Assuming you've got an AMD video card, and you probably do, you can still run local AI projects. The ToshLLM project claims "Native Metal acceleration, zero cloud, and no per-token bill." It won't be ...