Mesh LLM: distributed AI computing on iroh

Mesh LLM is a distributed computing system built on iroh that allows users to pool GPUs and memory across multiple machines and access them through a single OpenAI-compatible API, eliminating dependence on expensive cloud providers. The system can run models locally, route requests to peer machines, or split large models across multiple nodes in a pipeline, giving users control over costs, data privacy, and hardware. It offers an alternative to expensive centralized AI services by leveraging existing computing resources that businesses already own.