Physical Superintelligence (PSI), a startup lab, says it used early access to an unreleased Anthropic model, Fable 5.2, to solve the Aaronson-Ambainis conjecture, a quantum computing problem that has stood for 18 years. David Turturean, a recent MIT graduate and member of PSI's technical staff, led the work.
The conjecture concerns how far quantum computers can outpace classical ones. The result appears to rule out one route by which quantum machines might solve certain hard problems quickly, though not all such routes. Turturean described it as a negative result that tells researchers where to stop looking.
Scott Aaronson, who co-posed the problem, wrote that he does not fully follow the proof but believes it is correct. The solution was autoformalized in Lean, a language for checking mathematics, and PSI walked a group of leading researchers through it. Aaronson said he felt satisfied, awestruck, disappointed and fatigued, and noted he was surprised the problem had not appeared among the solutions OpenAI released a few weeks earlier.
PSI's approach pairs models with specialised harnesses, tools, memory systems and simulation infrastructure. CEO Matt Pines said the aim is to attack a problem with both breadth and depth. Today the company earns revenue by simulating heat, power, water and sound to optimise data centres. Its longer-term goal is turning theoretical physics findings into working technologies.
Pines acknowledged that physics is harder than mathematics because results must be validated against the real world, and said PSI is working on testing computational outputs with experiments efficiently. Separately, the startup Discovered Materials last week announced a $1M grand challenge for researchers using AI on problems such as superconductivity and heat dissipation.
Because PSI held its result until Anthropic's model was close to release, it could share the work with experts first. Jim Halverson, a Northeastern University professor who advises PSI, said best practices are still emerging and that transparency seemed the right course. He will attend an MIT summit next week where physicists will discuss how to respond to the faster pace of science.
