The problem isn't "good enough" software getting the ball rolling. It's when "good enough" continues to be the MVP at scale. The scale of a project, which can be any combination of complexity, user base size, lines of code, people working on it, etc, has a limit that is a result of the quality. The whole issue with a "ball of mud" isn't that it's difficulty to start, it's that it's difficult to maintain and extend. It is quite well studied that a system grows, if you have any centralized contention, you quickly plateau in productivity, and can very well achieve negative scaling.
Amdahl's law for code quality. High quality code has low contention and allows for massive scaling, but can be quite slow at small scales.