I think the difficulty with self education often boils down to nobody telling me, "here is what you will learn over the next three months on Tuesdays and Thursdays at 10."
It seems to me that those difficulties are more psychological than structural in that learning on one's own is a bit scary, but Tuesdays and Thursdays at 10 probably aren't going to work well for me because, like life, you know.
But it turns out that not having a timetable for completing the work and taking the test means I am free to take on learning mathematics or Lisp or Linux as long-term projects like learning architecture and the further in the rearview mirror college becomes the less all those years of habits in response to academic calendars seem like the natural order of the way things are learned. I'm by no means the smartest person to suggest this. [1]
Not knowing what to learn next has two issues, one is that outside of academia knowledge is not an iterable. There is no 'next' or rather there are infinitely many and photography is a next of Python and Linear Algebra is a next of FIFA 17 and all four can be learned concurrently. The second is that 'you will probably never use that after you get out of school' applies to things that are learned after getting out of school too.
My explicit advice: when you find the right thing to learn next you will look at high school and college and post-doc level resources with the similarly high levels of enthusiasm because you cannot get enough of the subject. You will also discard and skim resources at all those levels when the learning becomes smally incremental and later, perhaps several years, some you will circle around to because you did not understand them the first time or because you understand them differently now that you have learned more.
Good luck.
[1]: http://norvig.com/21-days.html
