Jordana Cepelewicz reports in Quanta Magazine that OpenAI's September 8, 2026 announcement — that its models had solved one of mathematics' most famous problems — felt sudden and violent to many mathematicians. Online reactions mixed grief, confusion, and fear. Scott Aaronson of the University of Texas, Austin wrote that human mathematicians are "forevermore dethroned as the main theorem-proving entities on planet earth."
At a Berkeley lecture, former professor Ken Ono — now at the AI startup Axiom Math — told roughly 150 students, postdocs, and professors to brace for a profession that "might not even exist" in its current form. The audience responded with anger and frustration. The scheduled 50-minute talk stretched past two hours, with one student asking how AI companies would take responsibility for "the shameful way" they are treating mathematics.
The deeper issue, the article argues, is that mathematics has always prized understanding, beauty, and the journey of discovery, not just final answers. AI proofs are checked for logical correctness but are nearly impossible for humans to read or connect to existing work. As one researcher put it, "It was never about the theorems" — but with AI skipping the climb and teleporting everyone to the mountaintop, that tradition is now in question.
The source presents a single perspective, so there are no differing accounts to compare. Its tone is reflective and openly concerned about what mathematics will become if human understanding is no longer needed to produce results.