The way to tell the difference between the past and the future is to look at the derivative. For example, in an attempt to get to know Windows better, I recently asked the only top Windows-using dev I know what his environment and tooling look like. "Oh, I switched to a Mac last year," he said. That migration is typical, and suggests that people who care about computers will continue preferentially switching from Windows to Mac and Linux.

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.

[1]: http://paulgraham.com/startupideas.html