5 ms·

Humans do a lot of things that computers don't, such as be born, age (verb), die, get hungry, fall in love, reproduce, and more. Computers can only metaphorical

Humans do a lot of things that computers don't, such as be born, age (verb), die, get hungry, fall in love, reproduce, and more. Computers can only metaphorically do these things, human learning is correlated with all of them, and we don't confidently know how. Have some humility.

onlyrealcuzzo 1y ago
The point is that if it's mathematically possible for humans, than it naively would be possible for computers. All of that just sounds hard, not mathematically impossible. As I understand it, this is mostly a rehash on the dated Lucas Penrose argument, which most Mind Theory researches refute.
ninetyninenine 1y ago
We don’t even know how LLMs work. But we do know the underlying mechanisms are governed by math because we have a theory of reality that governs things down to the atomic scale and humans and LLMs are made out of atoms. So because of this we know reality is governed by maths. We just can’t fully model the high level consequence of emergent patterns due to the sheer complexity of trillions of interacting atoms. So it’s not that there’s some mysterious supernatural thing we don’t understand. It’s purely a complexity problem in that we only don’t understand it because it’s too complex. What does humility have to do with anything?
- bigyabai 1y ago
  > We don’t even know how LLMs work Speak for yourself. LLMs are a feedforward algorithm inferring static weights to create a tokenized response string. We can compare that pretty trivially to the dynamic relationship of neurons and synapses in the human brain. It's not similar, case closed. That's the extent of serious discussion that can be had comparing LLMs to human thought, with apologies to Chomsky et. al. It's like trying to find the anatomical differences between a medieval scribe and a fax machine.
  - hnfong 1y ago
    Pretty sure in most other contexts you wouldn't agree a medieval scribe knows how a fax machine works.
    - ninetyninenine 1y ago
      Analogies aren’t proof. Like if an analogy doesn’t apply in certain context it is not a reflection of the actual situation. It just means the analogy is bad and irrelevant. Often people who don’t know how to be logical end up using analogies as proof. And you can simply say that the analogy doesn’t apply and is inaccurate and the whole argument becomes garbage because analogies aren’t logical basis for anything. Analogies are communication cools to facilitate easier understanding they are not proofs or evidence of anything.
  - ben_w 1y ago
    > Speak for yourself. LLMs are a feedforward algorithm inferring static weights to create a tokenized response string. If we're OK with descriptions so lossy that they fit in a sentence, we also understand the human brain: A electrochemical network with external inputs and some feedback loops, pumping ions around to trigger voltage cascades to create muscle contractions as outputs.
    - bigyabai 1y ago
      Yes. As long as we're confident in our definitions, that makes the questions easy. Is that the same as a feedforward algorithm inferring static weights to create a tokenized response string? Do you necessarily need an electrochemical network with external stimuli and feedback to generate legible text? No. The answer is already solved; AI is not a brain, we can prove this by characteristically defining them both and using heuristic reasoning.
      - ben_w 1y ago
        > The answer is already solved; AI is not a brain, we can prove this by characteristically defining them both and using heuristic reasoning. That "can" should be "could", else it presumes too much. For both human brains and surprisingly small ANNs, far smaller than LLMs, humanity collectively does not yet know the defining characteristics of the aspects we care about. I mean, humanity don't agree with itself what any of the three initials of AGI mean, there's 40 definitions of the word "consciousness", there are arguments about if there is either exactly one or many independent G-factors in human IQ scores, and also if those scores mean anything beyond correlating with school grades, and human nerodivergence covers various real states of existance that many of us find incomprehensible (sonetimes mutually, see e.g. most discussions where aphantasia comes up). The main reason I expect little from an AI is that we don't know what we're doing. The main reason I can't just assume the least is because neither did evolution when we popped out.
- IAmGraydon 1y ago
  >We don’t even know how LLMs work. Care to elaborate? Because that is utter nonsense.
