I've always been told this with mathematics. You won't really "get" it until you've done enough exercises, especially the more abstract you get (I'm looking at you, Category Theory proofs.) I don't know why it took me this long to apply the same principle to learning new programming languages. I always thought, well if I just start on this monolithic hobby project, I'll learn the language after having read xyz. Well, turns out, it's a lot more entertaining and faster for me to just blow through a bunch of medium-sized throw-away examples for a while first. That monolithic hobby project comes along much easier with less of a "I give up rate" afterwards, for me anyway..