What We Know About PSI's AI-Solved Quantum Computing Proof

A researcher at the startup lab Physical Superintelligence says an unreleased Anthropic model helped him solve a quantum computing problem that had stumped scientists for 18 years, according to TechCrunch.

David Turturean, a recent MIT graduate and member of the technical staff at Physical Superintelligence, known as PSI, used early access to Anthropic's Fable 5.2 model to address the Aaronson-Ambainis conjecture. Turturean told TechCrunch he deferred plans for a master's degree at MIT's computer science and AI lab to join the startup. "I'm totally accepting the idea that this might be the last paper I ever put out," he said. "In a few months, like, why would you even put out a paper?"

What is the Aaronson-Ambainis conjecture?

The conjecture, posed nearly two decades ago, concerns whether quantum computers can solve certain problems meaningfully faster than classical computers. The result PSI produced appears to rule out one path researchers had explored for achieving that speedup, though not every possible route. "It's a negative result that allows people to know, okay, we're no longer looking there," Turturean said.

How did researchers verify the AI-generated proof?

Scott Aaronson, one of the two researchers who originally posed the problem, wrote that he does not fully understand the proof himself but believes it is correct because it was autoformalized in Lean, a coding language used to check mathematical arguments, and because a team of researchers reviewed it after PSI shared the solution. Aaronson described feeling "satisfied, awestruck, disappointed, and fatigued" by the news. "A proof of full [Aaronson-Ambainis conjecture] was one of the last remaining open problems in quantum query complexity that everyone seemed to agree was centrally important," he wrote, adding that he was surprised OpenAI's recent "Mathocalypse" release of solved problems had not included it.

Why does PSI think its approach works where others haven't?

"An effective kind of research system should have lots of things related to harnesses, lots related to tools, memory systems, simulation infrastructure," PSI CEO Matt Pines told TechCrunch. "You want to be able to really go after a problem at breadth and depth, and be able to consider a wide range of potential alternatives."

Pines framed the result as evidence for PSI's bet that specialized model harnesses and scaffolding, not raw model scale alone, can crack open problems general-purpose systems miss.

How does PSI make money right now?

The company's current revenue comes from a narrower application: using its tools to optimize data centers by simulating variables such as heat, power, water and sound, factors Pines said determine both a site's performance and its effect on surrounding communities.

What is PSI's longer-term business plan?

The company's stated goal is converting theoretical physics breakthroughs into deployable technologies. Pines said that pitch was a harder sell to investors as recently as six to nine months ago. "I had to sort of stretch to validate that conditional, right?" he said. "Like, 'no, no, no, trust me, pretty soon, theoretical physics is going to be mostly the domain of machines.' And it's not there yet, but it's on its way."

TechCrunch's full report, including additional detail on the proof's scope, is available here.