An AI found the proof to a quantum computer puzzle that stood open since 2008
A research system running a non-public Anthropic model proved the Aaronson-Ambainis conjecture, and a computer program checked every step.

An important math puzzle about quantum computers has been solved, and an AI system found the proof. Physical Superintelligence, or PSI, says its research system proved the Aaronson-Ambainis conjecture. The puzzle had been open since 2008. The system ran on an Anthropic model that is not available to the public.
The question was easy to ask and hard to answer. Quantum computers can be far faster than normal computers on some problems. But those big wins usually need the input to hide a rare pattern. Could a quantum computer still win big on random inputs? Nobody knew for sure.
Now there is an answer. On yes-or-no questions, a quantum computer cannot be hugely faster than a normal one unless the input has a rare pattern. Here, hugely means a gap that grows very fast as the problem gets bigger.
PSI says the system had a complete proof on the third day of its run. A program called Lean, which checks every step of a proof by computer, accepted it. The system then improved the result twice, and Lean accepted the final version.
Four researchers at Stanford University are reading the proof and have sent comments. They called it "a huge result."
People still guided the work. A PSI researcher built the system, ran it, reviewed the draft and is responsible for the paper's correctness. An advisor also audited the proof. As PSI puts it, quantum computing still has plenty of mysteries, but now there is one fewer.
Open it and check the claim yourself. That is the point.