Philosophy Discussion Forums | A Humans-Only Club for Open-Minded Discussion & Debate

[amorphos_ii]

AI might perceive redness as a wave of amplitude and frequency correlating with the appropriate language. Can AI “think” of the experience as gratifying? What if gratifying is a loop of redness that can be controlled to an optimal point by an intrinsic method? How would AI describe it? ,,, would that be the same as human gratification with the premise that such an evolving loop exists? From 1 to ten….

If the AI only knows the physical information, as with the example of red amplitude of frequency, then it cannot know what redness [qualia/quality] is. It is observing a wavelength of transparent light, and that is its informational reality.
If however, if the red sky at night does have the qualia/quality of redness, and one which can be replicated in terms of the physical wavelength then emitted at the other end e.g. in our visual cortex. The question then becomes; how can it know what the qualia is? Ergo as I see it, there has to be a mind present, because it is only mind that registers qualia!
There are then two kinds of intelligence, one that is being experienced and one that is not, but the latter is doing the same thing – in all other contexts of intelligence. However we can say what our experiences are, write about them and whatnot. Ergo experience itself is a facet of the intellect. Seems to be the same for observation and subjectivity. I guess there will be no true AI poets just mimics.

...unless AI does learn what qualia are and experience them. Should be no different for them as it is for the sky or maybe us. One could grow a human brain and simultaneously manufacture a synthetic imitation – a 3d processor perhaps. What would give one thing life and mind, but not the other thing. Would the cyborg ‘receive’ an ‘observer’ too. Is this something which automatically occurs e.g. if you manufactured an organic wasp cell by cell, would it not just be a living wasp like any other?

_

[Sy Borg]

AI might perceive redness as a wave of amplitude and frequency correlating with the appropriate language. Can AI “think” of the experience as gratifying? What if gratifying is a loop of redness that can be controlled to an optimal point by an intrinsic method? How would AI describe it? ,,, would that be the same as human gratification with the premise that such an evolving loop exists? From 1 to ten….

If the AI only knows the physical information, as with the example of red amplitude of frequency, then it cannot know what redness [qualia/quality] is. It is observing a wavelength of transparent light, and that is its informational reality.
If however, if the red sky at night does have the qualia/quality of redness, and one which can be replicated in terms of the physical wavelength then emitted at the other end e.g. in our visual cortex. The question then becomes; how can it know what the qualia is? Ergo as I see it, there has to be a mind present, because it is only mind that registers qualia!
There are then two kinds of intelligence, one that is being experienced and one that is not, but the latter is doing the same thing – in all other contexts of intelligence. However we can say what our experiences are, write about them and whatnot. Ergo experience itself is a facet of the intellect. Seems to be the same for observation and subjectivity. I guess there will be no true AI poets just mimics.

...unless AI does learn what qualia are and experience them. Should be no different for them as it is for the sky or maybe us. One could grow a human brain and simultaneously manufacture a synthetic imitation – a 3d processor perhaps. What would give one thing life and mind, but not the other thing. Would the cyborg ‘receive’ an ‘observer’ too. Is this something which automatically occurs e.g. if you manufactured an organic wasp cell by cell, would it not just be a living wasp like any other?

_

Re: redness, AI will know what humans identify as red, but you'd expect it to also see colours that we can't.

It may take many iterations for AI to achieve sentience, maybe millions of years, just as life operated just fine without sentience for a very long time until it reached a point where sentience was advantageous in some settings. Maybe ti will happen faster, but that seems more iffy.

If different AIs end up competing for the same resources in the far future, there could be an evolutionary "arms race", resulting in a chain of adaptations and counter-adaptations.

[Lagayascienza]

We might, for example, start by thinking about what a neuron is and the electro-chemical processes that occur in networks of neurons. Then we might look at how those processes produce awareness and intelligence. That’s what Harkins and Bennet do and it is that sort of thing I’d like to discuss if we could.

Sure, we can also go to SEP and look at the philosophy around consciousness and intelligence and what it all might mean but, for me, it would be easier to do that once we are on firm ground in respect of what is known and what is yet to be discovered.

What is known about what? About intelligence? How will you do that without stating what you mean by intelligence, in the first place? And neurons, why should we be looking at neurons first? That’s probably because you already have a preconception of why and how neurons enter the picture pf intelligence, which is fine, but I already placed a couple of challenges to that, one of them being: we find neurons all over the bodies of organisms, yet it seems that not always involved in processes that fall within what some define as intelligence. The parasympathetic nervous system is a neural network controlling organic processes unconsciously. Does that point to intelligence in action? Surely, we might clear up any doubt if we started by defining what is intelligence.

