The exchange concerns claims about the Riemann Hypothesis and the credibility of proposed proofs. One commenter says the arguments against Masullo’s proof are correct, but argues that they resemble what language models produce and that Masullo, whom they call a known crank, should not receive credit. Another says that if AI helped find counterexamples, that use should be disclosed, noting claims that people found counterexamples quickly with Astra. A further commenter says Masullo also claimed to disprove an AI-generated proof associated with Levent Alpöge.
The main disagreement is over whether the AI-generated proof deserves to be treated as established. One commenter calls it more or less accepted, citing Lean formalization and reported checks by five independent Lean kernels, as well as confidence among mathematicians and Lean experts. Others reject that characterization: they stress that the proof had only just been released, is roughly a million lines long, and that machine-checkability does not itself settle how the mathematical community regards it. One says it is viewed as possibly correct partly because of its author’s credentials; another points out that an enormous formal proof is difficult to assess in practice. A commenter also makes an extreme call for Anthropic executives to be imprisoned.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.