Aug 24, 2026 · 44 min
Ryan Carson turns autonomous agents into a management system
I spent $20,000 on Devin in a month. Here’s what I learned | Ryan Carson (solo founder)
The episode examines how solo founders can use coding agents at scale without confusing more software output with better product decisions.
- 1Managing autonomous agents requires visible priorities, structured delegation, and workflows that prevent decision fatigue.
- 2Customer conversations and product judgment remain human responsibilities because coding agents cannot reliably decide what to build.
- 3Automated reviews, design systems, and executive assistants expand leverage while human oversight remains essential for higher-risk work.
Don't miss
Carson explains how his Watchdog playbook monitors customer cases, product issues, and operational work while he coordinates engineering, sales, and customer success.
The brief
Ryan Carson describes spending $20,000 on Devin while building Untangle, a family-law SaaS company, and argues that the emerging founder skill is managing many autonomous workers at once.
His system organizes agent work into folders, priorities, and threads, while a Watchdog playbook monitors customer cases, product issues, and operations across a solo-founder workload.
Claire Vo and Carson draw a boundary around automation: agents can generate and verify code, but product-market fit still depends on talking with customers and choosing what matters.
Their workflows divide labor among Devin, Codex, Claude, and OpenClaw for engineering, design, email, and executive work, with automated pull-request review escalating higher-risk changes to humans.
The episode’s central lesson is a constraint on AI enthusiasm: more generated code does not guarantee a better product, and visible priorities matter more than maximizing output.
Featuring
Devin
Claude
OpenClawListen to the full episode and explore every guest, topic, and moment on PodLume.