the larger context sizes specifically are what I’m fascinated by. imagine running an LLM locally and feeding it all your data. appointments, relationships, notes whatever. you could also connect it to smart Home devices. I really need to get my hands on a GPU with 16 gigs of vram