Decentralized GPU compute network for distributed LLM inference. Aggregate consumer GPUs to run large language models at a fraction of the cost.