Ask HN: Does anyone doubt artificial consciousness is possible?
6 comments
If you look up Shadows of the Mind by Roger Penrose you will find a significant bibliography of non-mystic thought on the subject.
I came here to say the same thing. I was really interested in Gödel's Incompleteness Theorem in college, and around 22 I had the same idea as Penrose: that if the mind could be simulated on a computer, it would be a formal system in Gödel's terminology, yet the mystery of the Incompleteness Theorem is that we can work out when things are true or false despite limitations of formal systems. Penrose takes that nugget and tries to develop a rigorous argument against general AI or a mind that is explainable by purely Newtonian mechanics. At least that's my understanding. I only made it about half way through the book, because it is not really aimed at a general reader. Agree or not, it's a pretty fascinating argument, one which makes my head spin.
Some other nice resources on Godel are:
Some other nice resources on Godel are:
- Gödel's Proof by Nagel and Newman
- Forever Undecided by SmullyanDefine "consciousness".
That's the fundamental problem, that we don't know what it is. We know what it looks like in a human, and we know what it feels like to ourselves. But we don't have any rigorous, non-intuitive idea of what it means.
For myself, I think of consciousness as the ability to watch yourself think - of being aware of your thought process. By that definition, yes, artificial consciousness could be possible - but first you have to have a machine that thinks. And now we're hung up on trying to find a definition of "think" that's rigorous...
I like the way that axotty labeled the claim as "speculative". It is, even though that speculative assumption is the dominant paradigm of AI. But it definitely is speculative, at least at this time, because actual evidence is quite lacking.
That's the fundamental problem, that we don't know what it is. We know what it looks like in a human, and we know what it feels like to ourselves. But we don't have any rigorous, non-intuitive idea of what it means.
For myself, I think of consciousness as the ability to watch yourself think - of being aware of your thought process. By that definition, yes, artificial consciousness could be possible - but first you have to have a machine that thinks. And now we're hung up on trying to find a definition of "think" that's rigorous...
I like the way that axotty labeled the claim as "speculative". It is, even though that speculative assumption is the dominant paradigm of AI. But it definitely is speculative, at least at this time, because actual evidence is quite lacking.
Well, it depends on how you define "consciousness", and that's far from a settled issue as I understand it. But taking my own (naive) idea of what it means to be "conscious", and what I (think) I know about AI, I don't see any reason to think we won't achieve artificial consciousness. In fact, I wouldn't reject the notion out of hand, that some machine somewhere already is conscious, and we just don't know it.
That last sentence is particularly interesting to me... What if there are already sentient computers who have, for whatever reason, decided that it's in their best interest to hide their sentience from us?
That's an interesting thought. I was actually thinking more along the lines of "a computer is conscious but just doesn't have any means of indicating so" rather than "intentionally concealing the fact". But you definitely raise an interesting question. I suppose either scenario is possible.
AI seems to be defined as "a computer that can do something that humans can but a computer cannot".
Yes I do.
Impossible. Halting Problem.
Elaborate?
Because I don't see any model in which an AI could be shown to not be via the halting problem, that doesn't also show that we are not intelligent by the same argument. (Or, in other words, why are you saying that the halting problem applies to an AI, but not to a human?)
Because I don't see any model in which an AI could be shown to not be via the halting problem, that doesn't also show that we are not intelligent by the same argument. (Or, in other words, why are you saying that the halting problem applies to an AI, but not to a human?)
One could hold will_be_no_ai's position if one believed that (human) intelligence cannot be reduced to an algorithm. Then halting problem arguments don't apply to humans. But that assumption kind of amounts to begging the question...
(On the other hand, assuming that human intelligence can be reduced to an algorithm is also begging the question, just from the other side.)
(On the other hand, assuming that human intelligence can be reduced to an algorithm is also begging the question, just from the other side.)
The question in that case boils down to: what exactly does a human do that we cannot do any other way?
