PM Who Codes: Why I Stopped Waiting
My job as a PM is still the same: pick the right problems, define value clearly, and reduce risk fast. What changed is execution speed. With AI + coding, I can move from product intent to working software in hours, not sprint cycles.
Core shift: I no longer treat implementation as a black box. I treat it as a controllable system with clear interfaces and tight feedback loops.
What PM judgment still owns
- Choosing the right user and the right pain point
- Defining success metrics before building
- Cutting scope ruthlessly when signal is weak
- Knowing when quality/risk is good enough to ship
What coding unlocks
I can test ideas immediately with real interactions, not mockups. I can run production-like edge cases before asking engineering teams to commit. This improves requirements quality and reduces churn later.
My operating loop
- Write a one-page product brief
- Build the narrowest end-to-end slice
- Run with real users or realistic internal usage
- Measure and decide: cut, iterate, or scale
PM who codes is not about replacing engineers. It is about arriving with higher-quality decisions and fewer unknowns.