Hacker News Reader

Are AIs good enough at doing novel work that they have yet to encounter?

My guess is that the GNU/HURD people are doing a lot of novel work that doesn't yet exist.

My opinion is Sol or Fable could absolutely "finish" Hurd (whatever the definition of "finished" is), but it would be difficult to review and understand everything it did, nobody would trust it, people would raise copyright concerns, doing so would violate all sorts of GNU cultural norms, and it would sit in a repo unused.

Thus my "blasphemous" comment.

How novel is it really?

You think what they're doing isn't covered in the intersection of OS books, OS papers, other OSs, osdev, random GitHub projects?

AI to some extent learns APIs. If they are doing new things they are probably introducing new kinds of interfaces.

Reply on news.ycombinator.com