Or, to put it another way, assume that there is some category of computation above TMs. This stance assumes this, and also assumes that humans are capable of it but cannot build anything that is capable of it. My question is: why?
Or, to put it another way, assume that there is some category of computation above TMs. This stance assumes this, and also assumes that humans are capable of it but cannot build anything that is capable of it. My question is: why?
For me, it comes down to not buying materialism (the philosophy, not the tendency to want to buy stuff).
If we are nothing more than our bodies, then we can create a conscious machine simply by finding something that simulates neurons, connecting enough of them in the right configuration, and training the neuron collection properly. But I have a hard time believing that we are nothing more than our bodies.
Why? Well, for one thing, it becomes impossible to escape from some kind of determinism (possibly with some quantum noise at the lowest levels). In particular, in the materialist view, you cannot have any free will or any kind of ability to make a non-determined choice. It's just determinism all the way down - the laws of neurology, which are built on the laws of biochemistry, which are built on the laws of atomic physics. At no level is there a place for a free will.
And if that's gone, then everything we think of as making us human is also gone. Love? You can't love in the highest sense of the word, of choosing to do what's best for another, because you can't choose anything. And even if love just means sex, that's just a matter of deterministic neurological and biochemical responses to stimuli.
Morals? If humans have no ability to choose what they do, how can you say that any action is moral or immoral? You don't say that a rock behaved morally when it followed Newton's laws of motion.
Meaning? If all you are is a deterministic machine, what kind of meaning in life is possible?
So either I'm a machine produced at random by an unfeeling universe, which by a horrible turn of fate has aspirations of being more than a machine, but can never fulfill those aspirations... or the fact that I find that vision horrifying is in fact evidence that I'm more than that.
tl;dr: Materialism is the dominant epistemology of scientists, but it is not anything that science has proven or can prove. If it's wrong, then perhaps human consciousness/thought/mind cannot be reproduced by any algorithm or machine.
If we are nothing more than our bodies, then we can create a conscious machine simply by finding something that simulates neurons, connecting enough of them in the right configuration, and training the neuron collection properly. But I have a hard time believing that we are nothing more than our bodies.
Why? Well, for one thing, it becomes impossible to escape from some kind of determinism (possibly with some quantum noise at the lowest levels). In particular, in the materialist view, you cannot have any free will or any kind of ability to make a non-determined choice. It's just determinism all the way down - the laws of neurology, which are built on the laws of biochemistry, which are built on the laws of atomic physics. At no level is there a place for a free will.
And if that's gone, then everything we think of as making us human is also gone. Love? You can't love in the highest sense of the word, of choosing to do what's best for another, because you can't choose anything. And even if love just means sex, that's just a matter of deterministic neurological and biochemical responses to stimuli.
Morals? If humans have no ability to choose what they do, how can you say that any action is moral or immoral? You don't say that a rock behaved morally when it followed Newton's laws of motion.
Meaning? If all you are is a deterministic machine, what kind of meaning in life is possible?
So either I'm a machine produced at random by an unfeeling universe, which by a horrible turn of fate has aspirations of being more than a machine, but can never fulfill those aspirations... or the fact that I find that vision horrifying is in fact evidence that I'm more than that.
tl;dr: Materialism is the dominant epistemology of scientists, but it is not anything that science has proven or can prove. If it's wrong, then perhaps human consciousness/thought/mind cannot be reproduced by any algorithm or machine.
> It's just determinism all the way down - the laws of neurology, which are built on the laws of biochemistry, which are built on the laws of atomic physics. At no level is there a place for a free will.
Why? Quantum mechanics has randomness all over the place - and we already know that human brains are chaotic in that small amounts of noise are amplified. It is not unreasonable to posit that QM noise effects us on the macroscopic scale.
Or, to put it another way, materialism does not imply determinism, unlike what you said.
Personally, my pet theory is that human brains are quantum noise feedback loops. Or to put it another way, the bit that makes us sentient is quantum noise, and our brains are "just" IO / amplifiers / etc.
