Postscript: The macOS kernelspace CSPRNG ***screams***. They're using Fortuna with SHA-2 as their core primitive with hardware-accelerated SHA-2 on their M1, M2, and M3 chipsets.
It's ridiculous how crazy fast their kernelspace CSPRNG is. It was the inspiration for me sending in my kernel patch to LKML, which got rejected.
https://chaos.social/@revolvermann/111694685078465635