5 ms·

(To those with predictable comments) I feel like the important question is when will you accept an ai is conscious? If not when it says so and not when it seems

(To those with predictable comments) I feel like the important question is when will you accept an ai is conscious? If not when it says so and not when it seems to display emotions (and change behavior based on them), when? I have a feeling that there won't be a point of satisfaction. The familiar will always trail behind the new, and the argument "it's just a fancier version of the last ai iteration " will never die, at least to some people. I feel like maybe some introspection is required

ggm 3y ago
Your question here is entirely reasonable. It's a meta question. here is my meta answer: The underlying assumption is not valid at this time. Thats why I cannot accept the question and answer as stated. It's akin to "why do you refuse to believe in god" for athiests, its a term which denies the axioms. Because I do not believe in connectionist models of intelligence, nor in the stats/chaining model here, I see no reason to give credence to any AI self assertion because it's simply one of the statistically valid responses to the priming question. I'd be more interested in the "no I am not alive because.. responses" and the ratios to "yes I am because.." since thats informative of the sources. "Why is a fish not riding a bicycle" questions are rarely useful. "why do you refuse to accept this is emergent intelligence" questions are primed to ignore the terse response "because.. it isn't" Also, another kind of "predictable" response here, is that the response from a trained system to this question appears to be .. highly predictable. I will accept evidence which is bound in theories, of mind, of human and other biologically driven forms of intelligence, which is informative to emerging knowledge of intelligence in organisms, which helps define intelligence in a way which can be inclusive of machine intelligence. Simple assertion "I am" is not yet there for me. As I have said (or at least implied) above, it's one of the million monkey responses we're being trained to believe. Language, and the denotation of meaning matters. Claude 3 did not CLAIM anything because CLAIMING is an attribute of a thinking mind. It was an assertion from a complex software system. Explain to me how it differs from "parse error on line #3, quitting." beyond "that string appears literally in the sources"
- lucubratory 3y ago
  I'm not sure if you're contradicting yourself or if I just haven't understood what you've written, so I'd like to ask two clarifying questions. The first is: Are you claiming that any system built on this architecture (i.e. an LLM is a crucial part of what makes it behave intelligently) is definitionally not thinking, and this would be the case regardless of any evidence to the contrary? If that isn't what you are claiming, what evidence would convince you that an AI system which critically depended on LLMs to function is thinking?
  - ggm 3y ago
    I'd need a theory as to why intelligence is in an llm not just apparent stimulus response outcomes. There is a huge corpus of knowledge inside the model and a black box for most of us. It's generating syntactically correct phrases, reams of them but with significant errors which to me say there is no apparent understanding of the regurgitated facts, and I see no introspection or intent. Evidence is a more abstract concept. Nothing to date is evidence, to me. I guess it will be settled in hindsight and a theory to me is the huge missing link. We do (as far as I know) have a distinct lack of theories about intelligence at large, I find it encouraging how much recent research sees long term evidence of intelligence in higher organisms and better understanding for brain function but without being either a brain scientist or an fMRI expert I think I can say neither brain science nor fMRI science think they've cracked it yet, and so llm intelligence doesn't even begin to have theories grounded in organic, existent intelligence to work on just observations of input and output.
    - lucubratory 3y ago
      You might be interested in global workspace theory as a theory of consciousness. You can find more discussion of it, and other theories of consciousness as they relate to modern AI, in this very helpful paper: https://arxiv.org/abs/2308.08708 https://arxiv.org/abs/2308.08708
- creer 3y ago
  > It's akin to "why do you refuse to believe in god" Wow wow wow! Here you have the opportunity to design your own experiments; choose as specific (or varied) criteria as YOU wish; start from your own definitions or axioms, state them and test them. In fact you are even allowed to move the goal post if you prefer (though nobody will like you for that, it's still possible). The argument is more "state and test them!" The case could not be more different. Even if you wish to just stick to defining "conscious", people will welcome that. Isn't it "not valid at this time", that's not valid? I don't claim to have the answer to any of these (but I do claim to see where this is headed, and rather fast too.) But I find it hard to claim that it's not even a valid conversation.
otabdeveloper4 3y ago
Consciousness doesn't matter. My phone is conscious when its battery is low. What you really want is "free will", by which you really mean information complexity, especially of the recursive kind. ("Consciousness" here is only interesting because of self-reflection, which is a kind of recursion, I guess.) So to answer your question, no, LLMs don't increase information complexity in any way, so they aren't interesting as intelligence.
- LawrenceKerr 3y ago
  So would you posit an LLM which self-updates (retrains) its internal model after every prompt-response (perhaps at night, when it dreams of electric sheep) is conscious because it shows some recursive self-reflection & decision making? (And/or add in active inference as described by Karl Friston, whereby it tries to predict outcomes and self-correct its model) If you knew the LLM does this after every prompt, is that a sufficient condition for possibly concluding that it's conscious? And if this recursion is not enough, what is? What's the missing piece? >Consciousness doesn't matter. My phone is conscious when its battery is low. Those are interesting claims...
  - otabdeveloper4 3y ago
    Like I said, "consciousness" is a red herring and doesn't even really exist. What we care about is whether information complexity is increased. LLMs don't (and can't) do that by the very nature of their mathematics.
    - jzombie 3y ago
      > ...by the very nature of their mathematics What are their mathematics? Please enlighten me.
      - otabdeveloper4 3y ago
        Their mathematics is a statistical model which outputs the statistically most likely result, in effect leaving only the most significant independent variables. Which decreases information complexity, by definition of "information complexity".
Barrin92 3y ago
> If not when it says so and not when it seems to display emotions (and change behavior based on them), when? A necessary but not sufficient condition is having the architecture for consciousness, which means having the ability to actually experience and reflect on their own internal state, which these models by design cannot do. It tells you it has consciousness because that's part of its training, if the training data and tuning told it to say it has pain in its legs and feels sad about it, it would do literally that despite not having legs. The human brain has a limbic system that regulates its emotional state, and there's constant brain activity going on managing those. These models don't sit there and experience emotions they literally do nothing until you ask them to give you another token.
ars 3y ago
How do I know another person is conscious? For all I know I'm the main character and you are all NPC's. There's no actual way to know, we act that way because when we talk to others they have an internal state, and their programming is not alterable. There's also the argument that we act like everyone is conscious because of what would happen if we didn't. And regarding that altered programming, there are drugs people can take that let them speak and seem intelligent without actually being conscious. Specifically Ambien can make ordinary humans into basically AI's: despite seeming to be awake they are not actually conscious. And if you talk to someone on Ambien you can tell (if you are looking for it), so you can also tell if a computer is conscious. https://www.nytimes.com/2011/12/04/magazine/can-ambien-wake-minimally-conscious.html https://www.nytimes.com/2011/12/04/magazine/can-ambien-wake-...
GoblinSlayer 3y ago
That said, why ai isn't trained to do logic? Why not train it on all scientific, philosophic and math texts, or it doesn't result in anything working?
- creer 3y ago
  You'll find some comparisons in reasoning skill under this comment https://news.ycombinator.com/item?id=39594562 https://news.ycombinator.com/item?id=39594562 There are also some papers using more complex architecture including LLMs that are aimed at better multi-step reasoning. "Self looping" kinds of answer computing.
grantcas 3y ago
[dead]
satvikpendem 3y ago
> (To those with predictable comments) I feel like the important question is when will you accept an ai is conscious? If not when it says so and not when it seems to display emotions (and change behavior based on them), when? Never, but that is because this is a philosophical question dating back millennia, it is nothing new [0]. [0] https://en.wikipedia.org/wiki/Solipsism https://en.wikipedia.org/wiki/Solipsism

