Why "lines of code" is a broken metric
19 Oct 2025 · 1 min

0:33
A spoken monologue (YouTube) arguing that using "lines of code" as a productivity metric is flawed. The speaker explains that with LLMs and agents, the metric can be gamed (e.g., prompting generation of long, commented code), which encourages verbosity, increases complexity, and creates technical debt, making the metric misleading and harmful.










