Does it, really? A recent study showed that, for experienced open-source developers, genAI tools made them less effective, not more. The study, Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity, was conducted by the nonprofit AI research group METR. It involved 16 experienced open-source developers, averaging more than 10 years of experience. In short, these were experts set to working on genuine issues like bug fixes and features in projects they knew well in familiar code repositories.
Using such genAI tools as Cursor Pro and Claude 3.5/3.7 Sonnet, the programmers presumed that with AI’s help, they’d reduce their task time by 24%. Even after finishing the study, they believed it had sped them up by 20%. They were wrong, badly wrong.
In reality, it took them 19% longer to complete tasks compared to their compadres who hadn’t used genAI at all. Tasks expected to take roughly six hours or less showed the most pronounced slowdown, with AI-assisted work slowing completion times significantly.
Read the full article here
