The article explains how recent advancements in artificial intelligence models from OpenAI and Anthropic have solved long-standing mathematical problems, including invalidating a conjecture by Paul Erdős and offering new perspectives on the Riemann hypothesis. (translated)







