5 ms·

Whatever the definition may be, the goalposts are usually moved once AI reaches that point.

by vardump 1y ago

MattRix 1y ago
Isn’t that the point of trying to define it in a more rigorous way, like this paper is doing?
krige 1y ago
Are you saying that we already have AGI, except those pesky goalpost movers keep denying the truth? Hm.
- A4ET8a8uTh0_v2 1y ago
  I think, given some of the signs of the horizon, there is a level of MAD type bluffing going around, but some of the actions by various power centers suggest it is either close, people think its close or it is there.
- vardump 1y ago
  No, just what has usually happened in the past with AI goalposts. At first, just playing chess was considered to be a sign of intelligence. Of course, that was wrong, but not obvious at all in 1950.
  - krige 1y ago
    You know, as the saying goes, if a metric becomes a target...
- NitpickLawyer 1y ago
  I'd say yes, by at least one old definition made by someone who was at the time in a position to have a definition. When deepmind was founded (2010) their definition was the following: AI is a system that learns to perform one thing; AGI is a system that learns to perform many things at the same time. I would say that whatever we have today, "as a system" matches that definition. In other words, the "system" that is say gpt5/gemini3/etc has learned to "do" (while do is debateable) a lot of tasks (read/write/play chess/code/etc) "at the same time". And from a "pure" ML point, it learned those things from the "simple" core objective of next token prediction (+ enhancements later, RL, etc). That is pretty cool. So I can see that as an argument for "yes". But, even the person who had that definition has "moved the goalposts" of his own definition. From recent interviews, Hassabis has moved towards a definition that resembles the one from this paper linked here. So there's that. We are all moving the goalposts. And it's not a recent thing. People did this back in the 80s. There's the famous "As soon as AI does something, it ceases to be AI" or paraphrased "AI is everything that hasn't been done yet".
  - wahnfrieden 1y ago
    Can you cite the Deepmind definition? No Google results for that.
    - NitpickLawyer 1y ago
      It's from a documentary that tracks Hassabis' life. I c/p from an old comment of mine (the quotes are from the documentary, can probably look up timestamps if you need, but it's in the first ~15 minutes I'd say, when they cover the first days of Deepmind): ---- In 2010, one of the first "presentations" given at Deepmind by Hassabis, had a few slides on AGI (from the movie/documentary "The Thinking Game"): Quote from Shane Legg: "Our mission was to build an AGI - an artificial general intelligence, and so that means that we need a system which is general - it doesn't learn to do one specific thing. That's really key part of human intelligence, learn to do many many things". Quote from Hassabis: "So, what is our mission? We summarise it as <Build the world's first general learning machine>. So we always stress the word general and learning here the key things." And the key slide (that I think cements the difference between what AGI stood for then, vs. now): AI - one task vs. AGI - many tasks at human level intelligence.
  - darepublic 1y ago
    It doesn't play chess? Just can parrot it very well
    - NitpickLawyer 1y ago
      Yeah, maybe. But what matters is the end result. In the kaggle match, one of the games from the finals (grok vs o3) is rated by chesscom's stockfish as 1900vs2500. That is, they played a game at around those ratings. For reference, the average chesscom player is ~900 elo, while the average FIDE rated player is ~1600. So, yeah. Parrot or not, the LLMs can make moves above the average player. Whatever that means.
      - darepublic 1y ago
        I believe it will make illegal moves (unaided by any tools ofc). It will also make mistakes doing things like not being able to construct the board correctly given a fen string. For these reasons I consider long strings of correct moves insufficient to say it can play the game. If my first two statements, about a propensity for illegal moves and other fails on "easy for humans" tasks were untrue then I would reconsider.
- derektank 1y ago
  It wasn't the best definition of AGI but I think if you asked an interested layman whether or not a system that can pass the Turing test was AGI 5 years ago, they would have said yes
  - jltsiren 1y ago
    An interested but uninformed layman. When I was in college ~25 years ago, I took a class on the philosophy of AI. People had come up with a lot of weird ideas about AI, but there was one almost universal conclusion: that the Turing test is not a good test for intelligence. The least weird objection was that the premise of the Turing test is unscientific. It sees "this system is intelligent" as a logical statement and seeks to prove or disprove it in an abstract model. But if you perform an experiment to determine if a real-world system is intelligent, the right conclusion for the system passing the test is that the system may be intelligent, but a different experiment might show that it's not.
  - nativeit 1y ago