We know what intelligence is. The usual definitions, such as the one at Wiki, are fine. General intelligence (GI) is what we have. Any artificial neural network that can do all, or much of, what we can do, will be intelligent - it will be an AGI. What is known incontrovertibly about intelligence is that it emerges from the processes that occur in natural neural networks. The focus of such neural networks is the brain. However, our entire neural network is not necessary for consciousness or intelligence. For example, motor areas of the brain can be disconnected from nerves lower down in the body (as occurs in quadriplegia) while consciousness and intelligence remain intact. And much of what goes on unconsciously in the sympathetic nervous system (such as the control of the heart) is similarly irrelevant to consciousness and intelligence. We can be brain-dead and still have a heartbeat, but for consciousness and intelligence we need, at a minimum, a highly functional brain. Put the brain out of action (with general anesthesia, for example) and consciousness and intelligent behavior cease.

What neuroscientists and computer scientists will need in order to produce AGI is a more detailed understanding of how the brain does what it does. I do not think that AGI can be built with chess-playing machines or LLMs. That is just number-crunching and automation and not intelligence. But once we can build artificial neural networks of a similar complexity, connectivity and processing power to the brain, then will we be able to build AGI. We are some way from that now because the focus has been on machines that do only one or a limited number of things for which intelligence is unnecessary. And progress has been slow because the neuroscience and computer folks don't talk much to each other. But that is changing.

We already know quite a lot about our own neural network and the neural networks of other animals. What our brains do is learn, house memory, build models of the world we inhabit, and produce moment to moment updates about our interaction with that world. We know quite a bit already about the brain structures that house memory, and we know that learning involves making new synaptic connections. And we know how data that is fed to the brain via our sensorium enables the updating of our mental models moment to moment.

There is still a lot to learn but once we have a fuller picture of how natural neural networks do what they do, and once the technology to reproduce that in an artificial substrate is further developed, AGI should quickly follow.

[Gertie]

Count Lucanor

I have pointed out that there s not an undifferentiated, homogenous, continuous neural network in living beings, but a complex system organized in different anatomical parts with different functions, including hundreds of different types of neurons, and they operate several organic systems within an organism, which also differ between the whole range of organic forms, from insects to humans. I propose that this is very much unlike the computational neural networks, so there’s something there to talk about and try to understand in physical systems. At least it would help us test your hypothesis that the biological neural networks can be replicated artificially, independently of the fact that we have not reached a definition of what intelligence is.

Sorry to butt in.  It's my (very basic) understanding that  individual neurons across all species can vary in particulars like numbers of specific parts, lengths, coating which change speed of electrical activity and suchlike.  But they are all much the same in terms of what components they have.  Like Lego bricks might have different shapes, but are all the same basic structure.  And as far as we can tell neurons don't contain some special part which other cells don't. 

This strikes me as significant when trying to understand how ''something it is like to be'' conscious experience might manifest in a working brain. It suggests that the configuration and patterns of interactions between cells are the key to conscious experience manifesting, rather than some special component of a neuron. 

If this is the case, then that would support the proposal that a different substrate which can functionally electro-chemically mimic the brain's activity, would manifest conscious experience.  Ie that there could be 'something it is like' to be a machine which functionally mimics organic brain activity. 

Maybe even that certain patterns of interactions of any type of substrate could manifest conscious experience.  And even more speculatively, that any interaction of any matter could.  And perhaps does - we just don't recognise it.

I'd say our ignorance is such, and our inability to confidently infer from what scientific inquiry can (currently) offer, leaves these  questions wide open.

[Count Lucanor]

We know what intelligence is.

Disputable. I have given enough arguments of why it is disputable. Still waiting for an answer.

The usual definitions, such as the one at Wiki, are fine.

But that one from the Wiki appears not to be consistent with other definitions of intelligence you have endorsed here. Which one you will finally pick? Hawkins made his mind: intelligence is for him only human intelligence, the ability that only our human neocortex can provide and make us "the intelligent species". Do you agree with him?

General intelligence (GI) is what we have. Any artificial neural network that can do all, or much of, what we can do, will be intelligent - it will be an AGI.

