DHH runs 16 agents but still reads every line of code
David Heinemeier Hansson used to be a vocal skeptic of AI-assisted development. That has changed. He now runs sixteen agents simultaneously. But his position on where the line sits has not moved at all.
The Distinction That Matters
DHH draws a clean line between two modes of working with AI code generation. One mode: you use the agent as a kind of compiler you still review. You take the leverage, but you keep the veto. You look at the implementation.
The other mode: you accept whatever the agent produces without reading it. According to DHH, that second mode is not programming. It is hoping.
The framing comes from a clip via the Lex Fridman Podcast (episode 501), analyzed by FourWeekMBA. The fact that it comes from a convert rather than a skeptic is what makes it worth taking seriously.
Why This Matters for Operators
If you are shipping products with AI-generated code, the question is not whether to use the tools. DHH clearly does. The question is whether you are in the loop when it counts. Skipping the review step is not a productivity gain if it means shipping code you cannot reason about when something breaks.
The leverage is real. The veto still has to be yours.
I’m founder of BizStack at Cagri Sarigoz LLC and a passionate advocate for entrepreneurs.
With over 14 years in tech, marketing, and AI, including my role as Head of SEO at CitizenShipper and co-founder of TaleBot at Intale AI, I’m dedicated to sharing genuine, useful product insights and tips.
At BizStack, I aim to cut through the digital noise to provide clear, actionable advice.
And more than all else, I’m a father to a (always) little girl and a husband.
Contact me at [email protected] for assistance.