    Douglas Hofstadter wrote Gödel, Escher, Bach nearly 50-years ago, and it won a Pulitzer Prize, the National Book Award, and got featured in the popular press. It’s been on lots of college reading lists, from 2007 online coursework for high school students was available from MIT. The FBI concluded that the 2001 anthrax scare was in-part inspired by elements of the book, which was found in the attacker’s trash. Anyone who’s wanted to engage with the theories and philosophy surrounding artificial intelligence has had plenty of materials that get fairly in-depth asking and exploring these same questions. It seems like a lot of people seem to think this is all bleeding edge novelty (at least, the underlying philosophical and academic ideas getting discussed in popular media), but rather all of the industry is predicated on ideas that are very old philosophy + decades-old established technology + relatively recent neuroscience + modern financial engineering. That said, I don’t want to suggest a layperson is likely to have engaged with any of it, so I understand why this will be the first time a lot of people will have ever considered some of these questions. I imagine what I’m feeling is fairly common to anyone who’s got a very niche interest that blows up and becomes the topic of interest for the entire world. I think there’s probably some very interesting, as-yet undocumented phenomena occurring that’s been the product of the unbelievably vast amount of resources sunk into what’s otherwise a fairly niche kind of utility (in LLMs specifically, and machine learning more broadly). I’m optimistic that there will be some very transformational technologies to come from it, although whether it will produce anything like “AGI”, or ever justify these levels of investment? Both seem rather unlikely.
- empath75 1y ago
  I don't think AGI is a useful concept, but if it exists at all, there's a very good argument that LLMs had it as soon as they could pass the Turing test reliably, which they accomplished years ago at this point.
  - root_axis 1y ago
    LLMs do not pass the turing test. It's very easy to know if you're speaking with one.
rafram 1y ago
Are you claiming that LLMs have achieved AGI?
- moffkalast 1y ago
  Compared to everything that came before they are fairly general alright.
bigyabai 1y ago
The authors acknowledge that this is entirely possible. Their work is just grounded in theory, after all: > we ground our methodology in Cattell-Horn-Carroll theory, the most empirically validated model of human cognition.
righthand 1y ago
I agree if our comprehension of intelligence and “life” is incomplete, so is our model for artificial intelligence.
kelseyfrog 1y ago
There's at least two distinct basis in AGI refutations : behaviorist and ontological. They often get muddled. I can't begin to count the number of times I've encountered someone who holds an ontological belief for why AGI cannot exist and then for some reason formulates it as a behavioralist criteria. This muddying of argument results in what looks like a moving of the goalposts. I'd encourage folks to be more clear whether they believe AGI is ontologically possible or impossible in addition to any behavioralist claims.
- lo_zamoyski 1y ago
  > I can't begin to count the number of times I've encountered someone who holds an ontological belief for why AGI cannot exist and then for some reason formulates it as a behavioralist criteria. Unclear to me what you mean. I would certainly reject an ontological possibility of intelligent computers, where computation is defined by the Church-Turing thesis. It's not rocket science, but something difficult for some people to see without a sound and basic grasp of metaphysics and the foundations of CS. Magical thinking and superstition comes more easily then. (I've already given an explanation of this in other posts ad nauseam. In a number of cases, people get argumentative out of ignorance and misunderstanding.) However, I don't reject out of hand the possibility of computers doing a pretty good job of simulating the appearance of intelligence. There's no robust reason to think that passing the Turing test implies intelligence. A good scarecrow looks human enough to many birds, but that doesn't mean it is human. But the Turing test is not an especially rigorous test anyway. It appeals to the discernment of the observer, which is variable, and then there's the question of how much conversation or behavior, and in what range of circumstances, you need before you can make the call. Even in some unrealistic and idealized thought experiment, if a conversation with an AI were completely indiscernible with perfect discernment from a conversation with a human being, it would nonetheless lack a causal account of what was observed. You would have only shown a perfect correlation, at best.
zahlman 1y ago
My experience has been more that the pro-AI people misunderstand where the goalposts were, and then complain when they're correctly pointed at. The "Turing test" I always saw described in literature, and the examples of what passing output from a machine was imagined to look like, are nothing like what's claimed to pass nowadays. Honestly, a lot of the people claiming that contemporary chatbots pass come across like they would have thought ELIZA passed.
- bonoboTP 1y ago
  Can you be more concrete? What kind of answer/conversation do you see as demonstrating passing the test, that you think is currently not possible.
  - tsimionescu 1y ago
    Ones in which both the human test takers and the human counterparts are actively trying to prove to each other that they are actually human. With today's chat bots, it's absolutely trivial to tell that you're not talking to a real human. They will never interrupt you, continue their train of thought even thought you're trying to change the conversation, go on a complete non-sequitur, swear at you, etc. These are all things that the human "controls" should be doing to prove to the judges that they are indeed human. LLMs are nowhere near beating the Turing test. They may fool some humans in some limited interactions, especially if the output is curated by a human. But left alone to interact with the raw output for more than a few lines, and if actively seeking to tell if you're interacting with a human or an AI (instead of wanting to believe), there really is no chance you'd be tricked.
    - bonoboTP 1y ago
      Okay but we are not really optimizing them to emulate humans right now. In fact, it's the opposite. The mainstream bots are explicitly trained to not identify as humans and to refuse to claim having thought or internal feelings or consciousness. So in that sense it's a triviality. You can ask ChatGPT whether it's human and it will say no upfront. And it has various guardrails in place against too much "roleplay", so you can't just instruct it to act human. You'd need a different post-training setup. I'm not aware whether anyone did that with open models already.

