The show talks about Fedora’s AI Policy, a few thoughts:
- Fedora much like many FOSS projects have to deal with AI slop and bots spamming their repos with pull requests and whatever else. Daniel Stenburg (the guy behind cURL project) has out right banned and blocked any pull requests or anything that is suspected to be AI generated, as he has wasted his time chasing non-existing issues.
- I think using Ai is okay, it is a tool and if this tool helps bring people into FOSS then that is a net positive. However there must be a declaration that AI was used and specifically where in the code. This will help address any issues while parsing who is using AI in earnest ways to contribute to a project versus AI being used for malicious purposes.
- Since AI is a predictive non-deterministic tool, any user who contributes code that has AI should understand that the code can and likely has errors, therefore should be labelled or tagged as “suggested”.
- Many FOSS projects use Bugzilla to help squash code bugs, relying on AI to deal with these software issues is short sighted and adds to further unpaid work burnout by the project contributors.
Fedora needs to stand tough on this issue, to draw the line in the sand as it were, so that everyone understands using AI is not inherently bad but is not the responsibility free coder it is hyped to be when real people work real hard on software we all rely on.