I feel you. I also don’t like the “if you’re not embarrassed, you haven’t released early”.
We only get one chance at a first impression and it’s very important that we get it right. If we do, then that user is delighted and will mostly likely recommend it to someone else.
That’s why I really like “minimal delightful product” rather than “minimal viable product”.
But “minimal delightful product needs to solve the users job”. This means even if it’s slightly buggy but you manage to solve a painful problem, the user still is delighted with a smile.
So many products I see that are the work of long months of building but don’t solve the users problem.
The process I find works best is a strong canary program. “You don’t spend more than a week without iterating by showing something to a customer and collecting/observing feedback”
The first version is ugly and only meant for a few users, as more holes are fixed, it’s exposed to more and more users until you hit the gold pot and find new users are actually “getting delighted”. That’s when you’d go full GA on a feature with a limited scope that “gets the users job done in a way that delights them”