In practice, trying to satisfy everyone's target for perfection took forever and immediately after launch the vision of perfection was revealed to be incorrect anyway; this is the classic "waterfall" failure mode. It also (incorrectly) perceived the cost of delay to be zero.

If you are optimizing for something like "actual business value delivered over time" in most typical business contexts (maybe not SpaceX), you are better off putting a ceiling on everyone's idea of perfection, whether that's stakeholders, designers, devs, etc and focusing on launching quickly. From there, you can choose whether to continue to iterate, or move to something else, based on real world feedback.

It's not really being against perfectionism; there are just many projects and diminishing returns.