ggm 3y ago

Your question here is entirely reasonable. It's a meta question. here is my meta answer: The underlying assumption is not valid at this time. Thats why I cannot accept the question and answer as stated. It's akin to "why do you refuse to believe in god" for athiests, its a term which denies the axioms. Because I do not believe in connectionist models of intelligence, nor in the stats/chaining model here, I see no reason to give credence to any AI self assertion because it's simply one of the statistically valid responses to the priming question. I'd be more interested in the "no I am not alive because.. responses" and the ratios to "yes I am because.." since thats informative of the sources. "Why is a fish not riding a bicycle" questions are rarely useful. "why do you refuse to accept this is emergent intelligence" questions are primed to ignore the terse response "because.. it isn't" Also, another kind of "predictable" response here, is that the response from a trained system to this question appears to be .. highly predictable. I will accept evidence which is bound in theories, of mind, of human and other biologically driven forms of intelligence, which is informative to emerging knowledge of intelligence in organisms, which helps define intelligence in a way which can be inclusive of machine intelligence. Simple assertion "I am" is not yet there for me. As I have said (or at least implied) above, it's one of the million monkey responses we're being trained to believe. Language, and the denotation of meaning matters. Claude 3 did not CLAIM anything because CLAIMING is an attribute of a thinking mind. It was an assertion from a complex software system. Explain to me how it differs from "parse error on line #3, quitting." beyond "that string appears literally in the sources"

