Thank you for the shout-out of the Pinker-Aaronson talk- I will now go there to read it up. - But right now, I feel your post is kinda straw-manning Pinker (not just getting his first name wrong): Sure there is sense in talking "smart" and "smarter" among humans. Or among bonobos. Less when it is human-bonobo, much less when humanGI and AI. - Task: "calculate pi to its 100,000,000,000,000th decimal place in less than 6 months". Dumbo: fail, me fail, you fail, John von Neumann: fail. Computer: did it . https://thenewstack.io/how-googles-emma-haruka-iwao-helped-set-a-new-record-for-pi/ So what does that show about General intelligence?? - You say: more strength is better for any sport. Well, ever saw a tank play basketball? - Now AI, new task: "Join this group for improv acting". (And do not say this is not helped by general intelligence.) . - Short: the tasks AI is expected to do - when not playing (chess, go, jeopardy ...) - and what even smart humans do, may very well be too different for "GI" to be useful concept here, esp. to compare AI with HI. (As the "strength" of an aircraft carrier is not in the same league as M. Ali strength was. Not even same sport. Not even sport. - Obviously Klitschko can knock me out, and you can outsmart me. This can not be what Pinker is talking about. - Now I go to check. ;)
Interesting thanks, I think you should give Pinker a bit of a break because he is in academia and needs to walk a narrow line when talking about intelligence. I also think that a computer analogy can be useful here. g (and IQ) is a measure of your hardware. Our culture, language, mathematics is the software we run in our brain. So people with 'better' software might still do something better than someone with a higher IQ, but worse software.
Thank you for the shout-out of the Pinker-Aaronson talk- I will now go there to read it up. - But right now, I feel your post is kinda straw-manning Pinker (not just getting his first name wrong): Sure there is sense in talking "smart" and "smarter" among humans. Or among bonobos. Less when it is human-bonobo, much less when humanGI and AI. - Task: "calculate pi to its 100,000,000,000,000th decimal place in less than 6 months". Dumbo: fail, me fail, you fail, John von Neumann: fail. Computer: did it . https://thenewstack.io/how-googles-emma-haruka-iwao-helped-set-a-new-record-for-pi/ So what does that show about General intelligence?? - You say: more strength is better for any sport. Well, ever saw a tank play basketball? - Now AI, new task: "Join this group for improv acting". (And do not say this is not helped by general intelligence.) . - Short: the tasks AI is expected to do - when not playing (chess, go, jeopardy ...) - and what even smart humans do, may very well be too different for "GI" to be useful concept here, esp. to compare AI with HI. (As the "strength" of an aircraft carrier is not in the same league as M. Ali strength was. Not even same sport. Not even sport. - Obviously Klitschko can knock me out, and you can outsmart me. This can not be what Pinker is talking about. - Now I go to check. ;)
Interesting thanks, I think you should give Pinker a bit of a break because he is in academia and needs to walk a narrow line when talking about intelligence. I also think that a computer analogy can be useful here. g (and IQ) is a measure of your hardware. Our culture, language, mathematics is the software we run in our brain. So people with 'better' software might still do something better than someone with a higher IQ, but worse software.