AI model helps solve quantum computing conjecture

A research system utilizing an unreleased Anthropic artificial intelligence model has successfully proved the Aaronson-Ambainis conjecture, a longstanding problem in quantum computing. The proof was completed in three days by a system developed at Physical Superintelligence Inc., a physics research startup that launched in September 2026.

The conjecture, which was first posed in 2008 by Scott Aaronson and Andris Ambainis, addresses whether quantum computers can achieve exponential speedups over classical computers when processing random, unstructured inputs. The proof demonstrates that quantum computers cannot save exponentially many input reads for yes-or-no questions on random inputs unless the problem contains a specific, rare structure.

David Turturean, a member of the technical staff at Physical Superintelligence Inc., led the development of the research system. The resulting proof was verified by the Lean proof assistant, a software tool designed to check the correctness of mathematical proofs. Jim Halverson, an advisor to the startup, audited the work for accuracy.

Physical Superintelligence Inc. announced the breakthrough on Nov. 6, 2026, stating that the result resolves one of the remaining mysteries in quantum computing. While the proof has been accepted by the Lean system, it has not yet undergone full peer review. Researchers from Stanford University, including Guy Blanc, Jordan Docter, Carmen Strassle, and Li-Yang Tan, are currently reviewing the proof and preparing a summary for the expert community.