文章摘要
Neel Somani宣布仅使用GPT 5.2 Pro解决了第二个Erdos问题(#281),此前无解。陶哲轩称这"或许是AI解决开放问题最明确的例子"。
文章总结
尼尔·索马尼在X平台上宣布,他仅使用GPT 5.2 Pro成功解决了第二个埃尔德什问题(编号281),且该问题此前未有公开解决方案。著名数学家陶哲轩对此评价称,这可能是"人工智能解决开放性问题中最明确的案例之一"。
(注:根据编辑规范,已删除以下无关内容: 1. 社交媒体平台推广信息 2. 图片描述及嵌入代码 3. 趋势话题等无关板块 4. 重复的URL及发布时间信息 保留了核心学术突破的陈述及关键人物评价)
评论总结
总结评论内容:
- 对LLM解决数学问题的认可
- 陶哲轩高度评价LLM在证明过程中避免错误的能力:"Very nice! ... actually the thing that impresses me more than the proof method is the avoidance of errors" (评论3)
- 有用户通过实验验证多个LLM都能解决该问题:"I just gave Deepseek the same prompt and it solved it too" (评论8)
- 对LLM能力的质疑
- 担心LLM可能给出看似正确实则错误的解:"I've 'solved' many math problems with LLMs, with LLMs giving full confidence in subtly or significantly incorrect solutions" (评论2)
- 质疑解决的问题是否属于"低垂果实":"if these problems solved by LLMs are just low-hanging fruit" (评论9)
- 关于埃尔德什问题的讨论
- 陶哲轩指出埃尔德什问题难度差异很大:"Erdős problems vary widely in difficulty" (评论10)
- 有人询问为何选择埃尔德什问题:"why has the LLM math solving community been focused on the Erdos problems" (评论6)
- 关于AI发展的观点
- 乐观预测AI将超越人类:"I truly believe that ai will far surpass human beings at coding" (评论5)
- 认为AI已展现某种智能:"it's undeniable at this point that they have some form of intelligence" (评论7)
- 解决方案状态的更新
- 指出该问题已有传统解法:"a prior solution has just been found" (评论11)
- 解决方案状态可能有变动:"I'm assuming something was retracted" (评论12)