[AI] Could FSF use AI to complete and develop the GNU kernel, Hurd/Mach?
- Vous devez vous identifier ou créer un compte pour écrire des commentaires
I just though of that.
No.
And Hurd's already completed anyway. Go download & use it already - https://guix.gnu.org/en/download/latest/
That's a virtual machine image, not a ISO image for running the Hurd on a computer from the USB stick. That isn't being "complete" in the completest sense.
Go run it on real hardware then. For example: https://guix.gnu.org/en/blog/2024/hurd-on-thinkpad/ Since even then, Hurd supports x86-64. The Guix installer lets you select Hurd.
https://guix.gnu.org/en/blog/2026/the-64-bit-hurd/ They have it marked as experimental (in Guix). It's time to stop treating Hurd as if it's some mythical vaporware. :) Go use if you're interested in doing so - It has some neat features. The status section there has interesting data and links. Like the FOSDEM talk. And "Debian GNU/Hurd has been a reality for some years now, reaching 75% of Debian packages being available for the Hurd." So most of Debian main runs on Hurd.
The GNU Hurd is ready, but it needs drivers to become compatible with a wider range of devices than it currently does support. NetBSD's Rumpkernel helps solve this issue. There was a report of a contributor using claude code, so it is not like they haven't thought up with this idea. The problem is the sustainability of AI code. I think there is too little manpower to handle all the sloppy code.