For lots of people, that's Linux. For lots of other people, that's Windows.
A similar thing happened a few years ago when Subversion was still the dominant version control system and SourceForge was still the main repository for open-source projects. It was 2008, and I was at a dinner seated between the CEO of CollabNet, principal corporate backer of Subversion, and the outgoing CEO of SourceForge. I asked the CollabNet CEO what he thought about Git. His view was that there was a winner among open-source VCSes, and that winner was Subversion. Then I asked the SourceForge CEO if he'd heard of GitHub. He had, but was unconcerned, since SourceForge was was cutting million-dollar deals while GitHub was still relatively small, having launched only in April of that year.
Now, I was in Y Combinator at the time, so I definitely shared pg's "myopia", and at that point basically everyone I knew had switched to Git and GitHub. I thought to myself, "These guys have already lost, and they don't even know there's a war." And indeed things have played out exactly as I expected, with Git and GitHub crushing Subversion and SourceForge into irrelevance.
So you see that there's a fundamental asymmetry between what pg sees in front of him and what you might see. Being in YC means you're "living in the future" [1], which is why pg writes with such confidence about the direction things are going. Subversion, SourceForge, Windows; Git, GitHub, Mac, Linux. The latter may still be big in absolute numbers, but the derivative tells you that the smart money is on the latter going forward.
I myself have always had both Windows and Linux (and occasionally OSX) boxen around to help me do what I need to do - so what's right in front of me doesn't seem so one-dimensional as what PG writes about. I've also been in and out of some radically different industries (for example, appdev consulting vs aerospace & defense), and it seems pretty obvious to me that "best tool for the job" is the future and it always has been. That means there will always be a place for Windows and BSD/Linux+derivatives (I throw OSX in this bucket), and there will always be engineers of many stripes using each one (or multiples).
I mean, seriously: even Git - which really does seem to have utterly won the VCS wars from a certain perspective, is still just one of many in reality. Lots of people use Perforce. Legion upon legion still use TFS. BitKeeper is still kicking around. Subversion still has substantial usership. Even CVS is still around. Pretty much the only VCS technology that has truly been defeated is RCS/SCCS, and probably only because they never figured out how to handle directories or else they'd still be around too.
That's why I find these kinds of prognostications to be very silly (although fertile ground for gathering up fake internet points and/or blog readership).
Windows is the one and only OS that runs video games well.
Same with a lot of other industries dealing with professionals who are not programmers. Example is modeling software (https://www.rhino3d.com/mac-feature-compare).
Maybe Valve will succeed with their Steam Boxes, but it's not the case yet.
I wrote "slightly edgy", which obviously changes the meaning, or else you wouldn't have deleted a word. And yes, Python and Ruby are slightly edgy compared to, say, Java or C++.
All the computer people use Macs or Linux now. Windows is for grandmas, like Macs used to be in the 90s. So not only does the desktop no longer matter, no one who cares about computers uses Microsoft's anyway.
Obviously, pg is exaggerating slightly here for effect, but if you go to a conference for a "cool" language like Elixir or Go—or even for a more established but still slightly edgy language like Python or Ruby—what you'll see in the audience is a few islands of Windows in a sea of Linux and Macs. Go to a more corporate-y conference and you'll see a lot more Windows. So—which world is the past, and which is the future? (Hint: At the corporate-y conference, are there more or fewer Macs than there were five years ago?)