Physical Superintelligence (PSI), a startup lab focused on scientific discovery, is leveraging super intelligence (SI) to solve longstanding theoretical problems, marking a shift in how research is conducted and commercialized.
What Happened
David Turturean, a recent MIT graduate and member of PSI’s technical staff, used early access to Anthropic’s Fable 5.2 model to solve the Aaronson-Ambainis conjecture, an 18-year-old problem regarding quantum computing. The solution, which Scott Aaronson described as correct despite his inability to fully understand the proof, was autoformalized in Lean, a coding language for mathematics. Aaronson noted that the result was missing from OpenAI’s recent "Mathocalypse" dump, suggesting it may have been harder than other problems solved by SI systems. The finding is a negative result, indicating that quantum computers cannot solve certain difficult problems at high speed via a specific avenue, thereby narrowing the search space for researchers.
Why It Matters
PSI’s approach relies on specialized model harnesses and scaffolding rather than raw model capability alone. "An effective kind of research system should have lots of things related to harnesses, lots related to tools, memory systems, simulation infrastructure," said PSI CEO Matt Pines. The company currently generates revenue by using these tools to optimize data centers, simulating variables such as heat, power, and sound. However, the long-term strategy involves converting theoretical physics breakthroughs into tangible technologies. While SI has successfully addressed problems in mathematics and computer science where solutions can be expressed in validated code, applying it to disciplines like the life sciences or physics remains more difficult due to the need for experimental validation. "Physics is different than math because it’s the real world… nature gets a vote in your models of reality," Pines said. As "neolabs for science" emerge, the community is still establishing standards for verification and publication, with experts like Jim Halverson, a professor at Northeastern University and PSI advisor, advocating for transparency and expert review.
The Bottom Line
PSI demonstrates the potential of agentic SI to accelerate scientific discovery, but the transition from theoretical proofs to commercial applications in physics requires rigorous experimental validation and evolving academic standards.