In this case, I worked for a startup still trying to find product-market-fit, so we went through about 4 different app versions with completely different sets of features.

We knew that the first run-through of a feature was to test if people used it at all, and if it didn't stick, we'd remove it. If it stuck around, then it would be made better and better over time. This often occurred through "pre-refactoring" to make slotting in your feature easier.

We still invested in high-quality code, followed TDD as an entire team, and paired on most features. But we also knew when something was good enough, and we should walk away from it.

I completely agree. "We have to trust the process and keep the quality high." Exactly how high changes depending on what you're doing.