  - Workaccount2 1y ago
    We understand and build the trellis that the LLMs "grow" on. We don't have good insight into how a fully grown LLM actually turns any specific input into any specific output. We can follow it through the network, but it's a totally senseless noisy mess. "Cat" lights up a certain set of neurons, but then "cat" looks completely different. That is what we don't really understand. (This is an illustrative example made for easy understanding, not something I specifically went and compared)
    - EPWN3D 1y ago
      We don't know the path for how a given input produces a given output, but that doesn't mean we don't know how LLMs work. We don't and can't know with certainty which specific atoms will fission in a nuclear reactor either. But we know how nuclear fission works.
      - ben_w 1y ago
        We have the Navier–Stokes equations which fit on a matchbox, yet for the last 25 years there's been a US$1,000,000 prize on offer to the first person providing a solution for a specific statement of the problem: Prove or give a counter-example of the following statement: In three space dimensions and time, given an initial velocity field, there exists a vector velocity and a scalar pressure field, which are both smooth and globally defined, that solve the Navier–Stokes equations.
        bigyabai 1y ago
        And when that prize is claimed, we'll ring the bell on AGI being found. Gentleman's agreement.
        1y ago
- hnfong 1y ago
  > we have a theory of reality that governs things down to the atomic scale and humans and LLMs are made out of atoms. > So because of this we know reality is governed by maths. That's not really true. You have a theory, and let's presume so far it's consistent with observations. But it doesn't mean it's 100% correct, and doesn't mean at some point in the future you won't observe something that invalidates the theory. In short, you don't know whether the theory is absolutely true and you can never know. Without an absolutely true theory, all you have is belief or speculation that reality is governed by maths. > What does humility have to do with anything? Not the GP but I think humility is kinda relevant here.
  - ninetyninenine 1y ago
    >That's not really true. You have a theory, and let's presume so far it's consistent with observations. But it doesn't mean it's 100% correct, and doesn't mean at some point in the future you won't observe something that invalidates the theory. In short, you don't know whether the theory is absolutely true and you can never know. Let me repharse it. As far as we know all of reality is governed by the principles of logic and therefore math. This is the most likely possibility and we have based all of our technology and culture and science around this. It is the fundamental assumption humanity has made on reality. We cannot consistently demonstrate disproof against this assumption. >Not the GP but I think humility is kinda relevant here. How so? If I assume all of reality is governed by math, but you don't. How does that make me not humble but you humble? Seems personal.
    - hnfong 1y ago
      I guess it's kinda hubris on my part to question your ability to know things with such high certainty about things that philosophers have been struggling to prove for millenia... What you said is only true for the bits of humanity you have decided to focus upon -- capitalist, technology-driven modern societies. If you looked beyond that, there are cultures that build society upon other assumptions. You might think those other modes are "wrong", but that's your personal view. For me, I personally don't think any of these are "true" in the absolute sense, as much as I don't think yours is "true". They're just ways humans with our mortal brains try to grapple with a reality that we don't understand. As a sidenote, probability does not mean the thing you think it means. There's no reasonable frequentist interpretation for fundamental truth of reality, so you're just saying your Bayesian subjective probability says that math is "the most likely possibility". Which is fine, except everyone has their own different priors...
andyjohnson0 1y ago
TFA presents an information-theoretic argument forAGI being impossible. My reading of your parent commenter is that they are asking why this argument does not also apply to humans. You make broadly valid points, particularly about the advantages of embodyment, but I just dont think theyre good responses to the theoretical article under discussion (or the comment that you were responding to).
daedrdev 1y ago
Taking GLP-1 makes me question how much hunger is really me versus my hormones controlling me.

