I don't think the big picture amounts to much anyway. It's really just not very impressive that LLMs, which were trained on social media posts, can almost convincingly mimic social media. "programs designed to mimic conversation do that" just isn't surprising even when they do it with a chatbot on both ends.