NVIDIA is collaborating with Apple to enhance the performance of Language Models (LLMs) in AI.Apple's ReDrafter technique combines beam search with dynamic tree attention, resulting in faster LLM token generation.This collaboration aims to reduce latency for users, use fewer GPUs, and consume less power.The improved AI experience will benefit developers using NVIDIA GPUs for production LLM applications.