I think you don't really understand how these LLMs work, and I only skimmed through the text, but it isn't really self aware (yet).
It's fully trained and doesn't learn anything new while you talk to it and only remembers inside the context window.
Just imagine it being really good at coming up with answers that sound right.
That being said, the tech is amazing and it will become better quickly.