OpenAI recently achieved a solution to the Navier-Stokes existence and smoothness problem, one of mathematics’ most enduring challenges, the company says. The internal AI model reportedly cost more to produce the result than the $1 million prize offered by the Clay Mathematics Institute for its solution.
OpenAI research scientist Noam Brown predicts that within a year, an AI capable of solving problems of this caliber will be accessible to anyone with a $20/month ChatGPT subscription. “Massively scaling test-time compute gives us a glimpse of the future,” Brown stated on September 8, 2026, prompting both excitement and concern within the UC Berkeley math community.
OpenAI’s $1 Million Navier-Stokes Solution Disrupts Mathematical Norms
The internal AI model employed by OpenAI to solve the Navier-Stokes problem used an approach different from traditional mathematical methods, using approximately 10,000 AI agents working in parallel. The speed with which OpenAI arrived at a solution has prompted reassessment of timelines within the mathematical community.
UC Berkeley math professor Richard Borcherds, a Fields Medal recipient, expressed surprise at the rapid progress, admitting he had only anticipated such a breakthrough in another decade. Borcherds also predicted a future where human mathematical research may become redundant, a sentiment fueled by the AI’s accelerating capabilities.
He acknowledged the anxieties of current graduate students, noting, “Grad students are not stupid — this has occurred to all of them, and quite a lot of them are dropping out or getting very worried about it.” However, not all mathematicians share this concern. Robin Huang, a Ph.
D. candidate at UC Berkeley, remains skeptical of the AI’s true capabilities, describing it as “very stupid and very inefficient” despite its success with the Navier-Stokes problem. Huang views the pursuit of unsolved problems by AI companies as exacerbating existing pressures within the math community and supporting a climate of fear surrounding being “scooped.” Despite these reservations, Huang maintains a degree of optimism, stating that in a year AI could actually facilitate learning.
Borcherds, while acknowledging the potential disruption, emphasized the importance of intrinsic enjoyment in mathematics. “People are complaining that AI is pushing out this 100,000-line formal proof of some major conjecture. I don’t really see what harm this does because you can always ignore this if you want, and then you’re no worse off than before,” he said.
He further suggested that the current anxieties stem from a realization of intellectual limitations, adding, “I think what people are getting upset about is they’re slowly realizing that they are dumb and somehow blaming this on AI, even though it’s not really AI’s fault.” The accessibility of such powerful AI tools is also rapidly changing. This democratization of advanced problem-solving capabilities raises questions about the future of mathematical education and the role of human intuition in a field increasingly dominated by algorithmic solutions.
Borcherds even believes AI would be a more effective instructor for his graduate-level abstract algebra course, citing its patience and clarity as advantages over his own teaching style. He stated he receives “much better answers from it than he gets from any human mathematician” when posing complex questions. Huang, while valuing the creative aspects of mathematics, acknowledges the potential for AI to assist with routine tasks, but stresses the importance of independent work.
“I don’t care if you do your homework or not. But if you don’t do your homework by yourself, you will fail the exam,” she tells her students.
She also suggests the current situation may be a catalyst for the mathematical community, prompting a reevaluation of its standards and incentives. She added, “There are other things in life that make me happy. Even if — which I do not think will happen — but even if AI completely takes over math, there are things in life that may not be, and life goes on.”.
It’s fairly obvious that if things go like this, in a few years, humans doing math research will be redundant.
Borcherds, who currently teaches MATH 250A, a graduate-level course on abstract algebra, also
UC Berkeley Mathematicians Debate AI’s Impact on Research and Future Careers
Borcherds voiced concerns about the future trajectory of mathematical research, stating, “It’s fairly obvious that if things go like this, in a few years, humans doing math research will be redundant.” He noted a growing anxiety among graduate students, observing that many are either abandoning their studies or experiencing significant worry regarding their future prospects within the field. The implications for aspiring mathematicians are prompting difficult conversations, as Borcherds believes AI will consistently outperform new entrants to the discipline.
“If you’re starting off as a graduate student in math, it’s likely that AI will always be better than you at mathematics,” he said, acknowledging a fundamental shift in mathematical expertise. This assessment has led to observable changes within the UC Berkeley program, with a noticeable number of students reconsidering their commitment to advanced study.
Huang said she thinks many graduate students and postdocs in math are concerned about job prospects, but she remains optimistic about the future. In fact, in a “perfect vacuum,” she said, if AI’s abilities were to stay the same, she would be able to learn a lot more about mathematics. She also criticized the motivations behind AI companies’ focus on solving high-profile mathematical problems, describing it as an action that encourages competition and fear of being preempted within the mathematical community. Huang emphasized the uniquely creative aspect of mathematics, distinguishing it from experimental sciences.
She believes this creative element is currently beyond the reach of artificial intelligence, stating that mathematics is fundamentally about developing individual perspectives. Despite her reservations, Huang encourages her students to use AI as a tool, stating, This pragmatic approach highlights a willingness to adapt to the changing environment while still valuing the core principles of mathematical inquiry.
If you’re starting off as a (graduate) student in math, it’s likely that AI will always be better than you at mathematics.
Borcherds, who currently teaches MATH 250A, a graduate-level course on abstract algebra, also
See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
