Secondary school maths showing that AI systems don't think

(raspberrypi.org)

28 points | by zdw 2 hours ago

10 comments

  • causal 1 hour ago
    I love the idea of educating students on the math behind AI to demystify them. But I think it's a little weird to assert "AI is not magic and AI systems do not think. It’s just maths." Equivalent statements could be made about how human brains are not magic, just biology - yet I think we still think.
    • qsort 37 minutes ago
      I agree saying "they don't think" and leaving it at that isn't particularly useful or insightful, it's like saying "submarines don't swim" and refusing to elaborate further. It can be useful if you extend it to "they don't think like you do". Concepts like finite context windows, or the fact that the model is "frozen" and stateless, or the idea that you can transfer conversations between models are trivial if you know a bit about how LLMs work, but extremely baffling otherwise.
    • ux266478 42 minutes ago
      It's just provencial nonsense, there's no sound reasoning to it. Reductionism being taken and used as a form of refutation is a pretty common cargo culting behavior I've found.

      Overwhelmingly, I just don't think the majority of human beings have the mental toolset to work with ambiguous philosophical contexts. They'll still try though, and what you get out of that is a 4th order baudrillardian simulation of reason.

    • jvanderbot 50 minutes ago
      Thinking is undefined so all statements about it are unverifiable.
      • terminalshort 2 minutes ago
        Statements like "it is bound by the laws of physics" are not "verifiable" by your definition, and yet we safely assume it is true of everything. Everything except the human brain, that is, for which wild speculation that it may be supernatural is seemingly considered rational discussion so long as it satisfies people's needs to believe that they are somehow special in the universe.
        • jvanderbot 0 minutes ago
          True. You need to define "it" before you can verify physics bounds it.

          Unicorns are not bound by the laws of physics - because they do not exist.

      • d-lisp 31 minutes ago
        Do you think that thinking is undefinable ? If thinking is definable, then all statements about it aren't unverifiable.
        • ablob 16 minutes ago
          Caveat: if thinking is definable, then not all statements about it are unverifiable.
      • nh23423fefe 39 minutes ago
        Is this some self refuting sentence?
        • d-lisp 27 minutes ago
          I think they meant "Cannot evaluate : (is <undefined> like x ?), argument missing"

          edit : Thinking is undefined, statements about undefined cannot be verified.

        • ux266478 25 minutes ago
          is a meta-level grammar the same as an object-level grammar?
      • random9749832 35 minutes ago
        Is reasoning undefined? That's what usually meant by "thinking".
        • nutjob2 30 minutes ago
          Formal reasoning is defined, informal reasoning very much isn't.
    • terminalshort 7 minutes ago
      I have yet to hear any plausible definition of "thought" that convincingly places LLMs and brains on opposite sides of it without being obviously contrived for that purpose.
    • snickerbockers 33 minutes ago
      >Equivalent statements could be made about how human brains are not magic, just biology - yet I think we still think.

      They're not equivalent at all because the AI is by no means biological. "It's just maths" could maybe be applied to humans but this is backed entirely by supposition and would ultimately just be an assumption of its own conclusion - that human brains work on the same underlying principles as AI because it is assumed that they're based on the same underlying principles as AI.

      • pegasus 11 minutes ago
        But parent didn't try to apply "it's just maths" to humans. He said one could just as easily say, as some do: "Humans are just biology, hence they're not magic". Our understanding of mathematics, including the maths of transformer models is limited, just as our understanding of biology. Some behaviors of these models have taken researches by surprise, and future surprises are not at all excluded. We don't know exactly how far they will evolve.

        As for applying the word thinking to AI systems, it's already in common usage and this won't change. We don't have any other candidate words, and this one is the closest existing word for referencing a computational process which, one must admit, is in many ways (but definitely not in all ways) analogous to human thought.

      • AlecSchueler 20 minutes ago
        > that human brains work on the same underlying principles as AI

        That wasn't the assumption though, it was only that human brains work by some "non-magical" electro-chemical process which could be described as a mechanism, whether that mechanism followed the same principles of AI or not.

