Luís Silva on Nostr: I'm going to say something that will shock a lot of people. Even taking into account ...

I'm going to say something that will shock a lot of people. Even taking into account the limitations on the context size and lack of inference of most LLMs, I believe they are, in fact, superior to the cognitive capabilities of the majority of human beings. I know they can't pass the Turing test, but these tests are conducted using humans who are well above the world average. No LLM will beat you in a chess game if you're an average player, and they may do some nonsensical things, but if you look around, you'll find people who behave even worse.