None of this rings true to me. Not at any startup I've worked at/founded at least(6 in total over the last 12 years). In reality good startups are relatively even keel.
Hell I would argue that startups, almost by definition, think MORE in the long-term than their larger counterparts. In a startup you are always progressing towards a long-term goal, constantly taking in feedback and making adjustments.
It's much like sailing to a small island thousands of miles away. A mistake in navigation in the beginning that goes uncaught will put you farther off course than one farther along on your journey (although sometimes that mistake means you discover North America and then pretend like you meant to do that all along).
The point being, if your startup is constantly in panic mode or is constantly trading long-term stability for short-term satisfaction then you should probably be thinking about leaving. At least from my point of view.
Compare to the startup that can just “pivot” at will and/or just end it with one of those “it was an amazing ride” blog posts.
Not trying to be smart, just presenting some perspective.
You better not, who cares about UML diagrams?
But startup code doesn't have to be messy. I don't write messier code (to be used in a company that pays me), just because I'm in a hurry, and neither should anyone.
If code is particularly messy, unplanned, and untested you might be feature searching as opposed to having a vitamin or medicine feature that a customer will pay for.
Release early, get feedback often. Choose a core competency.