Radiolab
Radiolab

Math Vs Machine

October 9, 2026

AI Summary

5 min read

In late September 2026, OpenAI announced it had solved one of the seven Millennium Problems in mathematics—the Navier-Stokes problem, a 150-year-old set of equations describing how fluids flow. For mathematicians watching this unfold, it was not a moment of triumph. Cornell mathematician Steve Strogatz described himself as "terrified." The problem wasn't that AI had gotten good at math. It was that AI had gotten good in a way that threatened something fundamental about what it means to be a human who tries to understand things.

The rapid ascent of a machine mathematician

Strogatz had been tracking AI's math abilities since ChatGPT first appeared in late 2022. At first, it couldn't do elementary school arithmetic. Over time, it progressed through developmental stages: clumsy algebra, then high school math, then college-level reasoning. By spring 2026, it was as good as a first-rate mathematician. Then in May, OpenAI's model solved the unit distance problem—a long-standing puzzle in discrete geometry about the maximum number of pairs of pegs on a flat board that can be exactly one unit apart. The solution was not a brute-force calculation. Experts called it "beautiful" and "elegant." It connected two branches of mathematics that no one had realized were linked. "If a person had come up with this," Strogatz said, "we would have called that person brilliant."

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (01:29) **Opening the Existential Question** - Host Soren Wheeler pulls mathematician Steve Strogatz away from Twitter/X to discuss a bombshell AI announcement that has the math community in crisis.
  • 2 (03:41) **The Slow Climb: AI's Math Journey** - Steve traces the evolution of AI's mathematical ability from November 2022 to mid-2026, showing a shocking acceleration.
  • 3 (06:42) **The First Breakthrough: The Unit Distance Problem** - In May 2026, OpenAI's model solves a long-standing open problem in discrete geometry, stunning mathematicians with its creativity.
  • 4 (11:04) **The Traumatizing Summer** - Following the unit distance solution, AI systematically solves harder and harder problems throughout the summer of 2026, creating a "traumatizing effect" on the math community.
  • 5 (11:48) **The Millennium Problem: Navier-Stokes** - Just two days before the conversation, OpenAI announces it has solved the Navier-Stokes problem, one of the seven most cherished "mountaintops" in mathematics.
  • 6 (14:44) **The Human Context and Opacity** - The Navier-Stokes solution builds on the work of two Spanish mathematicians, but the AI's method is so complex it's nearly incomprehensible.
  • 7 (17:14) **The Value of the Struggle** - Steve argues that the quest itself, the struggle to understand, is what makes mathematicians better, even if they never reach the summit.

+ Full timestamped outline available in the app

Show Notes

In less than four years, mathematician Steve Strogatz watched AI go from flailing at basic arithmetic to solving problems that baffle the world’s best mathematicians. In September of 2026, Open AI announced it had solved a Millennium Prize Problem, one of math’s highest achievements. In short, mathematics is right now on the cusp of a reckoning that may be coming for all of us. In this episode, Steve grapples with the idea that the window might be closing on humanity’s effort to unpuzzle, or learn, or understand the world around us.

Special thanks to Ravi Vakil, Alex Kontorovich, Bob Wright.

EPISODE CREDITS: 
Reported by - Maria Paz Gutierrez and Soren Wheeler
Produced by - Soren Wheeler and Maria Paz Gutierrez
Fact-checking by - Diane Kelly
and Edited by  - Pat Walters

EPISODE CITATIONS:

Videos -  

Articles - 

Books - 

Sign up for our newsletter!! It includes short essays, recommendations, and details about other ways to interact with the show. Signup (https://radiolab.org/newsletter)!

Radiolab is supported by listeners like you. Support Radiolab b

Radiolab

More from this podcast

Radiolab →