The real difficult in arranging agents is not to run them, but to identify the state change - to judge whether an agent stops to wait for you, or is stuck, or finished, and whether the two should be handled automatically or someone should be found. You have made this judgment for 12 agents, and you need to know how unreliable it is.
zhoujinliang · · focus · HN ↗