OpenAI and Anthropic nearly at the same time declared that their in-house models had debunked the Exponential Time Hypothesis (ETH), a theory that had held sway in theoretical computer science for 27 years. Remarkably, Anthropic's breakthrough was achieved autonomously by Claude, without any human input, earning Claude the distinction of being the sole author credited on the paper. The Exponential Time Hypothesis, first put forward by scholars in 1999, suggests that there is no sub-exponential time algorithm for 3SAT, a cornerstone assumption for numerous studies in related fields. Now that this hypothesis has been disproven, many research findings that relied on it will need to be re-examined. At present, both companies have made their operational Lean formal proof codes publicly available, and the original proponents of the hypothesis are working alongside top-tier teams to verify these codes. In the meantime, industry experts have distilled the lengthy AI-generated proofs into more digestible forms. This milestone indicates that AI is no longer just a tool for human scientific research; it has now become a key driver of cutting-edge theoretical exploration. Looking ahead, the scientific research paradigm may undergo a transformation, with human researchers being freed from the tedium of proof derivation and instead focusing on identifying core issues and forging interdisciplinary real-world connections.