PCIe Bandwidth: Key to Fast Inference. Quantization: Smaller Models, Surprising Results. Local LLMs for Creative Assistants. Pair Programming, Local Copilots, and Code AI. Data, Pretraining, and Ethical Foundations