In 2025, a research nonprofit called METR ran a careful experiment. They took 16 experienced open-source developers, gave them 246 real tasks on codebases they'd worked in for years, and randomly let them use AI tools on some tasks and not others. Then they timed everything.
Source: [Dev.to](https://dev.to/nazar_boyko/how-to-measure-if-ai-agents-actually-improve-developer-productivity-1hpp)