> Haven't the latest LLMs shown us that a neural net trained to "just string together readable words" lead to at least simple intelligence? LOL no. At least no

> Haven't the latest LLMs shown us that a neural net trained to "just string together readable words" lead to at least simple intelligence? LOL no. At least not for anyone outside of a few echo chambers. I have literally not spoken with a single human outside of one subset of the tech industry who says stuff like this. > I agree. On the other hand, if I tell a bot "I'm going to turn you off now" and it tries to stop me, that implies it feels fear. I wrote a Perl program (an IRC bot) that did this as a teenager. AI achieved /s