Lenny's Podcast: Product | Career | Growth · Lenny Rachitsky

How to measure AI developer productivity in 2025 | Nicole Forsgren

·1 hr 8 min·5 clips

Nicole Forsgren says most productivity metrics are a lie—AI can game them by generating verbose, complex code.

This episode of Lenny's Podcast features Nicole Forsgren, a leading expert on measuring developer productivity and creator of the DORA and SPACE frameworks. They explore how AI tools are changing how engineering teams work and how to accurately measure their impact. Forsgren argues that traditional productivity metrics like lines of code are now obsolete because AI can easily game them. She explains that AI accelerates coding but introduces new bottlenecks like code review and trust issues with non-deterministic outputs. A key shift is that developers now spend more time reviewing AI-generated code than writing it, which changes workflows and cognitive load. Forsgren observes that some senior engineers create advanced workflows where they assign architectural pieces to different AI agents to work in parallel. She cites research by Gloria Mark suggesting humans have about four hours of deep work daily, and AI might make shorter, 45-minute work blocks more productive by helping developers regain context. One surprising insight is that for regular users of AI coding tools, the code *they* write often increases twice as much as the code the AI generates for them. Forsgren states most teams can move faster, but speed must be paired with good strategy to avoid "shipping trash faster." Common signs a team has room for improvement include constantly breaking builds, flaky tests, and high friction when switching projects. Her simplest advice for immediate improvement is to talk to developers and listen to their daily pain points, as fixing a broken process often requires no new technology. The conversation is educational and practical, blending high-level frameworks with specific, actionable examples. Forsgren maintains a measured, research-informed perspective on AI's potential and pitfalls. Engineering managers, product leaders, and developers interested in productivity metrics will find this episode highly valuable. Listeners seeking a purely technical dive into AI coding tools or those who prefer narrative storytelling might find it less engaging.

As heard by us

A sharp reset for measuring AI-era speed without fooling yourself.

Nicole Forsgren makes the episode land by pushing past the vanity of lines-of-code metrics and forcing a better question: what should AI productivity actually measure?

Read the full review in PlayNext →

Why you'd press play

If you're trying to tell whether your team is really moving fast enough, this episode is a useful reset.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Lenny's Podcast: Product | Career | Growth