I think most of the difficulty I've had with learning #mathematics from books over the years has been notational polysemy (or abuse). Like, there have been (and continue to be) many times when I get hung up on an equation because I can't tell that it's a definition or a derivation or expansion that's just skipping intermediate steps, not to mention the times when an equation is actually a *temporary* initialization or assignment of value. Then, let's not even get started on all the nasty stuff folks get up to with matrices or eliding variables from an expression.
If I'm in a certain subfield, I can more often figure it out because I've seen the notation before, but it's still a pain.
@2ck I just started taking a second run at Knuth's "The Art of Computer Programming," four years after my first attempt.
Got past the first chapter, but only because I'd taken a detour in the intervening years through the mathematics of mechatronic control systems and had seen what a "state vector" is.
And even then, I want to yeet the phrase "It is obvious that" and all its variants into the sun.