I think the usefulness of insisting on reserving the term "MVP" for a successful experiment is that it could help prevent someone from spending months in develo

I think the usefulness of insisting on reserving the term "MVP" for a successful experiment is that it could help prevent someone from spending months in development, telling themselves "we're building our MVP, we're building our MVP", and thus thinking, when they've built something but not yet shown it to anyone, that it is an MVP. No; you hope it is, but it may not be (and in fact, the odds are against you if you haven't discussed it with potential customers). So this is not just a semantic quibble, but actually supports the thrust of the article. While you're doing market research experiments, you should call them that. When you actually strike gold, then you have an MVP.