The REWORK Podcast — selected coding-agent conversations
Bring your AI agents to Basecamp
Seven episodes on agent-assisted development, durable software, product judgment, and agent-ready interfaces
A manually reviewed selection from The REWORK Podcast following 37signals’ move from cautious LLM assistance to terminal agents, AI-accelerated Basecamp development, agent-accessible product interfaces, and agent-built competition. Jason Fried and David Heinemeier Hansson emphasize disposable prototyping, senior review, architecture, product scope, and consequence-aware use. Their examples are first-party experience and forecasts, not controlled measurements.
In this collection Browse 7 summaries 4 of 7
REWORK host Kimberly Rhodes speaks with 37signals co-founder and CTO David Heinemeier Hansson about Basecamp's command-line interface for external agents. Hansson frames CLIs and skills as an accessibility layer that makes existing product capabilities faster and more direct for agents than visual browser control.
Key Points Covered
- Browser success was still too slow: Agents could navigate Basecamp visually, but the interaction latency motivated a direct command-line interface 02:09–05:56(opens in a new tab).
- Tools turn prose into shared work: A launch request becomes to-dos, comments, messages, assignments, and schedule items inside a Basecamp project 05:56–07:37(opens in a new tab).
- Feedback loops matter: As with coding agents, terminal tools let an agent act, inspect errors, adjust the call, and continue 08:25–09:28(opens in a new tab).
- Accessibility can precede embedded AI: Hansson recommends exposing a product through CLIs and skills rather than adding native AI features that do not yet improve the workflow 10:23–14:05(opens in a new tab).
- Agents can build their own on-ramp: He reports that agents produced much of the Basecamp CLI and says the approach will extend to HEY and Fizzy 15:04–17:02(opens in a new tab).
Full video: https://www.youtube.com/watch?v=oE1Y_6W4Mj8(opens in a new tab)