Qualcomm offers an extensive developer network and robust tools for deploying AI on the new Snapdragon® X Elite chips. However, I found the process daunting. The platform seems particularly suited for running smaller models tailored to specific inference tasks. While there are examples demonstrating how to deploy large language models (LLMs), they often fall short of being practically usable. From less than informed perspective, certain layers of these LLMs could significantly benefit from the Neural Processing Unit (NPU), but this would require redesigning the runtime specifically for Qualcomm's chip architecture.