AI 导读
在迈向 RSI（递归自我改进）的 L4 阶段，AI 从部署后发生的事情中学习，并将有用的经验转化为对其记忆、工具或代码的持久改变。 这样，智能体就能持续从真实世界经验中适应，而无需等待人类手动更新它。
正文
At L4 towards RSI (Recursive Self-Improvement ), the AI learns from what happens after deployment and turns useful lessons into lasting changes to its memory, tools, or code.
So the agent can keep adapting from real-world experience instead of waiting for humans to manually update it.
来源：@rohanpaul_ai · x.com