Who is "we"? Humans? If so, then doing what a beetle does, that is not intelligence, right? "An animal doesn't need a neocortex to live a complex life", says Hawkins. And again, there's just too much ambiguity: what do you mean by "what we can do"? Our observable, external behavior? Is it what our cognitive apparatus actually does?

What is known incontrovertibly about intelligence is that it emerges from the processes that occur in natural neural networks.

If we care about the wording, that is disputable. It seems you reduce intelligence to neural networks, but can it? What we do know is that organisms with a complex nervous apparatus, comprised of several organs with different functions, can be intelligent. The bodies of all complex animals house lots of systems regulated by nerves, neural networks if you like, yet only the portion related to a neocortex seems to be associated with intelligence, if we adopt that definition of intelligence. That puts a vast amount of neural networks out of the intelligence equation, so how can it be "incontrovertible" that intelligence emerges from neural networks? We could also say that incontrovertibly, intelligence arises from processes that occur in cells. Well, sure, but that doesn't say anything really useful.

The focus of such neural networks is the brain.

Why?

However, our entire neural network is not necessary for consciousness or intelligence. For example, motor areas of the brain can be disconnected from nerves lower down in the body (as occurs in quadriplegia) while consciousness and intelligence remain intact. And much of what goes on unconsciously in the sympathetic nervous system (such as the control of the heart) is similarly irrelevant to consciousness and intelligence. We can be brain-dead and still have a heartbeat, but for consciousness and intelligence we need, at a minimum, a highly functional brain. Put the brain out of action (with general anesthesia, for example) and consciousness and intelligent behavior cease.

So, you agree that the presence of neural networks does not necessarily guarantee the presence of intelligent processes.

What neuroscientists and computer scientists will need in order to produce AGI is a more detailed understanding of how the brain does what it does.

Still lacking a justification of why the brain alone.

But once we can build artificial neural networks of a similar complexity, connectivity and processing power to the brain, then will we be able to build AGI.

So far, it is not looking as that being the case. First, it is only a hypothesis that you will get intelligence out of neural networks alone, since they would have to be organized with the complexity of the nervous systems of living organisms. Otherwise, it's like saying that an organism is ultimately reducible to its cells, and that once you have replicated cells from a particular tissue and its functions, you have solved the mystery of how to make an artificial complex organism as a whole: just make a tissue of similar complexity.

We are some way from that now because the focus has been on machines that do only one or a limited number of things for which intelligence is unnecessary. And progress has been slow because the neuroscience and computer folks don't talk much to each other. But that is changing.

We are in the wrong path because the mainstream theoretical framework is not the right one. Because we started using a metaphor of cognition in living beings to apply to computer machines, and then we used that as another metaphor to think of our own cognition. Our intelligence is being redefined in terms of the computer technology, assuming the computational theory of mind is true. I don't see any evidence of that currently changing.

We already know quite a lot about our own neural network and the neural networks of other animals. What our brains do is learn, house memory, build models of the world we inhabit, and produce moment to moment updates about our interaction with that world. We know quite a bit already about the brain structures that house memory, and we know that learning involves making new synaptic connections. And we know how data that is fed to the brain via our sensorium enables the updating of our mental models moment to moment.

There is still a lot to learn but once we have a fuller picture of how natural neural networks do what they do, and once the technology to reproduce that in an artificial substrate is further developed, AGI should quickly follow.

That's disputable. Quoting Hawkins: "how intelligence arises from cells in your head is still a profound mystery. As more puzzle pieces are collected each year, it sometimes feels as if we are getting further from understanding the brain, not closer".

[Count Lucanor]

Count Lucanor

I have pointed out that there s not an undifferentiated, homogenous, continuous neural network in living beings, but a complex system organized in different anatomical parts with different functions, including hundreds of different types of neurons, and they operate several organic systems within an organism, which also differ between the whole range of organic forms, from insects to humans. I propose that this is very much unlike the computational neural networks, so there’s something there to talk about and try to understand in physical systems. At least it would help us test your hypothesis that the biological neural networks can be replicated artificially, independently of the fact that we have not reached a definition of what intelligence is.

Sorry to butt in.  It's my (very basic) understanding that  individual neurons across all species can vary in particulars like numbers of specific parts, lengths, coating which change speed of electrical activity and suchlike.  But they are all much the same in terms of what components they have.  Like Lego bricks might have different shapes, but are all the same basic structure.  And as far as we can tell neurons don't contain some special part which other cells don't. 

