“[I]f a no LLM policy forces contributors to try to hide their LLM contribution by: ⁃ Deeply understanding the code ⁃ Outputting changes at a pace that is digestible to the project ⁃ Entering into discussions with other contributors ⁃ Actively engaging in code review ⁃ Learning about the project and growing their understanding

Then that sounds like a great policy to me.”

https://www.colincornaby.me/2026/10/open-source-is-social-and-llm-generated-prs-are-antisocial/

I often try to think of similar tactics to get people to the right thing - or at least the thing I want. I believe that showing the limits, costs and problems that AI causes is more persuasive than arguing against it. (And who knows, you might learn a thing or two while engaging with LLMs!) NB: At the moment I have a “no LLMs for personal use” policy in place. What convinced me to do that (at least for now) was a mix of reading what critics and boosters had to say, a lot of personal reflection and intense personal and professional use of the tech for real world problems. I don't think I would have landed here without all of these pieces.

• ← Previous
  Seems like the best™...
• Next →
  Not sure if this'll...