On 7 Nov 2026, Physical Superintelligence lab announced it used an unreleased Anthropic AI model to prove the Aaronson-Ambainis conjecture, an open problem in quantum computing since 2008.
The problem asks: when can quantum computers be much faster than regular computers? The research system built by the lab proved the conjecture in 3 days.
The proof shows that huge quantum advantage needs special patterns and structures. That means quantum computers have clear limits, while they will still be much stronger in specific tasks.
And there is an important conclusion, which is that AI is no longer only a tool for calculation, it became a research colleague that can link scattered ideas in math and physics. And that means problems that scientists thought need one genius in one time, may now need a hybrid team of a human mind that asks the question, and a machine mind that tries thousands of paths in days. And the most striking part is that the limits of quantum computing were drawn by an AI tool, meaning the tool that helps us understand the future is the same tool that sets the ceiling of that future.
