Fahd Mirza walks through how to run the Qwen3.5 35B model entirely locally using llama.cpp and then integrate it with OpenClaw, Anthr...