onlyrealcuzzo 1y ago

The point is that if it's mathematically possible for humans, than it naively would be possible for computers. All of that just sounds hard, not mathematically impossible. As I understand it, this is mostly a rehash on the dated Lucas Penrose argument, which most Mind Theory researches refute.

ninetyninenine 1y ago

We don’t even know how LLMs work. But we do know the underlying mechanisms are governed by math because we have a theory of reality that governs things down to the atomic scale and humans and LLMs are made out of atoms. So because of this we know reality is governed by maths. We just can’t fully model the high level consequence of emergent patterns due to the sheer complexity of trillions of interacting atoms. So it’s not that there’s some mysterious supernatural thing we don’t understand. It’s purely a complexity problem in that we only don’t understand it because it’s too complex. What does humility have to do with anything?

bigyabai 1y ago
> We don’t even know how LLMs work Speak for yourself. LLMs are a feedforward algorithm inferring static weights to create a tokenized response string. We can compare that pretty trivially to the dynamic relationship of neurons and synapses in the human brain. It's not similar, case closed. That's the extent of serious discussion that can be had comparing LLMs to human thought, with apologies to Chomsky et. al. It's like trying to find the anatomical differences between a medieval scribe and a fax machine.
- hnfong 1y ago
  Pretty sure in most other contexts you wouldn't agree a medieval scribe knows how a fax machine works.
  - ninetyninenine 1y ago
    Analogies aren’t proof. Like if an analogy doesn’t apply in certain context it is not a reflection of the actual situation. It just means the analogy is bad and irrelevant. Often people who don’t know how to be logical end up using analogies as proof. And you can simply say that the analogy doesn’t apply and is inaccurate and the whole argument becomes garbage because analogies aren’t logical basis for anything. Analogies are communication cools to facilitate easier understanding they are not proofs or evidence of anything.
- ben_w 1y ago
  > Speak for yourself. LLMs are a feedforward algorithm inferring static weights to create a tokenized response string. If we're OK with descriptions so lossy that they fit in a sentence, we also understand the human brain: A electrochemical network with external inputs and some feedback loops, pumping ions around to trigger voltage cascades to create muscle contractions as outputs.
  - bigyabai 1y ago
    Yes. As long as we're confident in our definitions, that makes the questions easy. Is that the same as a feedforward algorithm inferring static weights to create a tokenized response string? Do you necessarily need an electrochemical network with external stimuli and feedback to generate legible text? No. The answer is already solved; AI is not a brain, we can prove this by characteristically defining them both and using heuristic reasoning.
    - ben_w 1y ago
      > The answer is already solved; AI is not a brain, we can prove this by characteristically defining them both and using heuristic reasoning. That "can" should be "could", else it presumes too much. For both human brains and surprisingly small ANNs, far smaller than LLMs, humanity collectively does not yet know the defining characteristics of the aspects we care about. I mean, humanity don't agree with itself what any of the three initials of AGI mean, there's 40 definitions of the word "consciousness", there are arguments about if there is either exactly one or many independent G-factors in human IQ scores, and also if those scores mean anything beyond correlating with school grades, and human nerodivergence covers various real states of existance that many of us find incomprehensible (sonetimes mutually, see e.g. most discussions where aphantasia comes up). The main reason I expect little from an AI is that we don't know what we're doing. The main reason I can't just assume the least is because neither did evolution when we popped out.
IAmGraydon 1y ago
>We don’t even know how LLMs work. Care to elaborate? Because that is utter nonsense.
- Workaccount2 1y ago
  We understand and build the trellis that the LLMs "grow" on. We don't have good insight into how a fully grown LLM actually turns any specific input into any specific output. We can follow it through the network, but it's a totally senseless noisy mess. "Cat" lights up a certain set of neurons, but then "cat" looks completely different. That is what we don't really understand. (This is an illustrative example made for easy understanding, not something I specifically went and compared)
  - EPWN3D 1y ago
    We don't know the path for how a given input produces a given output, but that doesn't mean we don't know how LLMs work. We don't and can't know with certainty which specific atoms will fission in a nuclear reactor either. But we know how nuclear fission works.
    - ben_w 1y ago
      We have the Navier–Stokes equations which fit on a matchbox, yet for the last 25 years there's been a US$1,000,000 prize on offer to the first person providing a solution for a specific statement of the problem: Prove or give a counter-example of the following statement: In three space dimensions and time, given an initial velocity field, there exists a vector velocity and a scalar pressure field, which are both smooth and globally defined, that solve the Navier–Stokes equations.
      - bigyabai 1y ago
        And when that prize is claimed, we'll ring the bell on AGI being found. Gentleman's agreement.
        1y ago
hnfong 1y ago
> we have a theory of reality that governs things down to the atomic scale and humans and LLMs are made out of atoms. > So because of this we know reality is governed by maths. That's not really true. You have a theory, and let's presume so far it's consistent with observations. But it doesn't mean it's 100% correct, and doesn't mean at some point in the future you won't observe something that invalidates the theory. In short, you don't know whether the theory is absolutely true and you can never know. Without an absolutely true theory, all you have is belief or speculation that reality is governed by maths. > What does humility have to do with anything? Not the GP but I think humility is kinda relevant here.
- ninetyninenine 1y ago
  >That's not really true. You have a theory, and let's presume so far it's consistent with observations. But it doesn't mean it's 100% correct, and doesn't mean at some point in the future you won't observe something that invalidates the theory. In short, you don't know whether the theory is absolutely true and you can never know. Let me repharse it. As far as we know all of reality is governed by the principles of logic and therefore math. This is the most likely possibility and we have based all of our technology and culture and science around this. It is the fundamental assumption humanity has made on reality. We cannot consistently demonstrate disproof against this assumption. >Not the GP but I think humility is kinda relevant here. How so? If I assume all of reality is governed by math, but you don't. How does that make me not humble but you humble? Seems personal.
  - hnfong 1y ago
    I guess it's kinda hubris on my part to question your ability to know things with such high certainty about things that philosophers have been struggling to prove for millenia... What you said is only true for the bits of humanity you have decided to focus upon -- capitalist, technology-driven modern societies. If you looked beyond that, there are cultures that build society upon other assumptions. You might think those other modes are "wrong", but that's your personal view. For me, I personally don't think any of these are "true" in the absolute sense, as much as I don't think yours is "true". They're just ways humans with our mortal brains try to grapple with a reality that we don't understand. As a sidenote, probability does not mean the thing you think it means. There's no reasonable frequentist interpretation for fundamental truth of reality, so you're just saying your Bayesian subjective probability says that math is "the most likely possibility". Which is fine, except everyone has their own different priors...

andyjohnson0 1y ago

TFA presents an information-theoretic argument forAGI being impossible. My reading of your parent commenter is that they are asking why this argument does not also apply to humans. You make broadly valid points, particularly about the advantages of embodyment, but I just dont think theyre good responses to the theoretical article under discussion (or the comment that you were responding to).

daedrdev 1y ago

Taking GLP-1 makes me question how much hunger is really me versus my hormones controlling me.