Problematic advice: So this wasn't really a game development advice per se, but my mom always used to tell me "do it right the first time around and you'll save time and effort" or something like that. This was actually more of a general advice to life itself, and at the time it probably applied primarily to cleaning my room or something, but I basically accepted it and lived by that advice for many many years, including when I started programming. As most of you will recognize, this can be an absolute disaster for productivity. Micro optimizations everywhere! Lets hand optimize that 100-iterations-per-frame-loop in assembly! Let's make sure this engine works equally well for a tetris clone as the next high-end MMO. Since then, I've shifted to a more "let's go with the simplest possible solution that will do the job for now" for almost everything, and then iterate and improve/optimize when needed. My productivity is orders of magnitude higher. You don't need "perfect", you need "good enough".
Good advice: Don't be afraid to throw out and rewrite problematic portions of code. Don't stick with stuff that doesn't do the job just because you spent significant time on it.