This week, I am testing Splash, a local LLM inference engine for Apple Silicon, on a MacBook Pro with an M5 Max 128GB.On Friday, I stepped away from the bench and decided to use Qwen3.8-27B-Splash as ...