On relying on libraries that do only the 80%.
This puts it into words so much better than I ever could why I'm skeptical towards the world of today's software engineering, both at the junior and senior levels. It is rare to find a piece of code written with enough thought put into it so that it does what it needs to do, does it well, and doesn't extend its reach into parts where it shouldn't.
And because of that, I tend to write my code from scratch instead of relying on libraries or frameworks that claim to do that for me. Sure, the tradeoff is that it takes much longer, but at least the end product is something that I can place trust in without any doubts.
Yet it is depressing that a lot of this industry doesn't care about everything in the same way. It's the obverse of the whole rest of engineering fields where innate passion for your work and all of its details matter—here we (tend to) only care about the outcome, no matter how poorly it actually works.