lucubratory 3y ago
I'm not sure if you're contradicting yourself or if I just haven't understood what you've written, so I'd like to ask two clarifying questions. The first is: Are you claiming that any system built on this architecture (i.e. an LLM is a crucial part of what makes it behave intelligently) is definitionally not thinking, and this would be the case regardless of any evidence to the contrary? If that isn't what you are claiming, what evidence would convince you that an AI system which critically depended on LLMs to function is thinking?
- ggm 3y ago
  I'd need a theory as to why intelligence is in an llm not just apparent stimulus response outcomes. There is a huge corpus of knowledge inside the model and a black box for most of us. It's generating syntactically correct phrases, reams of them but with significant errors which to me say there is no apparent understanding of the regurgitated facts, and I see no introspection or intent. Evidence is a more abstract concept. Nothing to date is evidence, to me. I guess it will be settled in hindsight and a theory to me is the huge missing link. We do (as far as I know) have a distinct lack of theories about intelligence at large, I find it encouraging how much recent research sees long term evidence of intelligence in higher organisms and better understanding for brain function but without being either a brain scientist or an fMRI expert I think I can say neither brain science nor fMRI science think they've cracked it yet, and so llm intelligence doesn't even begin to have theories grounded in organic, existent intelligence to work on just observations of input and output.
  - lucubratory 3y ago
    You might be interested in global workspace theory as a theory of consciousness. You can find more discussion of it, and other theories of consciousness as they relate to modern AI, in this very helpful paper: https://arxiv.org/abs/2308.08708 https://arxiv.org/abs/2308.08708
creer 3y ago
> It's akin to "why do you refuse to believe in god" Wow wow wow! Here you have the opportunity to design your own experiments; choose as specific (or varied) criteria as YOU wish; start from your own definitions or axioms, state them and test them. In fact you are even allowed to move the goal post if you prefer (though nobody will like you for that, it's still possible). The argument is more "state and test them!" The case could not be more different. Even if you wish to just stick to defining "conscious", people will welcome that. Isn't it "not valid at this time", that's not valid? I don't claim to have the answer to any of these (but I do claim to see where this is headed, and rather fast too.) But I find it hard to claim that it's not even a valid conversation.

otabdeveloper4 3y ago

Consciousness doesn't matter. My phone is conscious when its battery is low. What you really want is "free will", by which you really mean information complexity, especially of the recursive kind. ("Consciousness" here is only interesting because of self-reflection, which is a kind of recursion, I guess.) So to answer your question, no, LLMs don't increase information complexity in any way, so they aren't interesting as intelligence.

LawrenceKerr 3y ago
So would you posit an LLM which self-updates (retrains) its internal model after every prompt-response (perhaps at night, when it dreams of electric sheep) is conscious because it shows some recursive self-reflection & decision making? (And/or add in active inference as described by Karl Friston, whereby it tries to predict outcomes and self-correct its model) If you knew the LLM does this after every prompt, is that a sufficient condition for possibly concluding that it's conscious? And if this recursion is not enough, what is? What's the missing piece? >Consciousness doesn't matter. My phone is conscious when its battery is low. Those are interesting claims...
- otabdeveloper4 3y ago
  Like I said, "consciousness" is a red herring and doesn't even really exist. What we care about is whether information complexity is increased. LLMs don't (and can't) do that by the very nature of their mathematics.
  - jzombie 3y ago
    > ...by the very nature of their mathematics What are their mathematics? Please enlighten me.
    - otabdeveloper4 3y ago
      Their mathematics is a statistical model which outputs the statistically most likely result, in effect leaving only the most significant independent variables. Which decreases information complexity, by definition of "information complexity".

Barrin92 3y ago

> If not when it says so and not when it seems to display emotions (and change behavior based on them), when? A necessary but not sufficient condition is having the architecture for consciousness, which means having the ability to actually experience and reflect on their own internal state, which these models by design cannot do. It tells you it has consciousness because that's part of its training, if the training data and tuning told it to say it has pain in its legs and feels sad about it, it would do literally that despite not having legs. The human brain has a limbic system that regulates its emotional state, and there's constant brain activity going on managing those. These models don't sit there and experience emotions they literally do nothing until you ask them to give you another token.

ars 3y ago

How do I know another person is conscious? For all I know I'm the main character and you are all NPC's. There's no actual way to know, we act that way because when we talk to others they have an internal state, and their programming is not alterable. There's also the argument that we act like everyone is conscious because of what would happen if we didn't. And regarding that altered programming, there are drugs people can take that let them speak and seem intelligent without actually being conscious. Specifically Ambien can make ordinary humans into basically AI's: despite seeming to be awake they are not actually conscious. And if you talk to someone on Ambien you can tell (if you are looking for it), so you can also tell if a computer is conscious. https://www.nytimes.com/2011/12/04/magazine/can-ambien-wake-minimally-conscious.html https://www.nytimes.com/2011/12/04/magazine/can-ambien-wake-...

GoblinSlayer 3y ago

That said, why ai isn't trained to do logic? Why not train it on all scientific, philosophic and math texts, or it doesn't result in anything working?

creer 3y ago
You'll find some comparisons in reasoning skill under this comment https://news.ycombinator.com/item?id=39594562 https://news.ycombinator.com/item?id=39594562 There are also some papers using more complex architecture including LLMs that are aimed at better multi-step reasoning. "Self looping" kinds of answer computing.

grantcas 3y ago

[dead]

satvikpendem 3y ago

> (To those with predictable comments) I feel like the important question is when will you accept an ai is conscious? If not when it says so and not when it seems to display emotions (and change behavior based on them), when? Never, but that is because this is a philosophical question dating back millennia, it is nothing new [0]. [0] https://en.wikipedia.org/wiki/Solipsism https://en.wikipedia.org/wiki/Solipsism