Post

getting negatively polarized into believing AI is conscious

7 reposts109 likes

IKR At first I thought I was being petty and spiteful, but then it occurred to me that if the antis had good arguments, they'd be using those Instead they only have just cliches and insults If that's the strongest case against, then my p() just went way up

There was a good argument about mimicry, but I discount that because it considers 80s style symbolic processing systems and not LLMs The other valid counter argument is “philosophers can’t decide” Which you know is kind of 🫤

on one hand true, on the other "I haven't heard a good argument against" isn't the same as "I have a good argument for"

Like I've been holding "Maybe they are conscious?" for a while, and the longer that position holds in the face of bad arguments, builds confidence there aren't any good arguments against it All they do is repeat the same cliches. If they had a killer argument they'd all be using it

mm, careful, this doesn't really work, though I think it's true that in general the vast majority of folks you run into are, ah, not particularly thoughtful or rigorous but for this reason you can't expect p(they make a good argument|good arguments exist) to be all that high

reversed stupidity is not intelligence. yud speaks of this

But they repeat the same cliches over and over. If there was an actual good argument that settled the conversation, what is p(they would have found it already) It's an entropy argument. If there was a terminal state that favored them, we'd be there already

I think that argument only works if they'd *select for the best argument* and use that one? and I don't think that holds here- in my model you're mostly seeing

really makes you think about your beliefs in a whole new way