Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

IQ tests results for AI

trackingai.org

Post reply on HN
Print view View original

121–130 of 359 posts

Re: IQ tests results for AI

#121

Post by A4ET8a8uTh0_v2 »

pona-a wrote: post #56
brabel wrote: post #48

Earlier quoted context omitted.

Really? If not genetics then what is it? Just random??

Education? Or more directly, socio-economics. The many of the subjects tested never had any experience with this kind of formal testing, had little to no education, and of course predictably failed on several abstract tasks. It might be that the very pattern of sitting down and intensely focusing on apparently meaningless problems isn't as innate as expected.

Eh, I wish we could retire socioeconomics as a valid term. It can mean just about any factor and for it to be used it scientific publications and even normal discourse that goes beyond Facebook level discussion, feel a little counter-productive.

What, exactly, did you mean though?

Re: IQ tests results for AI

#122

Post by dwroberts »

> Note: VERBAL models are asked using the verbalized test prompt. VISION models are asked the test image instead without any text prompts.

Just glancing at the bar graphs, the vision models mostly suck across the board for each question. Whereas verbal ones do OK.

And today's example of clock faces (#17) does a good job of demonstrating why: because when a lot of the diagrams are explained verbally, it makes it significantly easier to solve.

Maybe it's just me, but #17 for example - it's not immediately obvious those are even supposed to represent clocks, and yet the verbal prompt turns each one into clock times for the model (e.g. 1:30) which feels like 50% of the problem being solved before the model does anything at all.

Re: IQ tests results for AI

#123

Post by breakyerself »

BoingBoomTschak wrote: post #88
cedilla wrote: post #41

Earlier quoted context omitted.

"The Bell Curve" is, let's say, highly controversial and not a good introduction into the topic. Its claim that genetics are the main predictor of IQ, which was very weakly supported at the time, has been completely and undeniably refuted by science in the thirty years since it's publication.

How much of that controversy is manufactured, though? https://en.wikipedia.org/wiki/The_IQ_Controversy,_the_Media_... (if even Wikipedia can't drag this through the mud...)

It's well deserved. The bell curve isnt a good work of science.

Re: IQ tests results for AI

#124

Post by breakyerself »

hemabe wrote: post #72
cedilla wrote: post #41

Earlier quoted context omitted.

"The Bell Curve" is, let's say, highly controversial and not a good introduction into the topic. Its claim that genetics are the main predictor of IQ, which was very weakly supported at the time, has been completely and undeniably refuted by science in the thirty years since it's publication.

And yet, in the US, the first start-ups are offering the possibility of testing embryos for their IQ. https://www.theguardian.com/science/2024/oct/18/us-startup-c...

So if a start up makes a claim it must be real? Theranos investors would like a word.

Re: IQ tests results for AI

#126

Post by twodave »

nsoonhui wrote: post #31
gpt5 wrote: post #6

The way human IQ testing developed is that researchers noticed people who excel in one cognitive task tend to do well in others - the “positive manifold.” They then hypothesized a general factor, “g,” to explain this pattern. Early tests (e.g., Binet–Simon; later Stanford–Binet and Wechsler) sampled a wide range of tasks, and researchers used correlations and factor analysis to extract the common component, then norm…

Another component of this theory concerning g is that it's largely genetic, and immune to "intervention" AKA stability as you mentioned. See the classic "The Bell Curve" for a full exposition. Which makes me wonder what's the point of all the intervention in the form of teaching/parenting styles and whatnot, if g factor is nature and immutable by large? What's the logic of the educators here?

One’s ceiling may be more or less stable, but there are many instances where individuals have certain underdeveloped cognitive skills (of which there are a litany), undergo training to develop those skills, and then afterwards go on to score (sometimes much) higher on IQ tests. Children with certain disabilities such at autism or FASD tend to see more dramatic differences. This isn’t to say they “became more intelligent,” but rather that the testing is unable to measure intelligence directly, rather relying on those certain cognitive skills as a proxy for intelligence.

Re: IQ tests results for AI

#127

Post by root_axis »

nialse wrote: post #63
cedilla wrote: post #41

Earlier quoted context omitted.

"The Bell Curve" is, let's say, highly controversial and not a good introduction into the topic. Its claim that genetics are the main predictor of IQ, which was very weakly supported at the time, has been completely and undeniably refuted by science in the thirty years since it's publication.

Do note that The Bell Curve is not considered controversial in general. The part about race and genetics is. Also genes being the sole predictor of IQ is not an accurate description of the book’s premise.

The controversial part is specifically the policy recommendations that we avoid investing resources in certain racial groups because problems in their communities can be explained by genetics.

Re: IQ tests results for AI

#128

Post by A4ET8a8uTh0_v2 »

terminalshort wrote: post #97
nsoonhui wrote: post #31

Earlier quoted context omitted.

Another component of this theory concerning g is that it's largely genetic, and immune to "intervention" AKA stability as you mentioned. See the classic "The Bell Curve" for a full exposition. Which makes me wonder what's the point of all the intervention in the form of teaching/parenting styles and whatnot, if g factor is nature and immutable by large? What's the logic of the educators here?

There is no logic. Some people are just born much smarter than others and there is nothing you can do about it if you aren't one of them. The implications of that fact are too much for most people to accept so we pretend that it isn't true.

That is a very fatalist perspective. In HS, I was in a class with 3 kids, who were genuinely smarter than me ( not by a notch, mind; comparatively speaking I was just dumber ). I agree that pretending I was/am not dumb by comparison would not help, but suggesting 'nothing you can do about it' is not entirely accurate either. Frankly, the reason I slowly got less dumb is because of my exposure to them. There is something awe-inspiring about seeing a guy using cold logic to guide you through something and be 100% correct.

Re: IQ tests results for AI

#129

Post by TrackerFF »

When I took the WAIS test, some parts of it would be trivial for computers. Particularly working memory, like getting a sequence of unordered numbers and letters, and then sorting them, and telling them back.

Or the visual one, where you have two images that are very similar, and have to find the difference.

Re: IQ tests results for AI

#130

Post by harimau777 »

I could see AI scoring artificially high on IQ tests in a similar way to how it performs artifically well in chess.

As I understand it, chess AI isn't actually particularly good at playing chess intelligently, it's just that high level chess devolves into memorization and computers have an infinite ability to memorize scenarios.

In a similar way, from what I remember from taking an IQ test as a child, the tests are built on the assumption that the test taker has a finite memory and a finite amount of time to learn. In that case, having learned and remembered an unusually large amount for one's age could reasonably correlate with intelligence. However, without that limitation, the ability to answer the questions may not actually correlate with intelligence.

Post reply on HN