[sheepscreek]

In that regard, it’s even easier to use one Apple Studio with sufficient RAM and llama.cpp or even PyTorch for inference.