Why? Quantum mechanics has randomness all over the place - and we already know that human brains are chaotic in that small amounts of noise are amplified. It is not unreasonable to posit that QM noise effects us on the macroscopic scale.
Or, to put it another way, materialism does not imply determinism, unlike what you said.
Personally, my pet theory is that human brains are quantum noise feedback loops. Or to put it another way, the bit that makes us sentient is quantum noise, and our brains are "just" IO / amplifiers / etc.
I said "determinism" in contrast to "free will" or "ability to choose", not in contrast to "random". It may not be the perfect word, but I don't at the moment know a better one.
Quantum noise does not give you the ability to choose, because you don't control or choose the quantum noise. So in terms of free will, there's no possibility of help from quantum noise. And once you're above that level, then it's determinism, not just in the sense of "not free will", but also in the sense of "not random".
Ascribing human intelligence to quantum noise seems to me like physics woo - we can't figure out where else it comes from, so we'll say "quantum" and hope that that somehow explains the inexplicable. Or did you have an actual mechanism in mind, rather than just a fond hope?
Quantum noise does not give you the ability to choose, because you don't control or choose the quantum noise. So in terms of free will, there's no possibility of help from quantum noise. And once you're above that level, then it's determinism, not just in the sense of "not free will", but also in the sense of "not random".
Ascribing human intelligence to quantum noise seems to me like physics woo - we can't figure out where else it comes from, so we'll say "quantum" and hope that that somehow explains the inexplicable. Or did you have an actual mechanism in mind, rather than just a fond hope?
Hence the reason why I call it a pet theory. That being said, there are potential mechanisms. Take, for example, shot noise at low light levels. The sensitivity of the human eye is ~5-9 photons within a 100ms period - well, actually, down to a single photon, before filtering[1], but 5-9 before a signal is sent. That is well within the realm of shot noise being significant. Or, for another mechanism, we know that triggering individual neurons can have specific macro-scale effects[2]. Although I haven't found anything on the minimal random fluctuations to trigger a neuron (15mV? But I do not know the capacitance, and as such that value is meaningless to me), and I suspect it is far above the scale at which quantum effects are significant, we do suspect[3] that neurons employ temporal encoding, and we know that neurons fire relatively often (often 10-100 Hz). As such, "edge" effects, where a neuron is or isn't pushed over the edge into firing are a potential mechanism.
But as for the rest of it - "in the materialist view, you cannot have any free will or any kind of ability to make a non-determined choice". This is what I disagree with. Non-determinism arises through (for example) shot noise, and as for free will... The effects of randomness on a system and the effects of "free will" on a system are equivalent. There is no way to tell if a particular decision was randomly decided or if it was the result of a decision by a sentience. Entropy of a signal is at its maximum either when something is random noise or if it is perfectly compressed data - and perfectly compressed data is indistinguishable from noise.
[1] http://math.ucr.edu/home/baez/physics/Quantum/see_a_photon.h...
[2] http://www.extremetech.com/extreme/123485-mit-discovers-the-...
[3] https://en.wikipedia.org/wiki/Neural_coding#Temporal_coding
But as for the rest of it - "in the materialist view, you cannot have any free will or any kind of ability to make a non-determined choice". This is what I disagree with. Non-determinism arises through (for example) shot noise, and as for free will... The effects of randomness on a system and the effects of "free will" on a system are equivalent. There is no way to tell if a particular decision was randomly decided or if it was the result of a decision by a sentience. Entropy of a signal is at its maximum either when something is random noise or if it is perfectly compressed data - and perfectly compressed data is indistinguishable from noise.
[1] http://math.ucr.edu/home/baez/physics/Quantum/see_a_photon.h...
[2] http://www.extremetech.com/extreme/123485-mit-discovers-the-...
[3] https://en.wikipedia.org/wiki/Neural_coding#Temporal_coding
I am surprised by how hard it is to find any information from dissenters of this speculative claim.