>You spend a few minutes defining a task. Clear intent, clear scope, clear success criteria. You hand it to the AI and it starts working. But you’re not going to sit there watching a cursor move. So you jump to the next task. Set it up, define intent, delegate. Then the next one.

I make a couple of push up, or play a couple of minutes of Heroes III while the agent is working. I see no reason to run multiple agents.