In any case, I'm not sure that many (any?) product needs could dictate choice of language in the vast majority of cases. Yes, if you need to do a ton of realtime stuff then maybe think about Node etc., but most people want the same standard stuff.
So you can use that method when hiring the first developer.
When it's time to hire the second developer, are you looking for only developers that are comfortable with the same languages as the first developer? If so, you are hiring developers based on language, and no longer following your advice. If not.... then you get a bunch of developers comfortable with different languages, and now what?
Often the answer might be "good developers can learn to be effective at any language." Which while on the whole (if not always) true, nevertheless only sometimes works for an enterprise (sometimes you don't have time to wait for them to learn it). But even so, now that turns into different advice: Hire your first developer, have her use a language she's comfortable with, then hire subsequent developers that either already know that language or are capable of learning it, and have them learn it. Which doesn't sound quite so obvious as your original advice, although it's not neccesarily an unreasonable approach.
Getting something up & running is more important that everything else put together (except maybe determining what that something is.)
It's hard enough to get from "thought" to "thing" without worry about all the details. For an MVP, almost any technology can do almost anything you need. So just do it.
Although not ideal, you can always refactor, rewrite, rearchitect, or redeploy. But unless you have a time machine, you can't go back and do it again. So listen to untog and use what you already know.
"Never use a new language for the first time for a project you love or that's important to you - you'll end up hating both the language and the project."