Another Knuth-related one that is less programming that I found rewarding to read was 'Concrete Mathematics.' It is per its own description a lengthier version of the TAOCP, but it's mostly or entirely exempt of actual code. But because of the subject matter it spawned from it tends to cover math of 'what comes later' so there is a very specific perspective of Knuth style that comes from it and its easier to see the eventual application.
I had a professor who would joke that it was called Concrete Mathematics both because it was foundational math for computer science, but also because it was as hard as concrete.
