WIP Limits Still Matter in an AI-Powered World
My workflow makes it easy to spin up 10+ parallel agent threads, but 5 or 6 well-sized tasks is the sweet spot - for my sanity and to keep the final review pile manageable.
Artificial constraints like WIP limits are still really useful in an AI powered building world.
My workflow for established products now makes it really easy to spin up 10 or more parallel threads - but I'm finding about 5 or 6 "3 pointer" sized tasks in related areas in parallel is the sweet spot, both for my own context switching sanity and to avoid a giant pile of items to do the final pass + verification.
My robots now do the first 1-25 passes at review and verification, but I'm still accountable for what I ship.
Now I just need to build that artificial constraint into the workflow itself rather than having a rate limit enforce it for me. 😅
Related insights
Agents Are Going Multi-Player
Most people's experience with AI agents is single-player. We're moving into a multi-player era where specialized agents live in your team channels as true teammates.
Codify Senior Judgment to Scale AI Impact
Scaling impact with AI comes from codifying senior judgment into skills and agent workflows—which also becomes the training material for the people still climbing the ladder.
Codex Is Becoming My Daily Driver
After a year with Claude Code as my daily driver, the Codex desktop app's interface for managing many agent threads and automations has me reaching for it more and more.
Get More Like This
Follow along as I build and share what I learn
Found this helpful? Share it with your network!