To the surprise of nobody, code assistants continued their surge in use during 2024. According to Stack Overflow’s 2024 Developer Survey, 63% of Professional Developers said they currently use AI in their development process, with another 14% saying they plan to soon [2].
The 36,894 developers who answered the “most important benefit you’re hoping to achieve with AI” overwhelmingly picked “Increased productivity” [2]. Developers view AI as a means to write more code, faster. Through the lens of “does more code get written?” common sense and existing research agree that yes–AI assistants beget code lines written.
But if you ask a Senior Developer “what would let your team be more productive?” their answer won’t be “more code lines added.” To retain high project velocity over years, research suggests that a DRY (Don’t Repeat Yourself), modular approach to building is required [4] [5]. Canonical systems are documented, well-tested, reused, and periodically upgraded.
The aim of this research is to measure how today’s profusion of LLM-generated code influences the maintainability of tomorrow’s systems.
To investigate, GitClear analyzed 211 million changed lines of code from non-autogenerated files, authored between January 2020 and December 2024 [A1]. This is the largest known database of highly structured code change data that has been brought to evaluate code quality [A2].
We find multiple signals of accelerating degradation in code quality. In particular, the frequency of copy/pasted lines increased 6% faster than our prediction. Perhaps worse, the frequency of commits with large duplicated blocks grew with even greater velocity.
We conclude with suggestions for managers and developers seeking to maintain high code quality in spite of the countervailing forces of 2025.
If you would like to discuss the latest findings on a podcast or Youtube channel, please contact bill@gitclear.com. Last year's research caught a lot of attention. I'd love to geek out with people who care about the implications of this phase change in "what it means to be a programmer."
