Mathematics, the one field that has always promised certainty, is sliding toward a future where its most powerful tool delivers answers no human can explain.
The warning comes from applied mathematician Steven Strogatz and his Cornell colleague Alex Townsend, whose new book Big Math: The Hidden Codes That Run Our World contends that modern computation has quietly outpaced the minds that built it. Their term for the shift is big math: problems so vast, with billions of variables or more, that no human brain can hold them and only machines can.
For most of computing history, machines were servants to mathematicians, grinding through arithmetic too tedious or enormous for pencil and paper. By the 1950s they became scouts, venturing into unknown mathematical territory; one simulation by Enrico Fermi and his colleagues defied every expectation when energy that should have dissolved into disorder instead came roaring back in a strange echo, opening an entirely new branch of the field. Today the computer has become something else entirely: an oracle.
Modern AI now returns correct results while offering no insight into how it reached them. AlphaFold predicts the shapes of thousands of proteins and earned a Nobel Prize in chemistry, yet it learned no new physics or chemistry to do it. Translation engines convert languages flawlessly without revealing a single principle of how meaning works. The release of 722 AI-generated math papers in a single month crystallized the unease now spreading through the discipline.
For pure mathematicians, the stakes are existential. Their field has long prided itself on absolute proof, and Euclid's theorems still hold after more than 2,300 years, while nothing from the science of that era survives. Many now fear trading that eternal certainty for something more like medicine, where a treatment is trusted because it works in trial after trial, not because anyone understands its mechanism. Young researchers feel it most acutely, warned that the subject they devoted their lives to may already be over.
Yet the book resists despair. AI could become a patient teacher, potentially deepening human understanding of mathematics rather than ending it, while accelerating progress in medicine, climate and energy. Strogatz and Townsend close on Goethe's ambiguous Faust, redeemed not for selling his soul but for refusing to stop striving toward something larger than himself. The math keeps getting bigger. Whether we can grow with it without losing the joy of the search is the question the book leaves open.