As more and more central problems are nuked into oblivion by generated proofs, against the spreading AI-gloom the voices rise, emphasizing the internal value of deep mathematical thought and advocating for measures ranging from exercising extreme discretion with generative tools to seizing railway stations and the telegraph (or was it departments and classes of mathematical problems?) from the tightening grasp of the corporations.
This is understandable. Even if these pieces of advice seem hardly practical, we all need some venting as our dear field erodes beneath our feet, leaving soulless proof certificates where once the pure magic flowed. For what compels us to this calling, rather suboptimal in terms of earthly goals, if not its personal appeal, the sacred light of understanding, rising from hastily written notes in a dimly-lit room.
This intimate link is worth fighting for, perhaps by locking yourself in a cabin in the woods for the purposes of practicing pure thought (don’t try this at home; not without secured tenure, or at least a decent financial cushion). This way, no matter how much math were to succumb to the implacable march of AI, its spirit will be preserved.
It won’t.
No matter how elevated the guiding idea of a social enterprise, its body consists of boring stuff, such as career trajectories and job security, not necessarily reduced to, but surely resting upon the utility of the said enterprise. With the utility gone, the bottom to middle layers will be chipped away, their inhabitants turned into statisticians or worse.1 A masterful artisan, descending to the market for mass-produced furniture with the true understanding of a bedside table, will find little appreciation of his work.
Oh, come to me, math jobs market analyst, with hard data in hand; come and prove me wrong. But the September night is already offended by the sticky feeling of an exodus that has begun. An exodus that probably will be left unnoticed outside of the bubble, for the “I didn’t see you guys all the way over there” meme cuts both ways.
It is indeed sad. The utility I’ve been referring to is not necessarily, so to speak, utilitarian. Arguably, the authority of math is largely symbolic, grounded in its ability to find beauty emerging from triviality, poetry disguised as science. Those are smart guys, doing deep things, with integrity and a sense of purpose. I want to be a part of it. I don’t want to see it gone.
But even if AI progress in math were to come to an immediate halt, the high claim to possessing a unique and revealing insight into the nature of things would be shaken. The halo dims, and with it will dwindle the public attention and the funds. Maybe these smart guys will peacefully transition into areas with a more measurable impact; maybe they won’t. Maybe there won’t be anything to transition to.
Amidst a relatively mild disaster, as history turns its back on this land, I tell my young son: Learn mathematics, learn your French, learn your history.
Acknowledgements. Part of the blame for this text is on Yuval Filmus, who has drawn the author’s unhealthy attention to this resource.
Received 6 September 2026.
- This reference to Terence Tao’s quote from his interview, “On Math Education and Research”, is meant as a joke; no statistician was harmed in making this text. The full quote reads: “Most students who take math classes aren’t going to be mathematicians. They’re going to be engineers, statisticians – in many ways, that’s the more important mission of math education.” ↩︎
Add to the discussion