PSI used early access to Anthropic's Fable 5.2 to resolve the eighteen-year-old Aaronson-Ambainis conjecture

Physical Superintelligence (PSI), a startup science lab, used early access to Anthropic's Fable 5.2 model to solve the Aaronson-Ambainis conjecture, an eighteen-year-old problem about quantum computing, TechCrunch reports.

The work was done by David Turturean, a recent MIT graduate and member of PSI's technical staff. He deferred plans for a master's at MIT's computer science and AI lab to join the company.

The result

The conjecture concerns the ability of quantum computers to solve problems faster than classical ones. According to the report, the solution appears to close off one avenue quantum computers could follow to solve certain difficult problems at high speed, but not all of them. "It's a negative result that allows people to know, okay, we're no longer looking there," Turturean said.

Scott Aaronson, one of the two researchers who posed the problem, wrote that he does not fully understand the proof but believes it is correct. His reasons: it was autoformalized in Lean, a coding language for math, and explained by a team of top researchers with whom PSI shared the solution. He called a proof of the full conjecture "one of the last remaining open problems in quantum query complexity that everyone seemed to agree was centrally important." He described feeling satisfied, awestruck, disappointed and fatigued. Aaronson also said he was surprised the problem was missing from what he called OpenAI's "Mathocalypse dump" a few weeks earlier.

PSI held the result until Anthropic's model was nearly ready for release, which gave it time to share the proof with relevant researchers. Jim Halverson, a Northeastern University professor who advises PSI, said: "Best practices are still emerging... We figured that transparency is a good way to go, and contacting experts is a good way to go." He will attend an MIT summit next week where leading physicists will discuss a response to the new pace of science.

PSI's approach

PSI's model is to solve scientific problems with specialized harnesses and scaffolding around models. CEO Matt Pines said an effective research system needs "lots of things related to harnesses, lots related to tools, memory systems, simulation infrastructure."

The company's current revenue comes from using its tools to optimize data centers. It simulates heat, power, water and sound, where site performance and community impact are, in Pines' words, "a direct function of its physics." The longer-term goal is to convert theoretical physics breakthroughs into technologies.

Pines acknowledged that physics is harder than math for this approach. "Physics is different than math because it's the real world... nature gets a vote in your models of reality," he said. PSI is focused on validating computational results with experiments as efficiently as possible. He said theoretical physics being mostly the domain of machines "is not there yet, but it's on its way."

The report also notes that Discovered Materials, another startup, released a $1 million grand challenge last week for researchers using AI on problems such as superconductivity and heat dissipation.

Turturean said the delay before release let him rework an AI-elicited solution so it could be presented more naturally to humans. He expects the academic community to ask him to explain the result.

What this means

The result is a documented case of a model-plus-harness system resolving a named open problem in quantum query complexity, with a Lean formalization as the check on correctness. Aaronson's acceptance rests on that formalization and on expert explanation, not on his own full understanding of the proof.

The gap Pines describes between math and physics is the practical constraint. Lean can validate a mathematical proof, but physical claims need experiments, which is why PSI says it is building loops between AI-suggested ideas and experimental validation. Its current revenue comes from physics simulation for data centers, not from theoretical breakthroughs.

The release-timing decision also shows researchers working out norms in real time. PSI delayed publication to consult experts, and Halverson says best practices are still emerging.