Not by market share on desktops in particular Western countries, no.

In the world - yes, I'd expect there to be more Linux desktops really existing than MacOS desktops.

Apple is the West in a nutshell, though. When a real need or a problem is mentioned, it's mentioned only rarely as some selling point - like Jobs saying MacOSX is "very Linux-like" (sic). When a solution reaches reality, it's done in a neutered way - like Apple still regularly releasing sources. When a real need is fulfilled, you can expect twice the effort to have been covertly spent on negating its effect - like with Apple's "just works" and sterility and even privacy selling points, which come with walled gardens, closed APIs, short and bad support, and them actually still spying on you. A declaration of effort to tackle any problem existing in the wild is always aimed at portrayal of said effort to be persuasive enough to sell the product and the more efficient, the better, where efficiency means that actually doing anything other than persuasion is against the goal.

See, the difference between market mechanisms and geopolitics is not in people deciding that there'll be this set of rules here and that set of rules there. It's due to structure. The bigger that structure makes companies, the more similar market dynamic and scientific&technical progress will be to geopolitics.

Which is why I genuinely don't understand people who frown at "techno-luddites" like me dreaming of going back to decentralized, even if with worse compatibility, hardware production and simpler software and formats. I would even say that some degree of protectionism from less developed countries would help, where hardware can be imported only if there's domestic alternative production to match.

I'm more on the libertarian side, but if in XIX century weapons ownership was widespread, and in XX century it stopped being such, because of one's ability to kill much more people quickly, then the same with electronics would make sense by the same principle. Only the qualitative difference in power comes not with the device, but with the centralized production line it comes from.

Of course, that's not my proposition. It wouldn't work.

My proposition would be to try to design some "civilization minimized", where sufficiently usable computers can be produced in an area of 100k people, with society, industry, communications and warfare approaches optimized for that limitation. That kind of limitation for a self-contained unit of civilization, so to say.

So - one can buy a reasonably good laptop for 500$ today.

What kind of laptop can one produce for that sum in a 100k people settlement? Is that even possible to make it close to 500$ in relative value?

I'd expect we'd need to solve all heavy problems with specialized boards. I think lithography is actually applicable here - we don't need it very precise, it's far above our possibilities in such a hypothetical unit, and it would improve resource usage efficiency. And I think a more compact machine of PDP-11 level is possible, with specialized boards in Amiga ideology. I also think we can even have some kind of machine learning, but with analog storage of coefficients to reach anything close to sufficient efficiency. It would be a Fallout-kind computer, nothing fancy, but - possibly usable. For dedicated boards (some of which can, again, solve problems with analog approaches) there is a possibility of optronic elements being more accessible and energy-efficient.

There's a question of portability - one can expect lithium to just not be realistically available in any random area of the world populated by 100k people. I mean, trying to imagine some autarky ...

So maybe not a laptop, LOL.

Nah. That's bullshit. That won't work. At least not every 100k people. Maybe every 10mln.