AI helps prove a 16-year-old quantum computing puzzle in three days

A small physics lab says AI has settled a quantum computing question that stumped researchers for 16 years. Physical Superintelligence Inc., known as PSI, said Friday that a research system built around an unreleased Anthropic model proved the Aaronson-Ambainis conjecture. The system finished the proof in three days.

What the puzzle is

A conjecture is an educated guess that nobody has proven yet. This one has been open since 2008. It deals with when quantum computers can be significantly faster than regular computers. Quantum computers do exist today, but they are still in the early stages of development.

What the proof shows

According to PSI, the proof shows that certain huge speed advantages from quantum computing need special patterns or structures in the problem. In plain terms, quantum computers have limits. Researchers long suspected that. Now it appears to be proven. They would still be a major upgrade for some tasks, just not for everything.

Researchers who had worked on the problem before, Guy Blanc, Jordan Docter, Carmen Strassle and Li-Yang Tan of Stanford University, called it a "huge result." They also said, per PSI's announcement, that "everything is brought together in a rather magical way."

Part of a bigger shift

This is the latest math breakthrough credited to AI. In September, OpenAI said it had solved the 90-year-old Navier-Stokes equations problem, which it called one of the deepest at the frontier of mathematics.

The change is unsettling people in the field. Terence Tao, a mathematics professor at UCLA, said last month that math's "romantic era" was over. He added that new things are possible, but some things are being lost. MIT researchers who polled grad students and postdocs found that 76% were unlikely to stay in academic math. The same share said AI had made them less motivated to do research.

What is not known yet

• Which unreleased Anthropic model PSI used is not clear.
• Anthropic did not respond to a request for comment.
• The report does not say how the proof has been reviewed beyond the Stanford reaction.

David Turturean, a member of PSI's technical staff, developed the research system. The announcement comes from PSI itself, so the claims come from the lab's own account.