20 June 20264 min read
AI ships code by copy-paste. GitClear measured the bill.
GitClear analyzed 211 million lines of code and found that as AI tools spread, copy-pasted code overtook refactored code for the first time. The speed is real. So is the maintenance debt it quietly defers, and that bill comes due on your team, later.
Short version: AI feels fast because it produces a lot of working-looking code quickly. The question nobody asks in the demo is what kind of code. GitClear went and counted, across 211 million lines, and the answer is uncomfortable: as AI spread, teams stopped refactoring and started pasting. For the first time on record, copy-pasted code overtook refactored code. The output went up. The codebase got harder to live in.
What GitClear measured#
GitClear analyzed 211 million changed lines of code across five years of real repositories. The trend lines that matter:
- Copy-pasted lines rose from 8.3% of changes (2020) to 12.3% (2024) — a roughly 48% relative jump.
- Refactored ("moved") lines fell from 24.1% to 9.5%. In 2024, copy-pasted code exceeded refactored code for the first time they've recorded.
- Code clones grew roughly fourfold, and the share of code revised within two weeks of its commit climbed from 3.1% to 5.7%, a sign of more churn and rework.