This needs to be pointed out more often. It also applies to the “fail fast” mantra too. You’re only going to get to something better by believing that each release of what you have is the best there is. When you see it isn’t, you go round again.
On a practical note, when you say:
Releasing half-baked features smashed into a poorly constructed version of a product is not an MVP. It’s a trainwreck, and it’s counter productive.
I would also point out that in the vast majority of cases, looking at such things as visitor numbers or conversions is a standard method of judging the effect of MVPs released to the public. But it’s well to remember that if someone comes to your site once and then never returns, you won’t be able to measure them. So as in life, so on the Internet: first impressions are extremely important.
This is also why concentrating on form over function is such a terrible mistake to make, yet understanding of the practice of information architecture and interaction design is at an all-time low.