Vibe-coded ext4 for OpenBSD
21 points by atmosx
21 points by atmosx
I feel this article's title is not doing its contents justice - I was rather surprised vs. what I expected; IMO it'd be better if they rather titled it:
"No vibe-coded ext4 for OpenBSD"
Project was discussed here: https://lobste.rs/s/ggify2/openbsd_ext4fs_update#comments_story_01kketbbq7f11rn48j3tbnr771
Dev’s insistence on top-posting in the mailing list probably didn’t help.
Dev’s insistence on top-posting in the mailing list probably didn’t help.
It betrays the complete lack of understanding they have of the OpenBSD community. I even warned them about the licensing problems in the thread you linked.
Even if they rewrite it from scratch themselves, nobody on the OpenBSD lists are going to trust them ever again. Nobody with a commit bit is going to ok any patches they post to tech@. The fact that they think that forking OpenBSD is viable is laughable and naïve.
The fact that they think that forking OpenBSD is viable is laughable and naïve.
My reading of that is not that he is saying that forking OpenBSD is viable, but that "forking OpenBSD" would be easier than convincing them. It was an exaggeration from a frustrated person.
(Not taking sides here, I don't think LLM-generated code has any place currently in operating systems however)
Your reading is almost definitely the right one. I hadn't read the tech@ thread when I commented.
But I can't understand why the author of this patch set is even remotely surprised about this. They were warned.