DISTRIBUTED INFERENCE ON LLAMA CPP

Summary from the TLDR AI newsletter

Llama cpp now supports distributed inference across multiple machines. It is limited to FP16 for now, but is a great step for open source deployment.