A company called Physical Superintelligence, or PSI, says an unreleased model from the AI firm Anthropic has proved the Aaronson-Ambainis conjecture, a long-standing question in quantum complexity theory. If confirmed, it would be an example of artificial intelligence contributing to a mathematical problem that researchers have been trying to settle for years.
The claim was first reported by Crypto Briefing, which covered the announcement in a short piece that did not include the underlying evidence. No proof details, paper or peer review were made public, so readers and experts have no way yet to check whether the argument holds up.
That gap is the central issue. Because the model has not been released, outside mathematicians and computer scientists cannot examine it or test its reasoning. Conjectures in this field are typically judged only after careful scrutiny by specialists, and a result of this kind would normally require a written proof that others can review.
The episode also raises broader questions about transparency. Claims that AI systems have advanced scientific knowledge carry weight, and the credibility of such announcements depends on whether the companies involved are willing to share methods and results openly. For now, the accuracy of PSI's statement remains unconfirmed, and neither PSI nor Anthropic has supplied the verification that would settle it.
Credit. Based on reporting by Diego Almada Lopez for Crypto Briefing. Is this your work? Write to us.
19
