It's not really a massive time investment. Consider people you see every morning doing a sudoku or crossword on a patio at a coffee shop or on a subway. That's

It's not really a massive time investment. Consider people you see every morning doing a sudoku or crossword on a patio at a coffee shop or on a subway. That's basically what TAOCP is to me, a past-time/hobby, leisurely going through the psets and trying to do things like patch an ancient algorithm or solve some basic number theory puzzles (Vol 2). Djb often refers and quotes the volumes in talks which is one reason why I started reading them, to understand how to hand roll and optimize assembly libraries https://cr.yp.to/talks/2015.04.16/slides-djb-20150416-a4.pdf https://cr.yp.to/talks/2015.04.16/slides-djb-20150416-a4.pdf There is also the benefits of learning direct programming of hardware, for example MIXAL, used in the first 3 volumes (replaced by MMIXAL) is meant to be run without an operating system. Maybe you want to play around with direct control of a RISC-V chip or fpga or understand the quantum abstract machine Rigetti built and it's instruction language.