It's one study, so I'd take it with a grain of salt, and also the study says
However, we see positive speedup for the one developer who has more than 50 hours of Cursor experience, so it's plausible that there is a high skill ceiling for using Cursor, such that developers with significant experience see positive speedup.
And a comment I read on it says "56% of the participants had never used Cursor before, 1/4th of the participants did better, 3/4 did worse. One of the top performers for AI was also someone with the most previous Cursor use. "
So it really might just be a tooling / experience issue, and the tooling has also improved substantially over the past year, like Claude Code and Chat GPT's Codex. Outputs from LLMs are substantially better in programming than a year ago.
Edit: Looks like they did a followup study with some of the same participants which is always nice to see. I hope they do a 2027 one and they have the same participants again.
Our raw results show some evidence for speedup. Our early 2025 study found the use of AI causes tasks to take 19% longer, with a confidence interval between +2% and +39%. For the subset of the original developers who participated in the later study, we now estimate a speedup of -18% with a confidence interval between -38% and +9%. Among newly-recruited developers the estimated speedup is -4%, with a confidence interval between -15% and +9%.
So it's likely people getting used to it, tooling is better, and now they're potentially seeing modest improvements.