Interesting. I have a mac mini I run headless dedicated for LLM inference. Do you know if I could run project nomad on a linux box but use qwen on the mac for the AI stuff? I can't find any info on if this configuration is possible.