First, your basic understanding is wrong. There are hundreds of types of cells, although generally classified within three larger groups when talking about most of the nervous system. When it comes to the brain, it gets even more complicated.

Secondly, I don't know how that goes to my point. Your cognitive apparatus is not made of just a network of neurons, in other words, it cannot be reduced to it. It has different anatomical parts with different functions, so one could say that the neural networks in our bodies need to be organized physiologically in a certain way to operate and produce cognition. Dismissing the importance of that is pure reductionism aimed at facilitating the computational metaphor.

[The Beast]

In a speculative way, I could decide that a worm is born with the capabilities of being a worm fitting the allowed (capacities) intelligence to the worm form. Pub Med has interesting articles about nematodes. Specifically, TGF-beta signaling in parasitic nematodes evolving from the ground environment. Of course, there is an area of research that involves the use of compounds to manipulate TDF signaling. This is called: Artificial TGF-beta signaling… etc… and the cure of arthritis and other diseases of the immune system. How does it work? The most relevant part is that signaling exists. IMO there is a hierarchy of intelligence and of dangerous activities. Is there a comprehensive knowledge of how the worms interact with the ground? Optimistic scientists might construct steel mini excavators with evolving charts in the spectrum of directed evolution.

[Gertie]

Count Lucanor

I have pointed out that there s not an undifferentiated, homogenous, continuous neural network in living beings, but a complex system organized in different anatomical parts with different functions, including hundreds of different types of neurons, and they operate several organic systems within an organism, which also differ between the whole range of organic forms, from insects to humans. I propose that this is very much unlike the computational neural networks, so there’s something there to talk about and try to understand in physical systems. At least it would help us test your hypothesis that the biological neural networks can be replicated artificially, independently of the fact that we have not reached a definition of what intelligence is.

Sorry to butt in.  It's my (very basic) understanding that  individual neurons across all species can vary in particulars like numbers of specific parts, lengths, coating which change speed of electrical activity and suchlike.  But they are all much the same in terms of what components they have.  Like Lego bricks might have different shapes, but are all the same basic structure.  And as far as we can tell neurons don't contain some special part which other cells don't. 

First, your basic understanding is wrong. There are hundreds of types of cells, although generally classified within three larger groups when talking about most of the nervous system. When it comes to the brain, it gets even more complicated.

Would you agree then that the known difference between the 3 broad types of neurons and other cells, which looks relevant to conscious experience, is the ability to transfer 'neurotransmitter' ions via axons and dendrites (with some modifiers). And that neurons associated with hearing, vision, pain, memory, etc aren't apparently significantly different to each other in type.

Secondly, I don't know how that goes to my point. Your cognitive apparatus is not made of just a network of neurons, in other words, it cannot be reduced to it. It has different anatomical parts with different functions, so one could say that the neural networks in our bodies need to be organized physiologically in a certain way to operate and produce cognition.

Sure. I'm trying to get to what makes working neurons different to eg cells directly involved in digestion, which don't manifest conscious experience. This could help us identify whether neurons possess some key ingredient of consciousness, as opposed to the possibility that any cells with similarly interactive configurations could work.

The known notable thing to me about neurons is their role as neurotransmitters, facilitating the flexible transfer of ions within a highly complex interactive system via axons and dendrites.

If an artificial system could do that, then the question would be can we test if that system is conscious.

Dismissing the importance of that is pure reductionism aimed at facilitating the computational metaphor.

No. I don't find metaphors such as 'computation' or 'information processing' helpful here. Information isn't a 'thing in itself' with properties and causal powers which is computed by the brain. Embodied working brains are physical stuff and processes doing presumably physically explainable things.

[Pattern-chaser]

I'm trying to get to what makes working neurons different to eg cells directly involved in digestion, which don't manifest conscious experience. This could help us identify whether neurons possess some key ingredient of consciousness, as opposed to the possibility that any cells with similarly interactive configurations could work.

When I read this, I thought immediately of a holistic perspective, that might or could suggest that conscious experience is the product of all our cells, not just the ones that are most clearly and obviously involved in it. I wondered if my musing might spark some interest in this discussion?

[Gertie]

I'm trying to get to what makes working neurons different to eg cells directly involved in digestion, which don't manifest conscious experience. This could help us identify whether neurons possess some key ingredient of consciousness, as opposed to the possibility that any cells with similarly interactive configurations could work.

