@gbargoud I actually ran Hurd on real hardware way back when, was even my main development box for a bit1.

While I like a micro kernel architecture a whole lot more than monolithic kernels, and there are a lot of great stuff in the Hurd (translators, the concept of no UIDs, subhurds, etc), I don't think it will ever grow out of toy status.

That, and it's a GNU project. I have exactly zero trust in GNU, sadly.

I have my eyes on Redox OS, which - unlike GNU and the Hurd - has a strong anti-AI policy, and is a more capable, more modern microkernel OS. I think it has a far higher chance of becoming daily-driver material than the Hurd. It accomplished a lot more in a lot less time.

1. Evidence! ↩︎

@wyatt Yes :)

@StarkRG FreeBSD and OpenBSD have permissive policies, and tainted commits in core. NetBSD has a reasonable AI policy: it permits tainted code if core gives permission - they never did, and are unlikely to do so, based on official communication.

There are other smaller BSDs that have stronger anti-AI policies, but the major ones - except NetBSD - aren't suitable alternatives imo.

(Evidence links here)