Re: QEMU accepting AI code - viable alternatives for NVMM (bhyve/vmm)?

Mario Marietto <[email protected]> writes:
> but it's also true that it's not fair to make people wait decades to
> see your amazing features.

Nobody speaks about decades, but about weeks or months at least —
necessary to understand some problem's field or to make some necessary
work by the programmer(s). If the mentioned feature is really so
amazing, then the people definitely will wait — as I speak before, there
are no customers, shareholders, managers, etc to please them with the
speed of delivering new meaningless features.

E.g. there were some news about adding NetBSD-native jails instead of
existing sandboxctl solution. While it is definitely good (and
definitely not meaningless!) feature, I (and a lot of NetBSD users, I
think) still able to operate with our NetBSD servers and machines while
waiting for it.

And this was (and still is) normal mode of operation in OpenSource,
before it become invaded by corporations and people, thinking about
"profit" and "impact to community" more than about joy and fun.


> Technological progress continues even if you're a clean code purist.

This is nothing about code purism, it is just a common sense. If I'm
using a machine, perfectly munching the numbers, then I want from it to
perform these actions in a deterministic way.

What happens, when people select to write code in the non-deterministic
way, with the use of LLMs — we already saw with rsync[1]. Can't say that
this result was surprising on unpredictable.

[1] https://mastodon.gamedev.place/@JeremiahFieldhaven/116654345332213390

> And if you don't chase it (at least a little), your system will become
> increasingly niche.

Hmm, I checked "Cc:" field and I'm still writing to the NetBSD-users
list, strange!

The main selling point of NetBSD for some (exact value unknown for me)
part of it's users — is the niche-ism of this OS. If I wanted to use
some popular and widely adopted OS, then I'll install Windows.

-- 
Eugene Andrienko