When I read this, I thought immediately of a holistic perspective, that might or could suggest that conscious experience is the product of all our cells, not just the ones that are most clearly and obviously involved in it. I wondered if my musing might spark some interest in this discussion?

For sure every particular instantiation of a conscious entity will have an explanation which can be contextualised almost boundlessly. And there are plenty of broad cloth hypotheses about consciousness, which we struggle to choose from because they seem to be untestable. So I wonder how might that help us answer a question like Can AI Be Conscious do you think?

If we want to think about why Entity A is conscious and Entity B isn't (as far as we can tell), then it strikes me a sensible approach could be looking for similarities - and especially differences. To try to narrow down the necessary and sufficient conditions for conscious experience to manifest.

[Pattern-chaser]

I'm trying to get to what makes working neurons different to eg cells directly involved in digestion, which don't manifest conscious experience. This could help us identify whether neurons possess some key ingredient of consciousness, as opposed to the possibility that any cells with similarly interactive configurations could work.

When I read this, I thought immediately of a holistic perspective, that might or could suggest that conscious experience is the product of all our cells, not just the ones that are most clearly and obviously involved in it. I wondered if my musing might spark some interest in this discussion?

For sure every particular instantiation of a conscious entity will have an explanation which can be contextualised almost boundlessly. And there are plenty of broad cloth hypotheses about consciousness, which we struggle to choose from because they seem to be untestable. So I wonder how might that help us answer a question like Can AI Be Conscious do you think?

I don't see a holistic perspective being especially helpful in answering your final question, the one to which this topic is devoted. But perhaps it might be helpful in answering the question that comes before it?

Like humans do, we are trying to run marathons before we have started crawling. I think our understanding of consciousness, human or otherwise, is close to non-existent. Perhaps a holistic view might help us with this?

If we want to think about why Entity A is conscious and Entity B isn't (as far as we can tell), then it strikes me a sensible approach could be looking for similarities - and especially differences. To try to narrow down the necessary and sufficient conditions for conscious experience to manifest.

This is a very analytic approach to a subject that is, perhaps, less suited to that kind of approach?

[Pattern-chaser]

I just came across this AI-related article, that might be of interest?



The Conversation is a reputable site that carries all kinds of articles that relate to current affairs. It should be safe to visit and use.

[Empiricist-Bruno]

Please excuse me to reply in this slightly informal and unusual way, but otherwise it would've sprawled beyond comfortable reading size.

The same thing appears to be happening with AI. Some computer programmers discovered a way to build what appears to be intelligence. <Not "discovered", as though it was there, ready to be found, but "designed". Software designers created a program, or suite of programs, to meet a need.> Just ask a simple calculator for a complex multiplication and it will respond instantly. For us to be able to perform such a task, it would require us much more time and we'd need to apply our intelligence to come up with the right answer. <Mental arithmetic requires rather less than "intelligence", IMO. It's mechanical, even to the point where some of the first calculating machines were mechanical. Starting with Babbage's Difference Engine, and ending with the mechanical marvels I used to use, when I was in school, to do my homework. > Therefore, the calculator appears to be doing something intelligent, through projection. But yet, we know it's not there. There is no intelligence at work there. <> So, how are we going to qualify apparently intelligent work when it's done by electronic circuitry? <We could call it "that which resembles intelligence, but actually is just a seeming; a simulation"?> Calling it Artificial Intelligence makes sense because you can't call it projection intelligence because a projection is something abstract and you want to refer to the intelligence of circuitry as a thing. <And if I want to refer to the fairy dust that makes it work, will you let me describe it as magic, even though it is no such thing? Circuitry is not intelligent.> So, the next logical thing to do is to call it artificial intelligence because this helps you to understand that you are not referring to natural intelligence which is what real intelligence is but instead, you are referring to intelligence that comes from human culture and which is based on the things that we construct ourselves. And that's why it's called that way, and not because there is any intent to deceive any one. And there is no one deceived into thinking that there is actual intelligence in the artificial intelligence; we just call it that way because we don't know of any better option.

