Apple Silicon's Surprisingly Speedy SIMD

I just don’t understand - Apple decided to implement a more capable SIMD instruction set and just … didn’t document it? SIMD is hugely important in anything remotely HPC related. My first gen apple silicon laptop still favorably competes against Intel’s current best, years later. Why would they hide such an advantage?

can people publish blog posts (and other things) literally anywhere else? especially now with nitter/xcancel/etc getting taken down