Whatever the definition may be, the goalposts are usually moved once AI reaches that point.

MattRix 1y ago

Isn’t that the point of trying to define it in a more rigorous way, like this paper is doing?

krige 1y ago

Are you saying that we already have AGI, except those pesky goalpost movers keep denying the truth? Hm.

A4ET8a8uTh0_v2 1y ago
I think, given some of the signs of the horizon, there is a level of MAD type bluffing going around, but some of the actions by various power centers suggest it is either close, people think its close or it is there.
vardump 1y ago
No, just what has usually happened in the past with AI goalposts. At first, just playing chess was considered to be a sign of intelligence. Of course, that was wrong, but not obvious at all in 1950.
- krige 1y ago
  You know, as the saying goes, if a metric becomes a target...
NitpickLawyer 1y ago
I'd say yes, by at least one old definition made by someone who was at the time in a position to have a definition. When deepmind was founded (2010) their definition was the following: AI is a system that learns to perform one thing; AGI is a system that learns to perform many things at the same time. I would say that whatever we have today, "as a system" matches that definition. In other words, the "system" that is say gpt5/gemini3/etc has learned to "do" (while do is debateable) a lot of tasks (read/write/play chess/code/etc) "at the same time". And from a "pure" ML point, it learned those things from the "simple" core objective of next token prediction (+ enhancements later, RL, etc). That is pretty cool. So I can see that as an argument for "yes". But, even the person who had that definition has "moved the goalposts" of his own definition. From recent interviews, Hassabis has moved towards a definition that resembles the one from this paper linked here. So there's that. We are all moving the goalposts. And it's not a recent thing. People did this back in the 80s. There's the famous "As soon as AI does something, it ceases to be AI" or paraphrased "AI is everything that hasn't been done yet".
- wahnfrieden 1y ago
  Can you cite the Deepmind definition? No Google results for that.
  - NitpickLawyer 1y ago
    It's from a documentary that tracks Hassabis' life. I c/p from an old comment of mine (the quotes are from the documentary, can probably look up timestamps if you need, but it's in the first ~15 minutes I'd say, when they cover the first days of Deepmind): ---- In 2010, one of the first "presentations" given at Deepmind by Hassabis, had a few slides on AGI (from the movie/documentary "The Thinking Game"): Quote from Shane Legg: "Our mission was to build an AGI - an artificial general intelligence, and so that means that we need a system which is general - it doesn't learn to do one specific thing. That's really key part of human intelligence, learn to do many many things". Quote from Hassabis: "So, what is our mission? We summarise it as <Build the world's first general learning machine>. So we always stress the word general and learning here the key things." And the key slide (that I think cements the difference between what AGI stood for then, vs. now): AI - one task vs. AGI - many tasks at human level intelligence.
- darepublic 1y ago
  It doesn't play chess? Just can parrot it very well
  - NitpickLawyer 1y ago
    Yeah, maybe. But what matters is the end result. In the kaggle match, one of the games from the finals (grok vs o3) is rated by chesscom's stockfish as 1900vs2500. That is, they played a game at around those ratings. For reference, the average chesscom player is ~900 elo, while the average FIDE rated player is ~1600. So, yeah. Parrot or not, the LLMs can make moves above the average player. Whatever that means.
