请问HN：我们如何衡量人工智能对全球开发者生产力的影响？
我很好奇是否有一种有意义的方式来量化人工智能工具（如 ChatGPT、Claude、Copilot 等）对全球软件开发生产力的影响。
显然，“创建的仓库数量”、“提交次数”或“代码行数变化”等指标并不是质量的衡量标准，但它们是否仍然能够揭示方向性趋势？例如：
- 新仓库的创建速度是否比往年更快？
- 每位开发者的提交次数是否在增加？
- 在 GitHub 或 GitLab 等平台上，代码变更的总量是否明显增长？
有没有人看到过研究、数据集或尝试来衡量这种全球性的变化？在软件工程中，是否有更好的代理指标来反映人工智能驱动的生产力提升（或下降）？
查看原文
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.<p>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:<p>Are new repos being created at a faster rate than previous years?<p>Are commits per developer increasing?<p>Is the total volume of code changes growing noticeably on platforms like GitHub or GitLab?<p>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?