What was checked
The Riemann Hypothesis proof by Anthropic shows AI can now do 'inhuman' math tours no human brain could create alone and generalizes to making AI better at making AI. This is like a company producing enriched uranium - warning sign for extinction-level AI soon
Assessment summary
InaccurateA mathematician at Anthropic claimed to have discovered a proof with Claude, but reporting states that the claim was not independently verified and had not undergone outside expert review. The claim about whether the mathematical work is 'inhuman' or impossible for humans alone is subjective and not empirically verifiable. The proof involved collaboration between Claude and a human mathematician, with human review and formalization. Whether this represents something 'no human brain could create alone' is a philosophical claim without clear evidence.