AI solved one of math's hardest problems. Humanity learned nothing (so far)

Live updates

1. 1h ago

   AI Solves Math Problem, Leaving Humans Confused

   Artificial intelligence systems have solved complex mathematical challenges, including aspects of the Navier-Stokes problems, but human researchers struggle to understand the resulting proofs. While tech firms celebrate these computational milestones, mathematicians warn that the opaque methods used by artificial intelligence offer little new understanding for human science. Critics and editorial boards point out that these systems generate solutions that remain borderline incomprehensible to experts, highlighting a growing disconnect between automated problem-solving and human comprehension in modern scientific research.

   Why it matters

   The intersection of artificial intelligence and advanced mathematics has sparked intense debate among researchers regarding the true value of machine-generated proofs. Tech companies are increasingly applying advanced models to historical scientific challenges, yet the inability of human experts to decipher the reasoning threatens collaboration. Organizations like OpenAI are now attempting to bridge this gap by establishing advisory groups to consult elite mathematicians.

   What is confirmed

   • Artificial intelligence agents produced a solution to a difficult math problem related to the Navier-Stokes equations.
   • Mathematicians describe the artificial intelligence-generated proof as nearly incomprehensible, leaving human experts struggling to make sense of the methodology.
   • OpenAI established an advisory group on mathematics and artificial intelligence to consult elite mathematicians.

   Still unconfirmed

   • Anthropic's Claude discovered a novel enzyme system with properties reminiscent of CRISPR that experts missed, though its exact function remains unknown.
   • OpenAI specifically solved the wrong version of the Navier-Stokes problem during its computational run.

   What to watch next

   • Published findings from the Advisory Group on Mathematics and Artificial Intelligence
   • Further peer reviews and expert breakdowns of the artificial intelligence Navier-Stokes proof
   • Responses from tech firms regarding how they plan to make machine-generated proofs interpretable
   Sources used for this update (9)

   1. The Guardian — The Guardian view on AI v mathematicians: humans are still vital to the field, but tech firms refuse to see that | Editorial
   2. openai.com — Advisory Group on Mathematics and Artificial Intelligence
   3. Scientific American — Did OpenAI solve the wrong Navier-Stokes problem?
   4. The New York Times — Opinion | Why A.I. Is Dangerous for Science, Too
   5. NPR — AI solved one of math's hardest problems. Humanity learned nothing (so far)
   6. The Washington Post — Opinion | Is math doomed? Don’t count on it.
   7. The Verge — OpenAI wants to consult elite mathematicians about how to not fumble again
   8. futurism.com — Mathematicians Can’t Make Sense of How OpenAI’s Agents Solved One of the Toughest Math Problems Because the AI’s “Proof” Is Borderline Incomprehensible
   9. www.yahoo.com — Claude Found a Novel Enzyme System Experts Missed, and Nobody Knows What It Does
   confidence 90%