Systems Thinking

Markus Hartikainen runs a team at AMD that scales LLM inference on vLLM, and last week he asked a question I have been answering, in one form or another, for most of my career: …