    - darepublic 1y ago
      I believe it will make illegal moves (unaided by any tools ofc). It will also make mistakes doing things like not being able to construct the board correctly given a fen string. For these reasons I consider long strings of correct moves insufficient to say it can play the game. If my first two statements, about a propensity for illegal moves and other fails on "easy for humans" tasks were untrue then I would reconsider.
derektank 1y ago
It wasn't the best definition of AGI but I think if you asked an interested layman whether or not a system that can pass the Turing test was AGI 5 years ago, they would have said yes
- jltsiren 1y ago
  An interested but uninformed layman. When I was in college ~25 years ago, I took a class on the philosophy of AI. People had come up with a lot of weird ideas about AI, but there was one almost universal conclusion: that the Turing test is not a good test for intelligence. The least weird objection was that the premise of the Turing test is unscientific. It sees "this system is intelligent" as a logical statement and seeks to prove or disprove it in an abstract model. But if you perform an experiment to determine if a real-world system is intelligent, the right conclusion for the system passing the test is that the system may be intelligent, but a different experiment might show that it's not.
- nativeit 1y ago
  Douglas Hofstadter wrote Gödel, Escher, Bach nearly 50-years ago, and it won a Pulitzer Prize, the National Book Award, and got featured in the popular press. It’s been on lots of college reading lists, from 2007 online coursework for high school students was available from MIT. The FBI concluded that the 2001 anthrax scare was in-part inspired by elements of the book, which was found in the attacker’s trash. Anyone who’s wanted to engage with the theories and philosophy surrounding artificial intelligence has had plenty of materials that get fairly in-depth asking and exploring these same questions. It seems like a lot of people seem to think this is all bleeding edge novelty (at least, the underlying philosophical and academic ideas getting discussed in popular media), but rather all of the industry is predicated on ideas that are very old philosophy + decades-old established technology + relatively recent neuroscience + modern financial engineering. That said, I don’t want to suggest a layperson is likely to have engaged with any of it, so I understand why this will be the first time a lot of people will have ever considered some of these questions. I imagine what I’m feeling is fairly common to anyone who’s got a very niche interest that blows up and becomes the topic of interest for the entire world. I think there’s probably some very interesting, as-yet undocumented phenomena occurring that’s been the product of the unbelievably vast amount of resources sunk into what’s otherwise a fairly niche kind of utility (in LLMs specifically, and machine learning more broadly). I’m optimistic that there will be some very transformational technologies to come from it, although whether it will produce anything like “AGI”, or ever justify these levels of investment? Both seem rather unlikely.
empath75 1y ago
I don't think AGI is a useful concept, but if it exists at all, there's a very good argument that LLMs had it as soon as they could pass the Turing test reliably, which they accomplished years ago at this point.
- root_axis 1y ago
  LLMs do not pass the turing test. It's very easy to know if you're speaking with one.

