Read books with ideas you can apply, and use the ideas in practice. Don't read books that you cannot directly apply to something, that's consuming input without producing output. (I quite enjoy reading, and happily read books that I cannot apply, but this is not "efficient"...)

Value extracted, and even the best way to extract value from a book, will be heavily context dependent. E.g, here's a wildly contrived example: If you are trapped in a hut in a snowstorm you may be able to extract considerable value from a trashy novel or physics textbook in your backpack by tearing it apart to use as kindling.