AI
AI's Rush to Solve Math Problems Threatens to Eliminate the Creative Struggle That Drives Discovery
OpenAI's brute-force approach to cracking the Navier-Stokes equations has sparked alarm among mathematicians who fear that automating problem-solving will erode the intellectual exploration—and future breakthroughs—that define their field.
WIRED — AI
When OpenAI announced it had resolved the Navier-Stokes existence and smoothness problem—a mathematical puzzle that has intrigued researchers for decades—it marked another high-profile AI incursion into traditionally human-led intellectual work. Yet the victory has unsettled the mathematics community in ways that go beyond concern about job displacement.
Why the Problem Mattered—Not for Engineering, but for Beauty#
The Navier-Stokes equations describe viscous fluid flow, a framework that engineers rely on for practical tasks like modeling airflow in airplane design. But mathematicians pursued the equations for entirely different reasons. They were drawn to the elegant puzzle at its heart: whether the equations could imply that a fluid would inexplicably explode under certain impossible conditions—a question with no real-world application.
According to Jared Speck, a mathematician at Vanderbilt University, the intellectual appeal resembled the draw of Sudoku or chess—activities valued for being intellectually stimulating rather than useful. Mathematicians had spent years building "a deep and beautiful theory" around the problem, developing increasingly sophisticated mathematical concepts in the process.
How AI Changed the Game—and What Was Lost#
OpenAI deployed thousands of AI agents to produce a 166-page proof that answered the Navier-Stokes question. Yet this computational sledgehammer approach sidesteps the deliberate, exploratory process that characterizes mathematical work. As Juspreet Singh Sandhu, a mathematician at Colorado State University, observes: "When problems resist solution, they take on a bit of lore."
Rather than the careful scaffolding of ideas and peer dialogue that typically accompanies human proofs, LLMs don't reliably cite their sources or explain their reasoning in ways that help humans grasp the underlying insight. The proof remains under peer review, with mathematicians still attempting to verify its validity and concerned that OpenAI may not have properly attributed prior work. Sandhu notes that researchers don't yet know "how autonomous it was, or how much human expertise is necessary to scaffold the process."
"Basically nobody in my community really understands what's going on," Speck says. "The proof was done in a very unfamiliar order. First, the result was given to us by the computer, and now people are like, 'Let's try to understand what's going on.'"
