PSI Researcher Solves 18-Year Quantum Conjecture With Anthropic’s Fable 5.2
Updated
Updated · TechCrunch · Nov 10
PSI Researcher Solves 18-Year Quantum Conjecture With Anthropic’s Fable 5.2
3 articles · Updated · TechCrunch · Nov 10
Summary
David Turturean, a recent MIT graduate now at Physical Superintelligence, used early access to Anthropic’s Fable 5.2 to prove the 18-year-old Aaronson-Ambainis conjecture.
Scott Aaronson said he does not fully understand the proof but believes it is correct because it was autoformalized in Lean and reviewed with top researchers PSI consulted.
The result is a negative one for quantum computing: it shuts off one proposed route for quantum speedups on certain hard problems, while leaving other paths open.
PSI says the breakthrough supports its strategy of pairing frontier models with specialized research harnesses, tools and simulation systems rather than relying on a model alone.
The episode highlights AI’s growing role in math-heavy science, even as PSI and academics say standards for validation, disclosure and experimental follow-through are still being worked out.
If an AI solves a complex quantum problem but top experts barely understand the proof, who truly owns the scientific discovery?
How exactly does closing this theoretical quantum loophole impact the billion-dollar race to build practical, real-world quantum computers?
Will AI's ability to autoformalize complex math in Lean make traditional academic research papers completely obsolete within the next decade?
AI System Proves 18-Year Quantum Conjecture: PSI and Anthropic Resolve Aaronson-Ambainis, Setting New Limits on Quantum Speedups
Overview
In November 2026, PSI's Argon system, engineered by David Turturean and powered by Anthropic's Fable 5.2 model, produced a complete proof of the 18-year-old Aaronson-Ambainis conjecture, showing that quantum algorithms cannot achieve exponential speedups on unstructured random inputs. The proof was autoformalized in Lean 4 and independently verified by multiple kernel implementations, then published openly with code and audit scripts. Stanford researchers began auditing the work, while the result prompted a shift in quantum computing research and investment toward structured, domain-specific problems. This breakthrough is part of a wave of AI-driven advances resolving major open questions in mathematics and physics.