标签: Anthropic

How a Claude model made a math breakthrough during its unsuccessful 54-hour attempt to solve the Riemann hypothesis after a user repeatedly encouraged it (Ben Cohen/Wall Street Journal)

How a Claude model made a math breakthrough during its unsuccessful 54-hour attempt to solve the Riemann hypothesis after a user repeatedly encouraged it (Ben Cohen/Wall Street Journal)

一个 Claude 模型在用户长达 54 小时的反复鼓励下尝试攻克黎曼猜想,虽然最终没有成功,却在推理过程中意外做出了数学上的实质性突破。这件事值得关注,因为它展示了 AI 推理中“过程”与“结果”的复杂关系,也让人重新审视提示交互对模型能力的实际影响。