I’m curious whether there’s a meaningful way to quantify how much AI tools (ChatGPT, Claude, Copilot, etc.) are affecting software development productivity worldwide.

Obviously, metrics like “number of repos created,” “number of commits,” or “lines of code changed” are not quality indicators — but could they still reveal directional trends? For example:

Are new repos being created at a faster rate than previous years?

Are commits per developer increasing?

Is the total volume of code changes growing noticeably on platforms like GitHub or GitLab?

Has anyone seen research, datasets, or attempts to measure this kind of global shift? Are there better proxies for AI-driven productivity gains (or losses) in software engineering?