Because: 1. It's just not a huge deal to many people. Most people who want to do local ML training and inference can just buy a NVIDIA GPU. 2. AMD only has a

Because: 1. It's just not a huge deal to many people. Most people who want to do local ML training and inference can just buy a NVIDIA GPU. 2. AMD only has a skeleton team working on their solution. It's clear it's not a focus.