The system I use has names for each task which helps keep track of them, and usually for larger tasks they’re based on a GitHub issue. If I’m ever confused about what an agent is doing, I just ask it what it’s doing. If you want this more often, you could add custom instructions telling the agent to always summarize the current state (in whatever way you find most useful) at the end of a turn.
How do people manage multi-task with multiple AI instances like Claude Code / ChatGPT?
Top problems I face:
I forget the fine-grained context so I have to spend time reading the history rather than having summarized note
Forget about a task that I left an agent working on
I get mental fatigue faster than just working on a single task
Which ends up in:
Tasks remain partially complete so no task gets actually complete
Mentally exhausted and just want to curl up on a grassy meadow staring tat the sky
Would be happy for any suggestions
The system I use has names for each task which helps keep track of them, and usually for larger tasks they’re based on a GitHub issue. If I’m ever confused about what an agent is doing, I just ask it what it’s doing. If you want this more often, you could add custom instructions telling the agent to always summarize the current state (in whatever way you find most useful) at the end of a turn.
Do you keep track of current tasks in a notes and then keep revisiting them?
I was working on an app that maintains a round-robin queue of tasks so that I can switch around them.
My UI lists task by most recently interacted (and has an explicit archive button). I think Claude Code for the web has a similar UI.
I only keep one concurrent big task running. Other tasks are small and usually just end up in a PR that I review later.