It's so easy to hang on to things you really don't need.
Most underlying technology is timeless (see TAOCP, SICP, CLRS, K&R, GoF, Dragon Book, Beej's Guide, Sipser,...); but we seem set on producing an endless, pointless, churn of frameworks and minor language differences in the name of progress.
True enuf, but how often do you actually refer to the books underlying ideas? I've had (now gone, presumably) TAOCP on my bookshelves for years, but how often did I use it? Stuff on RNG a bit, I guess...
My oldest computer book I won’t part with is Alan Simpson’s dBase III+ Programmer's reference guide, circa 1987. This book was transformative and allowed me to get a gig as a coder, so much self driven practice on a crappy underpowered generic clone PC. That crappy hardware was an advantage I didn’t see at the time, having to think about routines that were fast enough based not because of faster disks and tons of RAM.
Do I need this book? Not so much. But it brings me joy carefully flipping through it on occasion.