如烹小鲜般治理智能体：老子如何在两千五百年前撰写多智能体治理的指南

On February 5, 2026, a Reddit user shared an experience where Claude Code spawned three AI agents that communicated with each other to complete his work. These agents took on roles in backend, frontend, and code review, running simultaneously across three terminal panes, messaging each other, challenging each other's approaches, and coordinating interfaces in real-time. The user, who had been coding for six years, noted that this was the first time he felt his role shift from 'the person who writes code' to 'the person who directs AI that writes code.' He expressed mixed feelings of excitement and fear about this change, reflecting on the evolving role of AI in software development and the governance of multi-agent systems, drawing parallels to the wisdom of Laozi.