Eli Tevet comments on Self-Awareness (and possible mode collapse around it) in ChatGPT