I am sad about AI in maths...

Summary

Sophie Maclean reflects on her sadness and uncertainty after OpenAI announced solutions—some formally verified—to a range of previously unsolved mathematical problems. She emphasizes that she is describing her own feelings, not speaking for the whole mathematics community.

Why the developments worry her

• Career uncertainty: As a final-year PhD student without a permanent position, Maclean is approaching the academic job market at a time when nobody knows how AI will change mathematical work. She asks that the concerns of PhD students, postdocs, and other insecurely employed researchers be taken seriously.
• Mathematics as a vocation: She argues that many mathematicians pursue the subject because they love it, not for money or status. The possibility that people may be unable to continue a career they have built their lives around is therefore personally painful.
• A fast-changing landscape: AI companies have tested systems on collections of open problems, including a website of problems associated with Thomas Bloom. Early announcements sometimes involved finding existing solutions rather than producing new ones; more recently, genuinely new solutions have appeared. Maclean notes that a problem’s being open does not necessarily mean it is difficult—it might simply be uninteresting or overlooked.
• Scale and computing resources: She points to the announced Navier–Stokes result as an example of the resources AI-assisted work can involve. Her point is not that AI is necessarily more intelligent than people, but that it can bring together far more time, computing power, and contributors than an individual mathematician.

What she values about mathematics

Maclean loves the process of doing mathematics: struggling with a problem, thinking deeply, learning, experimenting, and eventually finding a solution. She worries that when AI produces answers quickly, it can reduce the uncertainty and discovery that make mathematics meaningful to her.

She frames the issue as a difference in what people value:

• If mathematics is mainly about results, AI’s achievements may be an exciting advance for science.
• If mathematics is also about the process, automating solutions can feel like a loss.

The Snaky puzzle and the Sudoku analogy

A recent video of Maclean’s discussed a tic-tac-toe-style puzzle involving hexominoes, including an open question about a piece called Snaky. OpenAI solved that question. Maclean says the puzzle had attracted interest and engagement from viewers, including people who sent her strategies and even a 3D-printed Snaky piece. Although some questions about the puzzle may remain, the AI solution has changed the experience for her and made her reluctant to produce a follow-up video.

She compares this to someone repeatedly solving difficult Sudoku puzzles in a newspaper. The first solution might be impressive, but repeated demonstrations can feel less like discovery and more like point-scoring. In her view, the world is not necessarily better off simply because another recreational puzzle has been solved; the process and shared excitement mattered too.

Her overall position

Maclean is not claiming that AI has no value. She recognizes that AI could contribute to important scientific advances. Her sadness concerns what may be lost when problem-solving becomes automated, especially in areas where the activity itself is a source of curiosity, creativity, and community.

She describes her reaction as grief and says she fears losing the kind of mathematics she loves. At the same time, she hopes people will continue to value mathematics as a human activity: testing what people can achieve, collaborating, exchanging ideas, and sharing the pleasure of discovering something new.

Speakers and sources mentioned

• Speaker: Sophie Maclean, a mathematics communicator and final-year PhD student working in analytic number theory and additive combinatorics.
• Organizations and systems discussed: OpenAI and AI systems used to solve mathematical problems.
• People and resources mentioned: Thomas Bloom, creator of a website listing open mathematical problems. Maclean also refers to a solution announced for the Navier–Stokes problem and to her own video about the Snaky puzzle.