<I never thought, or wrote, that there was/is any intention to decieve. I wrote only what I believe to be the truth. Software designers tried for years to create something roughly worthy of the title "Artificial Intelligence". Their work was without significant success. It was just too difficult a task. And so the conscious decision was made to try to simulate intelligence, and that is the path that has lead to current AI as we know it. There was no intent to deceive, to the point that AI designers made clear their change of course, and explained it to anyone interested enough to listen. Hence there was and is no deception.>

And we don't know any better option because we generally fail to realize that the intelligence of circuitry is actually a paradox of intelligence. <The "intelligence of circuitry" is a terrible misunderstanding that is too deep, and too misguided, to properly refute here. It does not exist.>

<The "intelligence of circuitry" is a terrible misunderstanding that is too deep, and too misguided, to properly refute here. It does not exist.>

I have somewhat of an issue with your above statement but not because I support in anyway the existence of intelligence of circuitry but because your reply seems to be within the context of a rebuttal of an argument or thing that I have mentioned. When I talk about the intelligence of circuitry, I immediately claim that this intelligence is paradoxical and complain that others often fail to notice that. Now, in my mind, if any other reads this, he/she would immediately realize that I don't believe that intelligence of circuitry is in any way a form of intelligence like say, "emotional intelligence". If I say that the intelligence of circuitry is paradoxical, it means I certainly agree that it does not exist. It's just another way to say that. But here you go, apparently saying that my concept of intelligence of circuitry needs to be refuted as that it simply doesn't exist, which was what I was showing / explaining. And so I am left a bit perplexed. I will present you an example drawn from my own life which may help clarify why paradoxical intelligence isn't intelligence at all:

My dad used to claim that I was indisputably intelligent but that I never used my gifted endowment. Now that's a paradox. An intelligent person that does not use his or her intelligence is an idiot. There's no other way around it. But if the person that does that is intelligent then that intelligence must be paradoxical. That's a simple logical deduction. So, my dad used a paradox to imply that people shouldn't believe in my intelligence as it was paradoxical. It simply wasn't really there, although some might believe I have impressive amount of that. According to my dad, you would need to put a red flag on my intelligence as it wasn't what it appeared to be, according to him. That's from my understanding of paradoxes because he didn't actually say these last complementary sentences but I felt (and still feel) they were implied by him.

Now today, I talk about circuitry intelligence in the same way that my dad used to talk about my intelligence. And in reply, I get to notice that you say that I need to understand that circuitry intelligence doesn't exist. Ok, please explain to me what's your point, or why you think you need to inform me of this. Do you think that my dad thought I was really intelligent? Perhaps he is only half intelligent and didn't realize what he was saying about me?

<I never thought, or wrote, that there was/is any intention to deceive. I wrote only what I believe to be the truth. Software designers tried for years to create something roughly worthy of the title "Artificial Intelligence". Their work was without significant success. It was just too difficult a task. And so the conscious decision was made to try to simulate intelligence, and that is the path that has lead to current AI as we know it. There was no intent to deceive, to the point that AI designers made clear their change of course, and explained it to anyone interested enough to listen. Hence there was and is no deception.>

No, it's not a paradox. It's a deliberate (and transparent) deception. (Current) AIs are implemented to look intelligent, not to be so.

So, when you wrote that it was a "deliberate (and transparent) deception" you were using the wrong term? What you meant is "delusion" because when a deception is transparent what you have is actually an oxymoron or a short paradox, right? No deception is open and transparent about what it is. If it were that way, it would be a confession no?
Also, when you provided clarification about why you mentioned it was a "deception" here:

It is our human implementation of AI that is transparently deceptive. Deceptive because they emulate the appearance of intelligence without actually being intelligent, and transparent (i.e. widely known and appreciated) because no-one denies this, or claims it isn't so (because it is so).

You switched from the use of the term "deception" to "deceptive". So, it seems to me that you have failed to either mention that your initial use of the term "deception" was incorrect and that you meant deceptive. There clearly is a difference between the two terms although they have the same root. What is deceptive can occur naturally, without any intent. On the other hand a deception never occurs that way, to my knowledge.

I wrote only what I believe to be the truth.

Yes, and that makes you such an extraordinary engaging blogger. Keep this up.

The same thing appears to be happening with AI. Some computer programmers discovered a way to build what appears to be intelligence. <Not "discovered", as though it was there, ready to be found, but "designed". Software designers created a program, or suite of programs, to meet a need.>

Here, I find your comment very interesting because I have no idea as to why you are picking on this part of my text. So let me ask you, how would a design not be there, ready to be found? Wasn't calculus found or perhaps it would be better to say about it that it was designed? How does selecting such terminology matters?

