I’m just curious about the potential of running a small LLM locally on GrapheneOS, maybe some image processing too.