> We are likely to see LLMs integrated into every part of computing as infrastructure, not just as a product, in the next year or two.

Of course, for collecting better telemetry using local AI for analyzing video from camera and audio from a microphone.