E.g., easy file versioning and sync is fundamentally a solved problem, if you are willing to take the initial hit of workflow modification (git, sftp, etc).
So it's okay to teach people how to do it faster and better.
If and when they want to learn, sure. But the article is specifically about how you respond to people who don't want to contemplate change, let alone implement it.
Particularly, it's about the non-trivial cases where it would take hours to unravel the goals, features, problems and dependencies of their current 'system' to make an intelligent recommendation, to say nothing of breaking down time, capital and retraining costs to implement it.
This, however, would make the education system not agnostic, and drastically favor some companies.
Why? Perhaps you mean that there is versioning software, like git, that is so prevalent that it may be regarded as having dominated the ecosystem, so that that's what we have to teach; but that seems no different from teaching Java or, for an even more extreme example, the RSA cryptosystem.
(I know that there are objections to teaching Java, but I've never heard—which is not to say it doesn't exist!—the complaint that it favours Oracle.)
Easy file versioning and sync are only solved for a very specific use case that corresponds very poorly to most people's lives (photo library?) There's a priceless rant from John Siracusa about his sister ignoring Time Machine's warnings tha the backup device had failed for months because it didn't actually incapacitate her computer.