<Mental arithmetic requires rather less than "intelligence", IMO. It's mechanical, even to the point where some of the first calculating machines were mechanical. Starting with Babbage's Difference Engine, and ending with the mechanical marvels I used to use, when I was in school, to do my homework. >

Have you ever come across a mental arithmetic champion that was rather unintelligent? My own past online dealings with such a world champion suggest quite the opposite. I think people will say that you don't need to be intelligent to count because many animals can clearly do so and they want to distinguish themselves from animals through their possession of intelligence, which I find racist and offensive.

So, how are we going to qualify apparently intelligent work when it's done by electronic circuitry? <We could call it "that which resembles intelligence, but actually is just a seeming; a simulation"?>

I very much appreciate you attention to detail and attention to all my points and questions. Here, I notice that if we call it the way you suggest, then we have to deal with a margin: If something resembles another but is a simulation of the other you have to answer the question which is which? Which is the thing that imitates the other and which one is being imitated? You then become open to ideas that intelligence may actually come from things (designed circuitry) and not from people. Or maybe intelligence may come from both? And then you have to work on very hard to defend margin between the true intelligence of one and the simulated one of the other. All of that shows that the way to go is clear: call it paradoxical intelligence.

[Pattern-chaser]

If I say that the intelligence of circuitry is paradoxical, it means I certainly agree that it does not exist.

Then why not just say so?

The same thing appears to be happening with AI. Some computer programmers discovered a way to build what appears to be intelligence. <Not "discovered", as though it was there, ready to be found, but "designed". Software designers created a program, or suite of programs, to meet a need.>

Here, I find your comment very interesting because I have no idea as to why you are picking on this part of my text. So let me ask you, how would a design not be there, ready to be found?

I spent 40 years designing digital electronic hardware and software. "Design" is a subject close to my heart.

We can say, figuratively and artistically, that the design was there, waiting to be found. Just as Michelangelo's David was nestling inside a lump of marble, waiting to be 'released'. But outside such artistic use of language, designs do not sit around waiting to be discovered.

[Count Lucanor]

Would you agree then that the known difference between the 3 broad types of neurons and other cells, which looks relevant to conscious experience, is the ability to transfer 'neurotransmitter' ions via axons and dendrites (with some modifiers). And that neurons associated with hearing, vision, pain, memory, etc aren't apparently significantly different to each other in type.

Actually, there are many things associated with conscious experience, and neurons are obviously a key component, although not the only one. Glial cells, for example, don’t function as neurons do, but play a key role. Also, there are lots of neurons doing their job in the cognitive apparatus of complex organisms, including mammals, regulating the operation of several organic systems (such as heart, liver, kidney, etc.), while having nothing to do with conscious experience, So it seems obvious that you need neurons plus a lot of other things to produce consciousness.

Secondly, I don't know how that goes to my point. Your cognitive apparatus is not made of just a network of neurons, in other words, it cannot be reduced to it. It has different anatomical parts with different functions, so one could say that the neural networks in our bodies need to be organized physiologically in a certain way to operate and produce cognition.

Sure. I'm trying to get to what makes working neurons different to eg cells directly involved in digestion, which don't manifest conscious experience. This could help us identify whether neurons possess some key ingredient of consciousness, as opposed to the possibility that any cells with similarly interactive configurations could work.

Interestingly, neurons are not disassociated from digestive processes, since the parasympathetic nervous system controls visceral functions. Why then would you reduce neuronal functions to those of conscious experience?

The known notable thing to me about neurons is their role as neurotransmitters, facilitating the flexible transfer of ions within a highly complex interactive system via axons and dendrites.

If an artificial system could do that, then the question would be can we test if that system is conscious.

I repeat my question: why only consciousness? Neurons are involved in practically all functions of an organism, so, if you managed to replicate any or those functions in a machine, why wouldn’t you just say that it is an organic system? There are nerves in simple organisms. Would you then consider that any organism with neurons, such as anemones and corals, are conscious?

Dismissing the importance of that is pure reductionism aimed at facilitating the computational metaphor.

No. I don't find metaphors such as 'computation' or 'information processing' helpful here. Information isn't a 'thing in itself' with properties and causal powers which is computed by the brain. Embodied working brains are physical stuff and processes doing presumably physically explainable things.

I agree, but I will add: it’s not just a brain cognizing, but a complete organism.