Roughly one year ago I predicted wide-spread degradation in software quality in the near future, mostly thanks to over-zealous application of generative AI in the development process.
Looks like we are in the middle of it right now.
I'm seeing strange bugs manifest in ways that I've never seen before: wireguard module causing access violation randomly in linux 7.0 when compiled with clang 22, chromium refusing to show context menu when it doesn't feel like doing it, etc. Obviously this is nothing but anecdotal evidence, and I don't even know whether these problems are indeed caused by AI-generated code. But I can vividly feel the reliability of certain tools I deal with on a daily basis slipping away.
You may say this is nothing but my subjective feelings, or even, "my vibes". You are correct, and that's exactly my point. I'm wondering if there's a significant portion of the population feeling the same vibes of software quality degradation, and when that happens, whether we would finally reconsider some of the choices we made as a community.
Some folks are forking open source projects that adopted permissive AI policies. I don't know if these forks will last, or would they mirror the ffmpeg / libav situation. But one thing is clear: some people are not entertained by the idea of AI-generated code powering our entire infrastructure, and rightfully so.
Are you feeling the degradation?