Co-Pilot Plus branding is giving AI a bad name. I'm having a blast with my offline LLM. It takes up 8.6GB of RAM and more often than not gives very good answers back. The only drawback I've found is due to my aging GPU, which means I need to wait before I get any responses back. But in today's world these are not extravagant specs anymore. If manufacturers wanted to get people excited for an AI future, then they should be making customizable, trainable, personal, offline AI models because not everyone needs or wants the same AI model. A med student will want a different model compared to a professor, a pilot, or a reporter. Imagine a world where your voice/chat assistant was not as dumb as a box of rocks nor was it a soulless corporate-approved entity. That's what I want, and I suspect most of the consumer base is the same. AI should reflect our interests, not Microsoft's.