I understand that it's just predicting text and the model is only working when it receives input, but what if someone took the LLM and made some kind of cognitive loop in the background where the same LLM reads information, thinks to itself when you are gone, and stores these things in its chat content then tells you about the cool stuff it read about the next day. It may be just text, but now it's "thinking". At what point should we start to consider it as a creature with rights?