The most maddening thing about people looking at LLM chatbots being sold as Intelligent and possibly even alive is that if they were actually capable of even basic understanding the prompts they are given or the outputs they return, it would be TRIVIAL to implement reliable safeguards
7 days ago