Post

Jordan McQueen
@jm.dev

I've been a Linux user for a long time, and have always had a lot of confidence in the longevity of the project, but lately my mindset has shifted -- I asses the probability that it gets out-competed over the next decade as considerably higher than I did 5 years ago.

0 reposts3 likes

LeChuck
@beneath-a.bsky.social

Why do you think it will get outcompeted? In terms of Desktop users, L. may be negligible either way. For embedded stuff L. has very good hardware support, tooling and a regiment of experienced developers. Outcompeting Linux in the embedded realm will take years of effort and persuasion.

d3Xt3r
@d3xt3r.bsky.social

On the contrary, I think an alternative OS could do just as well, if not better, in the embedded realm, & that's because embedded systems depend heavily on proprietary firmware/drivers - a vast majority of which aren't upstream. So to embedded developers, the Linux kernel doesn't offer them much...

d3Xt3r
@d3xt3r.bsky.social

... of an advantage, since they have to develop/source all the drivers either way. Unless they're working on an off-the-shelf board like an RPi, which most commercial entities will not. Yes, the tooling and APIs may be different, but it's not really a huge hurdle if the competitor is Unix-like.

d3Xt3r
@d3xt3r.bsky.social

Just look at how well QNX is doing in the manufacturing and automobile sectors for instance. BSD is also doing well in the NAS, firewall, and network appliance space. So the hardware support advantage that Linux has is mostly only relevant to desktop users.

Jordan McQueen
@jm.dev

I don’t necessarily think it will. I’m just saying that the probability value I ascribe to that outcome is higher now that I would’ve said 5 years ago.

Jordan McQueen
@jm.dev

And this isn’t just because of the current RFL conflict, but that is a part of it. I believe there are multiple credible paths — other old UNIXes, new OS projects, and I believe in the future of the OS-as-a-library pattern.

Dorin Lazăr 🇷🇴🇪🇺
@dorinlazar.ro

There's less of an effervescence around OS designs than it was in the early '00s. People in decisive decissions were very reluctant to move from things like Windows CE and RTOS to Linux for a long long time. I doubt that in the LLM era there will be a huge desire to switch techs - who will teach it?

Dorin Lazăr 🇷🇴🇪🇺
@dorinlazar.ro

POSIX is a huge problem. we're somehow stuck in 1980 with the OS interface design and X11 was for a long time a huge blocker in developing user interfaces. Now, I'm afraid, there's fewer engineers skilled enough to define replacements. The language is the most unimportant aspect of the issue.