I'm surprised that this implementation of xz written by... well... random people has been adopted so widely. I would've expected a more 'industrial' implementation managed by Google or Meta or something, but there isn't one.
I'm still baffled xz took off so massively in the first place. The USP seems to be existing LZMA compression but made significantly more fragile and prone to never decompressing again.
The math and algorithms behind it are fun to learn but hard. And then you need to implement it both performant and correct.
Only a few people build up the algorithmic background to do this. And the gains once an implementation is there are marginal (optimizations).
The only larger one seems to be zstd, and I haven't wrapped my head around ANS/tANS...