Google and FOSS may be closer than you think...
I'm not entirely sure entirely I buy into the premise of this article, at least when it comes to comparing Google and FOSS/Linux.
(Can't really dispute the assessment of what's going on with Meta, though. I just wish I had some way to tap into the money tap; maybe I should create a startup which uses AI to generate avatar sock-textures, or somesuch...)
To me, Google's approach is essentially a commercialised take on the FOSS development process: take some smart people, give them a big budget and then see what happens.
(By all accounts, Valve's business model is even more anarchistic, which may help to explain why so little actually comes out from Valve themselves; in the last 5 years, they've released a total of 4 games, of which one was a tech demo, another was a commercialised mod and a third collapsed and died due to being a bit too blatantly pay-to-win...)
I'm also not sure that I agree with this statement:
> [FOSS developers] understand the importance of making things that matter and are very motivated to be a part of that. These resources are limited. Time spent developing, maintaining and testing for a processor that nobody uses is time not spent on current needs
There's certainly developers who do think like that, and Linus Torvalds is a pretty good example.
But I'd also note that most "flagship" FOSS projects are either sponsored by companies such as Google or Microsoft, or have decided to offer services around their technologies. E.g. Elastic, Libreoffice, Mozilla and Redhat. Which means that not only is there's some degree of commercial awareness in the actions these projects take, but a significant number of their developers are paid to work on those projects, rather than doing it just because they're "very motivated to be a part of it".
I'd also question how good it is to have someone working on code for ideological reasons. Because not only can ideologies can clash, but people's ideologies and motivations often change over time. And what happens if you can't find someone who shares that same set of motivations?
E.g. NTP was created by David L Mills, who's sadly now unable to significantly contribute to it any more due to visual disabilities. His code is described as both ideosyncratic and complex, but he agreed to let someone called Harlan Stenn take over the reference implementation
However, Stenn himself is now in his late 60s and has apparently burnt through his retirement funds supporting NTP, though he's also set up a foundation through which donations can be given.
And in the meantime, various issues with NTP has led to other people creating their own forks and implementations - NTPSec, Chrony, even Google with their "leapsmear" approach to time sync.
Which leads to the final point: there's also plenty of waste and duplicated effort in FOSS. I mean: Mysql and MariaDB, anyone? Or how about Openoffice or Libreoffice? Plex and Kodi (nee XBMC)? Or look at what happened to Perl, when Larry Wall decided it was better to start again with Perl 6 (aka Raku), rather than evolving Perl 5?
And as to abandoned FOSS projects? Where do you start?
FOSS has many useful attributes, but flagship projects such as Linux are very much an exception to the rule.