Startup Physical Superintelligence (PSI) Leverages AI to Solve Long-Standing Quantum Computing Conjecture
Physical Superintelligence (PSI), a startup focused on AI-driven scientific discovery, has successfully utilized Anthropic’s Fable 5.2 model to solve the Aaronson-Ambainis conjecture. This eighteen-year-old problem in quantum computing concerns query complexity and how quantum computers outperform classical ones. The solution was validated through autoformalization in Lean, a mathematical coding language. PSI's business model involves using specialized AI harnesses and simulation infrastructure to tackle complex scientific problems. While the company currently generates revenue by optimizing data center variables like heat and power, its long-term strategy is to bridge the gap between theoretical physics breakthroughs and practical technological applications. The startup aims to create a feedback loop where AI suggests physical discoveries that can be validated through real-world experimentation.
Topics: Aaronson-Ambainis conjecture AI-driven scientific discovery Large Language Models Physical Superintelligence (PSI) Quantum computing