[int_19h]

They do, but for inference at least, it's memory bandwidth that is the primary limiting factor for home LLMs right now, not raw compute.

[sroussey]

Wonder if the apple silicon ultra series will start using HBM3(e) on desktop in the future.