Gary Marcus扒出OpenAI的Astra证明可能有问题,一个数学博士生的反例讨论才十几个互动,和数千万播放反差太大,建议去看看。
Gary Marcus在X上批评,OpenAI新证明引发的欢呼有数千万观看,而一位数学博士研究生对疑似反例的讨论互动数不足12次。他引用自己的帖子指出,至少一个Astra证明可能出错,该帖子有3249次观看,但只有5条回复、3次转发和14个赞。Marcus用这个对比说明,证明的验证过程未得到足够关注。
jubilation about OpenAI’s new proofs: tens of millions of views consideration of a putative dispro...
jubilation about OpenAI’s new proofs: tens of millions of views consideration of a putative disproof of one of the putative counter-examples from a PhD student in math? less than a dozen 🙄 Gary Marcus @GaryMarcus Uh oh! At least one of the putative Astra proofs might turn out to be wrong: 🔗 View Quoted Tweet 💬 5 🔄 3 ❤️ 14 👀 3249 📊 5 ⚡