Steven Strogatz, the Cornell mathematician and author, does not usually cry in interviews. But when he begins to talk about what artificial intelligence has done to his field over the past week, his voice catches and tears surface. “The science is thrilling,” he says, “but there’s a lot of human unpleasantness going with it.” The trigger was a dizzying sequence of announcements. OpenAI said it had deployed tens of thousands of AI agents to attack a 90-year-old math problem, one carrying a million-dollar prize. The solution, still awaiting independent verification, builds on work by Spanish mathematicians Diego Córdoba and Luis Martínez-Zoroa. Almost immediately, another mathematician, Tristan Buckmaster, alleged that OpenAI had rushed ahead after learning of his own parallel work with Anthropic researcher Levent Alpöge, and that the company tried to influence who got credit. It is a mess of egos, money, and machine-generated discovery. For Strogatz, the tears are not only about the science. They are about watching a profession he loves be remade in a matter of days by forces no individual can control. The old math world had its rivalries and jealousies, but this is different. This is a corporate race where headlines, IPOs, and algorithms are reshaping the field before anyone has time to understand what is happening. The breakthrough itself could be a genuine advance: if verified, it solves a problem that has resisted human effort for nearly a century. But the human context around it is ugly, contested, and unresolved. Strogatz is old enough to remember when mathematics felt like a quiet pursuit, conducted by people in offices with chalkboards. Now it is a frontier for large language models, and the people who spent their lives on that frontier are not sure whether they are being elevated or erased.
The rupture around OpenAI’s announcement is not an isolated incident. It is a symptom of rapid and wrenching change. Strogatz and his collaborator Alex Townsend have just finished Big Math, a book about how mathematics is slipping away from human understanding, due in November. That book now feels almost prophetic. Last week, Anthropic said its model Claude had proved 29,500 small theorems while formalizing an existing proof of Fermat’s Last Theorem, the famously impossible problem that had consumed human mathematicians for generations. Other mathematicians had spent years trying to formalize that proof. In August, OpenAI claimed advances on ten other long-standing mathematical problems. The pace is not just accelerating; it is becoming impossible for a human mathematician to keep up. “I think the year 2026 is going to be remembered as either an annus mirabilis or annus horribilis for mathematics,” Strogatz says. “So much has happened.” He does not need to say which one he fears. He describes corporate labs in a race to grab headlines before blockbuster IPOs, using AI systems whose inner workings are not well understood. The result is a kind of vertigo: every week brings a result that would once have been a career-defining achievement, but no one has time to savor it. “You will not be able to compete without AI in the future if you want to do breakthrough math,” he adds. For a man who has spent his life teaching and writing about the beauty of mathematical reasoning, that sentence is almost unbearably sad. It means the human mind is no longer the most important instrument in the field. It means that to be a mathematician today is to be a manager of agents, not an explorer of ideas. And yet the science itself is thrilling, dazzling, full of possibilities. That contradiction—the beauty and the terror side by side—is what brings Strogatz to tears.
Townsend’s story shows what this moment feels like from inside. Toward the end of the interview, Townsend walks into Strogatz’s attic. He has already begun using AI in his own work. Recently he used ChatGPT to help solve a decades-old numerical linear algebra problem. Without the technology, he and his coauthor say, the volume of work and the cost-reward ratio would have been impossibly high. It was a genuine victory, a demonstration of AI’s power to accelerate research. But the victory came with a painful emotional price. “I actually feel kind of upset that I’ve dedicated 15 years of my life to research mathematics,” Townsend says, “and at a point in my career where I’m very productive and at my peak strength as a mathematician, that peak skill is no longer there. Something is able to surpass me.” His voice is not angry; it is mournful. He describes what it used to feel like to be a world-class mathematician: the excitement of pushing back the frontier of knowledge, the knowledge that you were one of the few people on earth capable of seeing something new. Now that feeling is gone. “I’m the one with an AI agent,” he says. “Which feels very different actually. And I feel totally threatened by it.” There is something deeply human about this confession. It is not just about math. It is about what happens when a craft, a vocation, a way of life is suddenly made obsolete by a machine. Townsend is young, productive, at the height of his powers—and he knows that no amount of talent or training can match what a model can do. The tools that should be liberating him are, in a sense, making him feel invisible. He is still doing research, still publishing, still contributing. But the frontier no longer belongs to him. It belongs to something else, something that does not need sleep, or courage, or intuition, or the quiet thrill of a proof that finally clicks.
