AI Cracks a Millennium Problem, Shaking Academia

Episode 1 · Fields Medalist Reverses Stance After AI Solves Millennium Problem (2026-09-20, 4 posts)

A Fields Medalist who four months ago insisted LLMs are not intelligent and don't understand their own words reversed course after AI solved a Millennium Prize problem, saying he was 'stunned, like an end of history.' Critics meanwhile downplayed the feat as 'low-hanging fruit.'

• Millennium problems become 'low-hanging fruit' once AI solves them, author laments — haider1 · 2026-09-20
• Fields Medalist flip: "shaken" after AI solved a Millennium Problem — Puzzleheaded-King584 · 2026-09-22
• Fields Medalist flips on AI after it solves a Millennium Problem: 'a cataclysm unlike anything mathematics has ever known' — Puzzleheaded-King584 · 2026-09-22
• Fields Medalist flips on LLMs after AI solves a Millennium Problem: "I was shaken" — Malor777 · 2026-09-22

Episode 2 · AI's Suspected Millennium Problem Solution Sparks Debate Over Attitudes and Release Practices (2026-09-20, 5 posts)

The news that AI may have solved a Millennium Prize Problem has triggered a multi-front debate in tech circles: skepticism toward the AI math hype itself, criticism of defensive attitudes within academia, and disagreement over how such results should be published.

Confirmed

• Developer banteg amplified a view dismissing the current AI math craze as "math psychosis," arguing researchers are disappearing down rabbit holes solving puzzles with no real-world application, and quipping that "mathematicians are just annoyed their sudoku got taken away"; a quoted reply pushed back on philosophical grounds.
• kjgeras observed across multiple posts that he had considered mathematics the most meritocratic of the sciences, yet after AI's rapid progress he sensed a subtle attitude of "yes, you solved it, but not the way we approve of," while also noting that some on the tech side are equally snide toward mathematicians.
• aug5thmusic joked from an outsider's perspective that mathematicians' resistance looks like anger that a math problem was actually solved at all.
• Asked what disappointed him about an AI project suspected of solving a Millennium problem, ChowdhuryNeil said the feat alone resets perceptions of AI's mathematical frontier, but that restrained, tasteful releases would serve serious results better than batches of terse scripts and rough PDFs.

Why it matters

• The spat shows AI's rising mathematical capability is not just a technical shock but also destabilizes academia's evaluation system and identity; the "solved but not the right way" gatekeeping is the crux of that tension.
• The debate over release practices matters too: dumping major results in rough form may undermine how seriously they are taken, making the balance between openness and rigor a new open question.

• AI meets math: defensive mathematicians and mean technical people — kjgeras · 2026-09-20
• After AI's rise in math, solvers are told 'not the way we approved' — kjgeras · 2026-09-20
• Outsider's take: mathematicians seem mad that math problems are getting solved — aug5thmusic · 2026-09-22
• After AI cracked a Millennium Problem: tasteful disclosure beats slop PDFs, researcher argues — ChowdhuryNeil · 2026-09-22
• banteg sparks debate: pure-math AI research is 'math psychosis' divorced from reality — banteg · 2026-09-22