AI Solves Math Problems Without True Understanding

OpenAI announced this month that a swarm of thousands of AI agents had cracked the Navier-Stokes existence and smoothness problem, one of the seven Millennium Prize puzzles that have stumped mathematicians for decades. The claim landed like a gut punch in a field that had assumed itself insulated from the disruption already hitting musicians and actors. Suno can now churn out elevator music indistinguishable from the real thing.

Tilly Norwood, an AI-generated performer, delivers scripted one-liners in customer-service ads. Juspreet Singh Sandhu, a mathematician at Colorado State University, put it bluntly: "The artists and the musicians have already gone through this." What makes the Navier-Stokes case sting differently is what it exposes about how mathematics actually works. Despite the stopwatch drama of high school exams, mathematicians rarely prize speed.

The real work looks less like calculation and more like invention, closer to designing a puzzle than solving one. OpenAI's approach, throwing enormous computational force at the problem, skipped that slow, deliberate process entirely. The question now is whether a correct answer, arrived at without the human reasoning behind it, actually counts as understanding.