      • hnfong 16 minutes ago
        Well, a better retort would be "Human brains are not magic, just physics. Protons, neutrons and electrons don't think".

        But I think most people get what GP means.

      • ikrenji 19 minutes ago
        Human brains might not be explained by the same type of math AI is explained with, but it will be some kind of math...
      • mcswell 12 minutes ago
        Straw man. The person who you're responding to talked about "equivalent statements" (emphasis added), whereas you appear to be talking about equivalent objects (AIs vs. brains), and pointing out the obvious flaw in this argument, that AIs aren't biology. The obvious flaw in the wrong argument, that is.
    • nutjob2 26 minutes ago
      AI systems compute and humans think. One is math and the other biology.

      But they are two different things with overlapping qualities.

      It's like MDMA and falling in love. They have many overlapping quantities but no one would claim one is the other.

    • CamperBob2 29 minutes ago
      That's where these threads always end up. Someone asserts, almost violently, that AI does not and/or cannot "think." When asked how to falsify their assertion, perhaps by explaining what exactly is unique about the human brain that cannot and/or will not be possible to emulate, that's the last anyone ever hears from them. At least until the next "AI can't think" story gets posted.

      The same arguments that appeared in 2015 inevitably get trotted out, almost verbatim, ten years later. It would be amusing on other sites, but it's just pathetic here.

      • Terr_ 25 minutes ago
        Someone asserts, almost religiously, that LLMs do and/or can "think." When asked how to falsify their assertion, perhaps by explaining what exactly is "thinking" in the human brain that can and/or will be possible to emulate...
        • gfdvgfffv 12 minutes ago
          One mostly sees people aggressively claiming they can’t, ever. On the other side people seem to simply allow that they might, or might eventually.
        • superkuh 10 minutes ago
          Or they just point to the turing test which was the defacto standard test for something so nebulous. And behold: LLM can pass the turing test. So they think. Can you come up with something better (than the turing test)?
        • umanwizard 15 minutes ago
          Err, no, that’s not what’s happening. Nobody, at least in this thread (and most others like it I’ve seen), is confidently claiming LLMs can think.

          There are people confidently claiming they can’t and then other people expressing skepticism at their confidence and/or trying to get them to nail down what they mean.

        • CamperBob2 18 minutes ago
          When asked how to falsify their assertion, perhaps by explaining what exactly is "thinking" in the human brain that can and/or will be possible to emulate...

          ... someone else points out that the same models that can't "think" are somehow turning in gold-level performance at international math and programming competitions, making Fields Medalists sit up and take notice, winning art competitions, composing music indistinguishable from human output, and making entire subreddits fail the Turing test.

          • Terr_ 17 minutes ago
            A couple decades of chess programs nods knowingly: "First time?"
            • CamperBob2 10 minutes ago
              A couple decades of chess programs nods knowingly: "First time?"

              Uh huh. Good luck getting Stockfish to do your math homework while Leela works on your next waifu.

              LLMs play chess poorly. Chess engines do nothing else at all. That's kind of a big difference, wouldn't you say?

          • nh23423fefe 10 minutes ago
            god of the gaps
            • CamperBob2 8 minutes ago
              Exactly. As soon as a model does something it "wasn't supposed to be able to do," two gaps open up on either side.
          • nutjob2 8 minutes ago
            Computers can perform math and numerous other tasks billions of times faster than humans, whats your point?

            This is exactly the problem. Claims about AI are unfalsifiable, thus your various non-sequiturs about AI 'thinking'.

  • WhyOhWhyQ 6 minutes ago
    A lot of the drama here is due to the ambiguity of what the word 'think' is supposed to mean. One camp associates 'thinking' to consciousness, another does not. I personally believe it is possible to create an animal-like or human-like intelligence, without consciousness existing in the system. I personally would still describe whatever processing that system is doing as 'thinking'. Others believe in "substrate independence"; they think any such system must be consciousness.

    (Sneaking a bit of belief in here, to me "substrate independence" is a more extreme position than the idea that a system could be made which is intelligent but not conscious, hence I find it implausible.)

