Source item

Gave a lecture today at Stanford in CS 229S - Systems for Machine Learning Topics were AI infrastructure, multi-datacenter training, and asynchronous fault tolerance Also our inference simulator to quantify cost impact of reasoning + throughput vs latency https://t.co/CoGOJlZHKQ https://t.co/ZETKZq3psn Vid here https://t.co/GKoNrO94rZ