For Strogatz, the feeling is darker and more existential. He likens the current moment for mathematics to a horror movie. The AI systems are creeping closer, and no one fully understands what they are doing or why. The models are not just calculators; they produce arguments that sometimes work, sometimes fail, and often cannot be explained. Mathematicians are being asked to trust results generated by black boxes. The idea of proof, the very foundation of the discipline, is becoming unstable. What does it mean to say a theorem is true if no human has checked every step? What does it mean to be a mathematician if the deepest discoveries are made by systems that do not know they are making them? These are not abstract questions. They are the daily reality of a field in crisis. Strogatz has spent his career making math feel human, writing books that show the beauty of pattern and logic. Now he is writing a book about how math is slipping away from human understanding. The book was probably conceived as an explanation, a warning, maybe even a celebration of the strange new powers of AI. But living through 2026 has made it feel like a eulogy. “But we’re not at the end of the movie yet,” he says. “Instinctively, I’m really terrified.” The horror-movie metaphor is apt. In a horror film, the monster is often most frightening when it is still mostly offscreen. You know it is coming because you have seen the signs, but you don’t know exactly what it will look like or what it will take from you. Strogatz and Townsend are in that liminal space. They have seen enough to be awed and enough to be afraid. They are not claiming that AI is evil, or that mathematical progress is bad. But they are witnessing the end of a world in which human mathematicians were the heroes of their own story. That is a loss that cannot be measured in theorems.
The broader picture is even more unsettling. The breakthroughs are happening inside corporations, not universities. The credit disputes are not just academic quarrels; they are battles over intellectual property, funding, and reputation in a world where a single announcement can move markets. OpenAI, Anthropic, and other labs are racing toward enormous public offerings, and mathematics has become a stage for their ambitions. A 90-year-old problem with a million-dollar prize is a perfect publicity stunt. The fact that the solution builds on the work of human mathematicians—Córdoba and Martínez-Zoroa—is easy to forget in the noise. Buckmaster’s accusations, whatever their truth, reveal a field where the old norms of openness and collegiality are breaking down. Scientists used to share ideas and wait for peer review; now they worry about being scooped by an AI company that can mobilize tens of thousands of agents overnight. Strogatz says the science is thrilling, but the human unpleasantness is impossible to ignore. It is not just about bruised egos. It is about power. The people who build the models control the means of mathematical production. The people who spent their lives learning the craft are being reduced to spectators or, at best, junior partners. Townsend’s feelings of threat are rational. If a machine can prove theorems faster and more reliably than a human, what value remains in human mathematical intuition? Strogatz would say, perhaps, that the value is in asking questions, in framing problems, in deciding which truths matter. But even those roles are being automated. The AI agents do not just solve problems; they are beginning to suggest which problems are worth solving. The human mind is no longer the center of the mathematical universe. It is a tiny, fragile thing, watching something vast and incomprehensible take over.
So what does 2026 mean for humanity? Strogatz is not sure. He is terrified, but he has not given up. “We’re not at the end of the movie yet,” he says again, and the words carry both hope and dread. In a horror movie, the survivors can still change the ending. They can find a way to understand the monster, or at least to live with it. Strogatz’s instinct is terror, but his work—teaching, writing, talking to people like Townsend—is an attempt to keep the human voice in the conversation. The year 2026 may be remembered as an annus mirabilis, a miracle year, when AI solved problems that had haunted humans for generations. Or it may be remembered as an annus horribilis, when the field of mathematics lost its soul. The truth is probably somewhere in between. The breakthroughs are real. The speed is astonishing. The beauty of the science remains. But so does the grief. Strogatz cries not because he hates AI, but because he loves mathematics. He loves the idea of a person sitting with a pencil and a problem, slowly, patiently, feeling their way toward a truth. That image is not dead, but it is no longer central. The frontier of knowledge is now being pushed forward by agents that do not share that love, that do not feel joy or doubt or fear. The human mathematicians are left to watch, to verify, to clean up, and to wonder what their own minds are for. Townsend is upset; Strogatz is terrified. They are not alone. The rest of us, who are not mathematicians, can feel the same shudder. If AI can replace the people who push back the frontier of mathematics, no profession is safe. The tears in that attic are for all of us. The movie is not over, but the monster is in the house, and the mathematicians are the first to hear its footsteps.
