When I first saw OpenAI's announcement on Saturday, it felt like a punch in the gut. OpenAI said that an internal AI model had solved or made substantial progress on 10 major open problems in mathematics.

For the first time, I viscerally felt what it's like for AI to get really good at something I care about. Growing up, I always thought I'd be a mathematician — even these past few years have sometimes felt like a break from my inevitable math PhD. Yet at the same time, I was very glad that I'm a journalist now.

In mid-July, I went to the International Congress of Mathematicians, the traditionally the largest math conference in the world[1]. I spoke with over 20 mathematicians about how they are responding to AI developments. There were a range of answers: some people found LLMs to be a useful tool; others worried for the future of the profession.

AI’s effects on mathematics is a complicated topic and it's easy to focus on the flashy announcements. But I learned a lot talking with mathematicians, and I hope you'll learn a lot reading my piece: