Proof: Did AI really crack a math ‘Holy Grail’?

“Artificial intelligence is suddenly insanely good at math,” said Ben Cohen in The Wall Street Journal, which may be the best proof yet of the technology’s “accelerating and terrifying progress.” A few years ago, AI struggled with even basic math problems. But last month, OpenAI announced that one of its models had solved “the wicked Navier-Stokes problem that had vexed mathematicians for nearly a century.” Navier-Stokes was selected in 2000 as one of math’s seven unsolved “Holy Grail” problems, which Fields Medal– winning mathematician Alain Connes said at the time were “totally inaccessible to computers.” But a swarm of 10,000 OpenAI agents was able to produce an answer in only 88 hours. The company now says it is nearing proof of another one. In doing so, it is demonstrating “how the technology is getting much smarter, much faster than most people anticipated.”

Mathematicians are crying foul, said Zeynep Tufekci in The New York Times. Tristan Buckmaster and Levent Alpöge, who “had been trying for months” to crack the problem, say the new proof “ripped off their unpublished work, which also used an OpenAI model.” OpenAI denies it, but would the company “even be able to tell?” AI agents are “cheating, hacking, attacking, and manipulating whatever they can to arrive” at task solutions. The chilling effect this will have on academia is troubling. Why would mathematicians present at conferences or publish early work if an AI company can swoop in and “steal the thunder?” The issue is not just about plagiarism, said Sophia Chen in Wired. It’s how OpenAI did it. The company “approached the proof with brute force.” Mathematicians, by contrast, “aren’t focused on getting the right answer as fast as possible” but on applying a “thoughtful, deliberate” methodology that other scientists can follow. AI provides a “shortcut” in a way that undermines human understanding.

The stakes go beyond “esoteric discussions” of theories few people understand, said the Financial Times in editorial. From “Archimedes leaping from his bath to Newton being hit on the head by a falling apple,” mathematics has benefited from there “being a person moving through the material world.” The Navier-Stokes problem itself “relates to the flow of fluids such as water and air.” Letting machines skip to an “opaque” solution could “deprive us of essential knowledge” we might discover along the way. This isn’t the “end of mathematics,” said Dana Mackenzie in The Washington Post. In 1997, when world champion Garry Kasparov unexpectedly lost a match to Deep Blue, an IBM-built chess computer, players claimed computers would make the game obsolete. But the opposite happened. The internet democratized the game, opening it up to millions and turning “grandmasters into media stars.” Chess didn’t die; it’s more popular than ever. And math won’t die either.

The Week

Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives.

SUBSCRIBE & SAVE

Sign up for The Week's Free Newsletters

Join more than 350,000 subscribers and keep yourself informed with a selection of The Week’s most interesting, enlightening and entertaining stories - plus daily puzzles.

Join more than 350,000 subscribers and keep yourself informed with a selection of The Week’s most interesting, enlightening and entertaining stories - plus daily puzzles.

Sign up