To me, a big reason LLMs aren’t conscious is that they only respond to user input, generate output and then stop. They don’t talk to themselves. They aren’t sitting their contemplating the meaning of their existence while you are away from the keyboard.