poincaré (11 rue Pierre et Marie Curie, Paris 5th arrondissement), the speaker list included Pierre Deligne, Peter Sarnak, Maryna Viazovska, and Don Zagier — a lineup that would headline any mathematics conference.

But the real headline was the last name: Jean-Pierre Serre, who turns 100 on September 15, is himself on that afternoon's program.

The two-day celebration runs September 15–16. The same day, the European Mathematical Society's *EMS Magazine* (issue No. 141) published an entire issue dedicated to Serre, including an interview spanning nearly five hours. A Hacker News thread appeared under the understated title: "Serre is 100 years old today."

A living centenarian with his body of work, his awards, and still giving lectures — mathematics history has no second example.

Key points

• From a pharmacist's son: Born September 15, 1926, in Bagnères (Pyrénées-Orientales), Serre first met calculus through old textbooks on his mother's shelf, computing "purely formally, in Euler's style" — he disliked and couldn't parse epsilons and deltas. At 19 he entered the École Normale Supérieure and became the youngest member in Bourbaki's history.
• The night train, 1950: Returning from vacation, he suddenly saw how to take path spaces themselves as objects of study — with the fibration's fibers as loop spaces, spectral sequences could compute. He was so certain he woke his wife. His 1951 Sorbonne thesis on the homology of fibrations followed, and those tools were promptly applied to a long-standing problem: computing homotopy groups of spheres.
• 27, still unbeaten: At the 1954 Amsterdam ICM, Fields Medal committee chair Hermann Weyl announced medals for Kunihiko Kodaira and the 27-year-old Serre, noting it was the first medal awarded to a mathematician outside analysis. Seventy years later, no one younger has won it — and the record is unlikely to fall.
• 1955–56: *FAC* (Faisceaux algébriques cohérents) systematically brought sheaf theory into algebraic geometry; *GAGA* proved complex algebraic and analytic geometry coincide in a precise sense. These papers, plus his duality theorems, became the grammar of modern algebraic geometry; much of Grothendieck's grand edifice rests on them. (Serre and Grothendieck met at the 1948 Cartan seminar; their twenty-year correspondence is among the twentieth century's most important.)
• A footnote feeds a medal: On page 243 of FAC, Serre casually asked whether finitely generated non-free projective modules exist — the "Serre conjecture," a name he protested. Quillen and Suslin independently proved it in 1976; it featured among Quillen's 1978 Fields Medal citations.
• Number theory: His modularity conjectures (1973, 1987), fully proven by 2009, became a key pillar of the bridge Wiles used to prove Fermat's Last Theorem. His 1977 booklet *Trees* made groups acting on trees a standard tool of geometric group theory. Fields, Balzan, CNRS Gold Medal, Steele, Wolf — and in 2003, the inaugural Abel Prize, presented by King Harald V.

The gray zone: machine-verified proofs

His most striking remark connects unexpectedly with the AI era. In a 2003 *Notices of the AMS* interview, asked about computer-assisted proofs, Serre said:

> "We are entering a gray zone: computer-assisted proofs. They are not proofs in the standard sense, because they cannot be checked line by line by hand. They are especially unreliable when they claim to enumerate exhaustively some class of objects. In the nineties I received a list of subgroups of given index in some discrete group; the computer found twenty. I know these groups well and easily found about thirty by hand. I wrote to ask. The authors explained that part of their computation was done in Japan and part in Germany, and a segment in between was forgotten. All too typical!"

Then the pivot: "That said, computer-assisted proofs are often more convincing than many standard proofs."

A gray zone, with no ground ceded on either side. Written over twenty years ago, it reads like a prophecy for the age of formal verification and AI proof — trust must be redistributed, but the right to doubt cannot be outsourced.

A related myth is worth settling: the internet impression that Serre "refuses computer proofs" is inaccurate. He never commented on Lean or any proof assistant — no such statement exists in public records. What he evaluated was only the above: lists must reconcile, proofs must be re-checkable.

How to write math, and how to grow old

A well-known video, "How to write mathematics badly," recorded at a Harvard seminar in November 2003, runs over fifty minutes on making papers terrible — in reverse. He counted FAC among his most important papers that were "written very smoothly."

The same playfulness runs through the new interview. After nearly five hours with interviewer Fresán, the two met again to go over the transcript line by line — "in the Bourbaki spirit: delete adverbs and adjectives, replace wrong theorems with correct ones." Asked whether age had dulled his interest in any area of mathematics, he answered: "I can't think of a single one. Quite the opposite. Age has made me appreciate things I was ignorant of, or even disliked, when young."

His hobbies at 100: skiing, table tennis, and climbing — Fontainebleau-style bouldering, though he says he can now only do "compensatory climbing," driving friends to the base of the boulders. Favorite films include *Pulp Fiction*; his reading runs from Giono to *Harry Potter*. On HN, the French call him "Tonton Serre" — Uncle Serre.

A detail pointing at tomorrow

On that 1950 night train, a 23-year-old woke his wife to say he had seen something. It became a thesis the next year, a Fields Medal seven years later, and a packed Paris amphitheater seventy-two years later — where Maryna Viazovska, the 2019 Fields medalist also on the program, studies sphere packing. From homotopy groups of spheres to the densest packing of spheres: seventy years apart, the methods have changed entirely, but the temperament underneath is the same — turn what cannot be computed into what can be seen.

Epsilon-delta once annoyed him; he learned it anyway, then helped write it into textbooks worldwide. The most practical lesson for ordinary readers: when you don't understand something, first remember that you don't. Happy birthday, Monsieur Serre. Two days of talks, two live streams, and the whole world watching.

---

*Sources: MacTutor History of Mathematics (St Andrews), Wikipedia and citations, serre100.sciencesconf.org, EMS Magazine No. 141 (2026-09-15, open access), Notices of the AMS 51(2) 2004 interview, hn.algolia 49718822.*