- A study on AI coding agents in GitHub repositories found no significant decline in newcomer participation, contrary to fears that AI would push beginners out.
- While code complexity increased (e.g., cognitive complexity rose about 11% in Python projects), it did not deter new contributors, and the active contributor base grew.
- The research had limitations, including focusing on established projects with pre-AI history and measuring adoption via configuration files rather than actual tool usage.
- Despite AI-generated code being more complex, newcomer retention held steady, and the main concern shifts to the growing maintenance burden on existing maintainers.
- The study suggests future work should explore AI reliance levels and projects 'born with AI,' emphasizing the impact on maintenance efforts over contributor numbers.