Distributed LLM inference. Connect home devices into a powerful cluster to accelerate LLM inference. More devices means faster inference.