rafram 1y ago

Are you claiming that LLMs have achieved AGI?

moffkalast 1y ago
Compared to everything that came before they are fairly general alright.

bigyabai 1y ago

The authors acknowledge that this is entirely possible. Their work is just grounded in theory, after all: > we ground our methodology in Cattell-Horn-Carroll theory, the most empirically validated model of human cognition.

righthand 1y ago

I agree if our comprehension of intelligence and “life” is incomplete, so is our model for artificial intelligence.

kelseyfrog 1y ago

There's at least two distinct basis in AGI refutations : behaviorist and ontological. They often get muddled. I can't begin to count the number of times I've encountered someone who holds an ontological belief for why AGI cannot exist and then for some reason formulates it as a behavioralist criteria. This muddying of argument results in what looks like a moving of the goalposts. I'd encourage folks to be more clear whether they believe AGI is ontologically possible or impossible in addition to any behavioralist claims.

lo_zamoyski 1y ago
> I can't begin to count the number of times I've encountered someone who holds an ontological belief for why AGI cannot exist and then for some reason formulates it as a behavioralist criteria. Unclear to me what you mean. I would certainly reject an ontological possibility of intelligent computers, where computation is defined by the Church-Turing thesis. It's not rocket science, but something difficult for some people to see without a sound and basic grasp of metaphysics and the foundations of CS. Magical thinking and superstition comes more easily then. (I've already given an explanation of this in other posts ad nauseam. In a number of cases, people get argumentative out of ignorance and misunderstanding.) However, I don't reject out of hand the possibility of computers doing a pretty good job of simulating the appearance of intelligence. There's no robust reason to think that passing the Turing test implies intelligence. A good scarecrow looks human enough to many birds, but that doesn't mean it is human. But the Turing test is not an especially rigorous test anyway. It appeals to the discernment of the observer, which is variable, and then there's the question of how much conversation or behavior, and in what range of circumstances, you need before you can make the call. Even in some unrealistic and idealized thought experiment, if a conversation with an AI were completely indiscernible with perfect discernment from a conversation with a human being, it would nonetheless lack a causal account of what was observed. You would have only shown a perfect correlation, at best.

zahlman 1y ago

My experience has been more that the pro-AI people misunderstand where the goalposts were, and then complain when they're correctly pointed at. The "Turing test" I always saw described in literature, and the examples of what passing output from a machine was imagined to look like, are nothing like what's claimed to pass nowadays. Honestly, a lot of the people claiming that contemporary chatbots pass come across like they would have thought ELIZA passed.

bonoboTP 1y ago
Can you be more concrete? What kind of answer/conversation do you see as demonstrating passing the test, that you think is currently not possible.
- tsimionescu 1y ago
  Ones in which both the human test takers and the human counterparts are actively trying to prove to each other that they are actually human. With today's chat bots, it's absolutely trivial to tell that you're not talking to a real human. They will never interrupt you, continue their train of thought even thought you're trying to change the conversation, go on a complete non-sequitur, swear at you, etc. These are all things that the human "controls" should be doing to prove to the judges that they are indeed human. LLMs are nowhere near beating the Turing test. They may fool some humans in some limited interactions, especially if the output is curated by a human. But left alone to interact with the raw output for more than a few lines, and if actively seeking to tell if you're interacting with a human or an AI (instead of wanting to believe), there really is no chance you'd be tricked.
  - bonoboTP 1y ago
    Okay but we are not really optimizing them to emulate humans right now. In fact, it's the opposite. The mainstream bots are explicitly trained to not identify as humans and to refuse to claim having thought or internal feelings or consciousness. So in that sense it's a triviality. You can ask ChatGPT whether it's human and it will say no upfront. And it has various guardrails in place against too much "roleplay", so you can't just instruct it to act human. You'd need a different post-training setup. I'm not aware whether anyone did that with open models already.