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

Sparks of Artificial General Intelligence: Early Experiments with GPT-4

arxiv.org

Post reply on HN
Print view View original

101–110 of 244 posts

Re: Sparks of Artificial General Intelligence: Early Experiments with GPT-4

#101

Post by KeplerBoy »

csdvrx wrote: post #84
reidjs wrote: post #65

Well of course Microsoft is going to say something sensational about it, aren’t they in charge of the project somewhat? This is just an advertisement for them, by them.

IDK, but Microsoft seems to be now what Google was a many many years ago: a company creating tech I like to use such as Bing, Edge, Windows Terminal, VSCode, etc. Their Surface hardware is nice too (even if I prefer thinkpads) Oh and they're also helping with the linux kernel. Why can't old people let go? Companies aren't people - they respond to market incentices. Yes, Microsoft did bad stuff in the 1990s, but now t…

you like to use stuff like the windows terminal and edge?

both are passable, but nothing to write home about, are they?

Re: Sparks of Artificial General Intelligence: Early Experiments with GPT-4

#103

Post by highduc »

worrycue wrote: post #94
hislaziness wrote: post #79

For whatever reason we seem to have set a very high expectation from AI as compared to NI (Natural Intelligence). I remember reading "There is a considerable overlap between the intelligence of the smartest bears and the dumbest tourists."

We get our expectations from fiction. AIs in shows like Star Trek are precise and accurate - the perfect complement to the unreliability of humans. That’s what we want.

You'd think, but most humans would rather have "someone" who's lying to them in a very pleasant manner. People don't like objective truth, they go to great lengths to avoid it.

Re: Sparks of Artificial General Intelligence: Early Experiments with GPT-4

#106

Post by blueorange8 »

js8 wrote: post #89
sterlind wrote: post #85

Earlier quoted context omitted.

can you solve general instances of SAT? can the average person?

With enough patience, yes. For example: You have a goat, a wolf, a cabbage and you want to cross a river...

That Goat wolf cabbage problem gpt-4 can solve already

Re: Sparks of Artificial General Intelligence: Early Experiments with GPT-4

#107

Post by otabdeveloper4 »

beoberha wrote: post #75

I know enough about how neural nets work to be absolutely blown away at how good the GPT are. I only skimmed the paper, but even chatGPT showed a lot of these “sparks”, IMO. We are certainly a long way off from any semblance of general intelligence, but for a model that just tries to predict the next word, I’m dumbfounded at how good it is.

Maybe the words we write aren't as smart as we think.

I mean, The Akinator can read your thoughts and that thing hasn't even graduated to a neural network from "a bunch of if/then statements".

Re: Sparks of Artificial General Intelligence: Early Experiments with GPT-4

#108

Post by ly3xqhl8g9 »

skybrian wrote: post #73
PaulDavisThe1st wrote: post #64

I can't help but the hear the distant, very very quiet echo of Clever Hans.

It would be interesting to figure out how Clever Hans does it, though. Don’t you want to know the tricks? Even when it’s a cheat, it might be a clever one. For example, researchers eventually figured out that image recognition algorithms pay attention to textures.

Apparently the horse 'knew' the right answer by inferring from the questioner's behaviour: "Pfungst (the debunker) then examined the behaviour of the questioner in detail, and showed that as the horse's taps approached the right answer, the questioner's posture and facial expression changed in ways that were consistent with an increase in tension, which was released when the horse made the final, correct tap. This provided a cue that the horse could use to tell it to stop tapping." [1] However, there are gene regulatory networks that can actually count up to 3, with the mechanism of counting up to 2 being curiously different than the one for counting up to 3. [2]

"Every intelligence test is also a test of the questioner" [3]: we don't regard a simple liver cell as intelligent, yet it performs a complex task in a large problem space.

[1] https://en.wikipedia.org/wiki/Clever_Hans#:~:text=Pfungst%20....

[2] 2013, Malte Lehmann, "Genetic Regulatory Networks that count to 3", https://pubmed.ncbi.nlm.nih.gov/23567648

[3] Michael Levin, "Bioelectric Networks: Taming the Collective Intelligence of Cells for Regenerative Medicine", https://www.youtube.com/watch?v=41b254BcMJM

Re: Sparks of Artificial General Intelligence: Early Experiments with GPT-4

#109

Post by hislaziness »

ftxbro wrote: post #56
hislaziness wrote: post #5

I remember reading this somewhere - "There is a considerable overlap between the intelligence of the smartest bears and the dumbest tourists.". Though I do not think GPT-4 is even close to AGI it can definitely claim to be better at faking it than many intelligent beings can.

so we are at the snapshot in time where people think 'AI is smarter than many people but not even close to being as smart as me'

It does not have to be "me". My point is we seem to have a different benchmark for Natural Intelligence vs Artificial Intelligence.

Re: Sparks of Artificial General Intelligence: Early Experiments with GPT-4

#110

Post by IanCal »

I'm increasingly convinced you can build an agi system with gpt4.

People are trying to get it to solve everything up front but I've had GPT3 do much better by taking it through a problem asking it questions. Then I realised it was good at asking those questions too so just hooked it up to talk to itself with different roles. Gpt4 seems much better overall and is very good at using tools if you just tell it how and what it has available.

With a better setup than reAct, better memory storage and recall, I think it'd be an agi. I'm not hugely convinced it isn't anyway - it's better than most people at most tasks I've thrown at it.

Oh, and gpt came up with better roles for the "voices in the head" than I did too.

Post reply on HN