I wanted to ask a more specific question about the team’s stance on AI use in development and contributions.
I did see @j0rge earlier post about AI, and that answered the broader question pretty well. It seems clear that Universal Blue is supportive of AI workloads and sees them as an important part of the space.
What I am more curious about is the development and contribution side of it.
I know there are pretty strong opinions out there about AI in general, especially around development work. Personally, I take a more pragmatic approach. I see real value in where it can save time or help unblock things, and it has definitely helped me learn. At the same time, I’m also aware it can produce bad or misleading results, and there’s no substitute for actually understanding what you’re doing (and actually doing it yourself!).
So I was curious how AI is currently viewed, if at all, in the actual development and maintenance of Universal Blue. Is it being used in any meaningful way for image development, infrastructure, repo maintenance, or related work? Is there a general philosophy from the team on where it is appropriate versus not appropriate on the development side?
From a contributor perspective, I’m also interested in how AI-assisted work is viewed. If someone finds a bug, uses AI to help generate a fix, and submits a PR, is that just evaluated on its merit like any other PR as long as it’s correct and tested?
I did see references in Bluefin docs around AI attribution, so I wasn’t sure if that reflects the broader Universal Blue stance or just that specific project.
I tend to treat AI-generated code and fixes VERY carefully. If I don’t fully understand what is being done, I simply don’t submit a PR. It would be crazy embarrassing to submit a PR, even if it works, and then have somebody ask a question, only to have to say “well, IDK… didn’t write it!” lol. On my own personal repos I’ve let it go wild and it’s created some monstrously complicated workflows and logic, but on somebody else’s repos I would never submit a PR with something like that.
Mostly just trying to understand the expectations and norms here before contributing more actively.
So… before looking at my Claude and OpenAI Codex percentages left for the week and pointing it at all the Universal Blue repos and saying “look for any and all bugs, and submit PRs” … NO… I would not do that ROFL, but I hope you can get what I’m asking in a roundabout way 
I really appreciate the proofread function in Discourse! So, AI for the win there!
BTW… in case anybody else knows about this (stumbled across it the other day), this automated wiki scans repos and tries to understand them, very cool and useful (if right!). Example: ublue-os/main | DeepWiki