
这期深读一篇关于 AI 递归自我改进的最新综述:AI 已经能参与修改答案、脚手架、训练数据和研究流程,但真正卡住它的不是会不会循环,而是谁来验证它确实变好了。文章同时用 harness engineering、GhostApproval 和 Friendly Fire 三个信号,解释为什么下一阶段的人机边界会落在验证器、权限和否决权上。
lilianweng.github.io
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.