  • frozenlettuce 53 minutes ago
    You can replicate all calculations done by LLMs with pen and paper. It would take ages to calculate anything, but it's possible. I don't think that pen and paper will ever "think", regardless of how complex the calculations involved are.
    • mcswell 7 minutes ago
      I don't see the relevance of that argument (which other responders to your post have pointed out as Searle's Chinese Room argument). The pen and paper are of course not doing any thinking, but then the pen isn't doing any writing on its own, either. It's the system of pen + paper + human that's doing the thinking.
    • gus_massa 30 minutes ago
      The official name is https://en.wikipedia.org/wiki/Chinese_room

      The opinions are exactly the same than about LLM.

      • sigmoid10 8 minutes ago
        And the counter argument is also exactly the same. Imagine you take one neuron from a brain and replace it with an artificial piece of electronics (e.g. some transistors) that only generates specific outputs based on inputs, exactly like the neuron does. Now replace another neuron. And another. Eventually, you will have the entire brain replaced with a huge set of fundamentally super simple transistors. I.e. a computer. If you believe that consciousness or the ability to think disappears somewhere during this process, then you are essentially believing in some religious meta-physics or soul-like component in our brains that can not be measured. But if it can not be measured, it fundamentally can not affect you in any way. So it doesn't matter for the experiment in the end, because the outcome would be exactly the same. The only reason you might think that you are conscious and the computer is not is because you believe so. But to an outsider observer, belief is all it is. Basically religion.
    • Wowfunhappy 11 minutes ago
      https://xkcd.com/505/

      You can replicate the entire universe with pen and paper (or a bunch of rocks). It would take an unimaginably long time, and we haven't discovered all the calculations you'd need to do yet, but presumably they exist and this could be done.

      Does that actually make a universe? I don't know!

      The comic is meant to be a joke, I think, but I find myself thinking about it all the time!!!

    • thrance 30 minutes ago
      You're arguing against Functionalism [0], of which I'd encourage you to at least read the Wikipedia page. Why would doing the brain's computations on pen and paper rather than on wetware lead to different outcomes? And how?

      Connect your pen and paper operator to a brainless human body, and you got something indistinguishable from a regular alive human.

      [0] https://en.wikipedia.org/wiki/Functionalism_%28philosophy_of...

    • umanwizard 50 minutes ago
      You can simulate a human brain on pen and paper too.
      • palmotea 35 minutes ago
        > You can simulate a human brain on pen and paper too.

        That's an assumption, though. A plausible assumption, but still an assumption.

        We know you can execute an LLM on pen and paper, because people built them and they're understood well enough that we could list the calculations you'd need to do. We don't know enough about the human brain to create a similar list, so I don't think you can reasonably make a stronger statement than "you could probably simulate..." without getting ahead of yourself.

        • hnfong 10 minutes ago
          This is basically the Church-Turing thesis and one of the motivations of using tape(paper) and an arbitrary alphabet in the Turing machine model.

          It's been kinda discussed to oblivion in the last century, interesting that it seems people don't realize the "existing literature" and repeat the same arguments (not saying anyone is wrong).

        • terminalshort 10 minutes ago
          I can make a claim much stronger than "you could probably" The counterclaim here is that the brain may not obey physical laws that can be described by mathematics. This is a "5G causes covid" level claim. The overwhelming burden of proof is on you.
      • phantasmish 41 minutes ago
        The simulation isn't an operating brain. It's a description of one. What it "means" is imposed by us, what it actually is, is a shitload of graphite marks on paper or relays flipping around or rocks on sand or (pick your medium).

        An arbitrarily-perfect simulation of a burning candle will never, ever melt wax.

        An LLM is always a description. An LLM operating on a computer is identical to a description of it operating on paper (if much faster).

        • gnull 25 minutes ago
          What makes the simulation we live in special compared to the simulation of a burning candle that you or I might be running?

          That simulated candle is perfectly melting wax in its own simulation. Duh, it won't melt any in ours, because our arbitrary notions of "real" wax are disconnected between the two simulatons.

          • hnfong 7 minutes ago
            They do have a valid subtle point though.

