I’ll write romance novels if AI solves all maths problems, Chinese Fields winner jokes

Deng Yu, a 37-year-old Chinese mathematician who recently won the 2026 Fields Medal, jestingly joked in a social media post that if artificial intelligence solves all mathematics problems, he would instead turn his talents to writing yuri novels—a genre of Japanese literature focusing on female friendships and romantic relationships.

Deng's comment echoed the growing unease within the mathematical community regarding the potential impact of AI on the field. Just days prior, on September 8, OpenAI claimed to have solved a Millennium Prize Problem using a yet-to-be-released next-generation model. This problem, alongside six other unsolved mathematical questions, carries a $1 million prize and has remained unsolved since 2000.

The Navier-Stokes Millennium Prize Problem, which concerns fluid mechanics, is particularly significant as it could greatly enhance our understanding of mathematical physics if solved. OpenAI's proof, generated through an 88-hour process consuming 130 billion output tokens valued in the millions of dollars, has raised questions about the role of AI in mathematics.

However, some mathematicians, such as New York University's Tristan Buckmaster and Anthropic researcher Levent Alpoge, claim they had already proven the Euler equations—a crucial step towards unraveling the Navier-Stokes problem—before OpenAI made its announcement. OpenAI later clarified that while they had not directly accessed the work of Buckmaster and Alpoge, their models may have benefited from de-identified data generated by these mathematicians using OpenAI's tools.

This incident is not unique; earlier this year, OpenAI researcher Su Weijie used AI to advance the twin prime conjecture, a significant problem in number theory. His progress, from 246 to 240 and then 188, was achieved with the help of OpenAI’s GPT-6 Astra. While Su acknowledged his role was largely auxiliary, he noted that the AI's mathematical capabilities were still far from their limits.

The American Mathematical Society has acknowledged the groundbreaking nature of OpenAI's proof, recognizing the significant contributions of humans behind the technology. They emphasized that mathematics is fundamentally about human understanding and that the insights gained will continue to shape the field for years to come.