周彦充Ai私教陪跑落地

The moment @claudeai Fable 5.2 was released, Google panicked and rushed out its previously announced (and now behind) Carbon 🤣 Keep shipping late 👏 @OfficialLoganK

朋友们，Anthropic 这次发布Fable 5.2，自己在公告里承认，训练的时候故意把模型的两项能力给压制了。 一个是给芯片写底层程序，让芯片训练模型跑得更快；另一个是改进大模型的架构。 他们给出的理由是，这两样最能直接加速 AI 研发。 说白了，就是怕 AI 太会造 AI，自己给自己升级，人类就跟不上了。 这个我之前也写过，AI 2027的作者也不建议太快实现RSI，否则会造成失控。 而上周 METR 刚测出来，其实Anthropic 的进步速度已经翻了倍，大约八成的提速要算在 AI 自己头上。 所以Anthropic现在这意思是，我们已经在手动放缓模型的RSI进程了，求求各位，别追着我们杀??? 可他们前几天不还在招股书里说，未来模型可能很快实现RSI，这又开始左右脑互搏了，是被谁压力了???@grok

We have dumped JAX/SGLang and rewritten the entire RL and inference stack from scratch in C/C++, which took some extra time. Training (aka pre-training) is done. Lot of cleanup was needed on the SFT data to get that mostly right. Now, 4.8 is in RL and the evals are starting to climb. We will do a soft rollout to Bot users later this week and see how it goes. Grok 4.8 will bear no resemblance to 4.7 (thankfully). It is not yet incredible, but, this time, the foundations are built right, so I expect it to evolve to be something special.

继续删内容，抖音删完了，现在删X，最后是小红书。只留爆款。这样可以提升网感，虽然比较缓慢，但是我心里觉得是有用的，因为大多数内容确实是垃圾内容。。。

Claude Fable 5.2 is now available in Cursor! At max effort, it's the top model on CursorBench at 60%. Fable 5.2 scores 56.9% on CursorBench at default effort, up from 49.2% for Fable 5.1.

今晚估计真要睡不着了！！！😱 做梦都没想到，我的帖子竟然被马斯克转发了！ 玩 X 以来最惊喜的一刻！！！ 被老马翻牌子了！😂