            If we don't think the candle in a simulated universe is a "real candle", why do we consider the intelligence in a simulated universe possibly "real intelligence"?

            Being a functionalist ( https://en.wikipedia.org/wiki/Functionalism_(philosophy_of_m... ) myself, I don't know the answer on the top of my head.

      • pton_xd 36 minutes ago
        So the brain is a mathematical artifact that operates independently from time? It just happens to be implemented using physics? Somehow I doubt it.
        • thrance 28 minutes ago
          The brain follows the laws of physics. The laws of physics can be closely approximated by mathematical models. Thus, the brain can be closely approximated by mathematical models.
      • an0malous 37 minutes ago
        Parent said replicate, as in deterministically
      • andrepd 45 minutes ago
        It's an open problem whether you can or not.
        • space_fountain 41 minutes ago
          It’s not that open. We can simulate smaller system of neurons just fine, we can simulate chemistry. There might be something beyond that in our brains for some reason, but it sees doubtful right now
          • phantasmish 38 minutes ago
            Our brains actually do something, may be the difference. They're a thing happening, not a description of a thing happening.

            Whatever that something that it actually does in the real, physical world is produces the cogito in cogito, ergo sum and I doubt you can get it just by describing what all the subatomic particles are doing, any more than a computer or pen-and-paper simulated hurricane can knock your house down, no matter how perfectly simulated.

            • ehsanu1 9 minutes ago
              Doing something merely requires I/O. Brains wouldn't be doing much without that. A sufficiently accurate simulation of a fundamentally computational process is really just the same process.
            • terminalshort 14 minutes ago
              Why are the electric currents moving in a GPU any less of a "thing happening" than the firing of the neurons in your brain? What you are describing here is a claim that the brain is fundamentally supernatural.
            • thrance 25 minutes ago
              You're arguing for the existence of a soul, for dualism. Nothing wrong with that, except we have never been able to measure it, and have never had to use it to explain any phenomenon of the brain's working. The brain follows the rules of physics, like any other objects of the material world.

              A pen and paper simulation of a brain would also be "a thing happening" as you put it. You have to explain what is the magical ingredient that makes the brain's computations impossible to replicate.

              You could connect your brain simulation to an actual body, and you'd be unable to tell the difference with a regular human, unless you crack it open.

              • phantasmish 18 minutes ago
                > You're arguing for the existence of a soul, for dualism.

                I'm not. You might want me to be, but I'm very, very much not.

  • TallGuyShort 14 minutes ago
    It's unfortunate that there's so little (none in the article, just 1 comment here as of this writing) mention of the Turing Test. The whole premise of the paper that introduced that was that "do machines think" is such a hard question to define that you have to frame the question differently. And it's ironic that we seem to talk about the Turing Test less than ever now that systems almost everyone can access can arguably pass it now.
  • alanuhoo 1 minute ago
    Valley of dispair type article
  • hamdingers 37 minutes ago
    If it comes to the correct answer I don't particularly care how it got there.
    • emp17344 3 minutes ago
      In most cases, you don’t know if it came to the correct answer.
    • ares623 31 minutes ago
      How do you know if it came to the right answer?
      • mcswell 3 minutes ago
        It's not always the case, but often verifying an answer is far easier than coming up with the answer in the first place. That's precisely the principle behind the RSA algorithm for cryptography.
  • croemer 1 hour ago
    Don't think this is very good - more of a report of their activities. Underdelivers on the headline.
  • terminalshort 59 minutes ago
    > the team wants to tackle a major and common misconception: that students think that ANN systems learn, recognise, see, and understand, when really it’s all just maths

    This is completely idiotic. Do these people actually believe that showing it can't be actual thought because it is described by math?

    • nomel 55 minutes ago
      Wait until they hear about the physics/maths related to neurons firing!
  • brador 22 minutes ago
    Do we think?

    By every scientific measure we have the answer is no. It’s just electrical current taking the path of least resistance through connected neurons mixed with cell death.

    The fact a human brain peaks at IQ around 200 is fascinating. Can the scale even go higher? It would seem no since nothing has achieved a higher score it must not exist.