Combining Nvidia DGX Spark and Apple Mac Studio for 4x Faster LLM Inference | Dark Hacker News