What your AI enablement strategy forgot about. I use Claude Code for daily operations: managing todos, tracking people and projects, processing meeting notes, building onboarding curricula. Here's my ~system~ for not losing my mind as a people manager.
My partner gave me his old gaming rig to turn into a home server. I dreaded dual-booting Linux again, so I SSHed Claude Code into the Windows machine and let it do the work. For intermediate hobbyists who've half-assed this stuff before and don't realize how fast they can move now.
I built a team of 11 AI agents to work on my projects autonomously. After 18 days and 1,159 commits, I pulled the plug and started over. Here's what worked, what broke, and what I learned about managing AI systems as an engineering leader.
The technical companion to my retrospective. How the system actually worked: the architecture, agent roles, NATS message bus, roadmap-driven development, and the patterns I used to coordinate 11 AI agents. DM me for the code.