It's not about thinking, it's about feeling. If half-baked but apparently working features inspire confidence and are perceived as progress by clients and investors, then that perception will drive the rewards and thus the process. The phrase around 1999 was "who cares if it crashes randomly, the user will just reload the page." Shortly afterwards, amount of bugs fixed becomes an additional metric of success, so delivering less features that also come with less bugs means hitting your success metric twice.

Sad, but remember that many startups never deliver profits, AKA actual tangible value, before they exit.