codingjoe  Mar 5

Q to my fellow #Django maintainers:

How do you deal with AI contributions? I believe I can fairly easily detect slop. However, it's usually dangerously close to code from an inexperienced engineer. How do I make sure not to insinuate and crush young engineers? I want to maintain the wonderful growth culture of the Django community while being mindful about my own capacity.

Show threadSarah BoyceMar 5
@codingjoe I don't have an answer but I'm also concerned about this
Show threadcodingjoe  Mar 5
@sarahboyce yes, it's making an already hard job even harder. I have been using GitHub Copilot prompts to get better assisted reviews. However, copilot won't address the elephant in the room by itself.
Besides, how wild is it that you can turn that darn thing off…
Show threadSarah Boyce
@codingjoe I was thinking about this today. With projects that get a lot of PRs, maybe it's worth prioritizing PRs where the contributor has expressed self-doubt, invited input on a specific part, or indicated that they tried several solutions and have landed on this one. A PR review is a mini collaboration and it's more enjoyable to collaborate with people who present themselves as "human" 🤔
Mar 9 at 5:29pmMastodon for Android
Show threadcodingjoe  Mar 9
@sarahboyce hm... I guess that makes a lot of sense for large projects. Since I mostly do medium projects, it's less so about priorities but communication. The wild mix of English being my second language, non-native contributors, AI and ADHD is a lot to navigate. I know people universally mean well, but all cultures have different ways to express (or not express) this 😅 Some might see self-doubt as a weakness, while I consider it a strength.
Show threadcodingjoe  Mar 9
@sarahboyce but you seem to navigate this fairly well, at least from my own experience. Should you ever find the secret ingredient, do share 😊