RE: [agi] How wrong are these numbers?
Alan The next question is: What's your corresponding estimate of processing power? To emulate the massively parallel information update rate of the brain on N bits of memory, how many commodity PC processors are required per GB of RAM? Ben G Ben Goertzel wrote: A short, interesting article on the information capacity of the brain was written by nanotechnologist Ralph Merkle; see http://www.merkle.com/humanMemory.html He gives figures between 10^9 bits and 10^15 bits. [ The upper bound, which in my oppinion is unrealistic, that I gave in my orrigional post was 1.6*10^10 bits. (8*2gb).] Yes, I am aware of those results. His information is based on counting synapses. -- I just re-read it... It seems my numbers aren't off base at all. What it means is that we are in the 'singularity window' _NOW_. To continue the line of thought: Lets say that the information content of the output of a cortical column can be expressed in firings / EEG period. The wakefull frequency is around 30hz. A neuron can fire as fast as 1khz. Doing the division yields an output capacity of 33 1/3 bits/EEG cycle. (it is probable that there are secondary IO channels for each cortical column). -- pain (n): see Linux. http://users.rcn.com/alangrimes/ --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
RE: [agi] How wrong are these numbers?
Kevin, About mind=brain ... My own view of that elusive entity, mind is well-articulated in terms of the philosophy of Charles S. Peirce, who considered there to be several different levels on which mind could be separately considered. Peirce used three levels, but inspired by Jung and others, I have introduced a fourth, and we prefer to think about: 1. First, raw experience 2. Second, physical reaction 3. Third, relationship and pattern 4. Fourth, synergy and emergence Each of these levels constitutes a different perspective on the mind; and many important mental phenomena can only be understood by considering them on several different levels. First corresponds roughly speaking to consciousness. On this level, analysis has no more meaning than the color red, and everything is simply what it presents itself as. We will not speak about this level further in this article, except to say that, in the Peircean perspective, it is an aspect that everything has even rocks and elementary particles not just human brains. Second, the level of physical reaction, corresponds to the machinery underlying intelligent systems. In the case of humans, its bodies and brains; in the case of groups of humans, its sets of bodies and brains. In fact, theres a strong case to be made that even in the case of individual human minds, the bodies and brains of a whole set of humans is involved. No human mind makes sense in isolation; if a human mind is isolated for very long, it changes into a different sort of thing than an ordinary human mind as embedded in society. Third, the level of relationship and pattern, is the level that is most commonly associated with the word mind in the English language. One way of conceiving of the mind is as the set of patterns associated with a certain physical system. By associated with we mean the patterns in that system, and the patterns that emerge when one considers that system together with other systems in its habitual environment. So, for instance, the human mind may be considered as the set of patterns in the human brain (both in its structure, and in its unfolding over time), and the patterns that are observed when this brain is considered in conjunction with other humans and its physical environment. This perspective may justly be claimed incomplete it doesnt capture the experiential aspect of the mind, which is First; or the physical aspect of the mind, which is Second. But it captures a very important aspect of mind, mind as relationship. This view of mind in terms of patterns may be mathematically formalized, as has been done in a loose way in my book From Complexity to Creativity. Fourth, the level of synergy, has to do with groups of patterns that emerge from each other, in what have been called networks of emergence. A mind is not just a disconnected bundle of patterns, its a complex, self-organizing system, composed of patterns that emerge from sets of other patterns, in an interpenetrating way. The notion of synergy is particularly important in the context of collective intelligence. The mind of a group of people has many aspects experiential, physical, relational and synergetic but what distinguishes it from the minds of the people within the group, is specifically the emergent patterns that exist only when the group is together, and not when the group is separated and dispersed throughout the rest of society. One thing all this means is that the number of bits needed to realize a mind physically, does not equal the number of bits in the mind. One cannot reduce mind to the Second level. The physical substructure of a mind is the key unlocking the door to a cornucopia of emergent patterns between an embodied system and its environment (including other embodied systems). These patterns are the mind, and they contain a lot more information than is explicit in the number of bits in the physical substrate. Regarding quantum or quantum gravity approaches to the mind, these are interesting to me, but from a philosophical perspective they're just details regarding how the physical universe organizes its patterns... they don't really affect the above general picture -- Ben G -- Ben G -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of maitri Sent: Tuesday, December 03, 2002 6:10 PM To: [EMAIL PROTECTED] Subject: Re: [agi] How wrong are these numbers? I've got a sawbuck in my pocket that says that you are seriously underestimating the capacity of the human mind. In fact, its questionable whether you can emulate a mouse brain adequately with that amount of power. I also think you guys are seriously underestimating the memory capacity of the human mind. Of course, I view the fundamental problem with your analysis as the mistaken assumption that mind=brain. There's a lot of anecdotal evidence that indicates the error in this line of thinking
RE: [agi] How wrong are these numbers?
Kevin, You raise a critical point, and my thinking on this point is a bit unorthodox, as well as incomplete... There is a big unanswered question close to the heart of my theory of mind, and this is the connection between Firstness and Fourthness. I sum up this question with the Orwell paraphrase All things are conscious, but some things are more conscious than others. I'm a Peircean animist, in the sense that I believe consciousness is everywhere. Yet, I believe that I'm more intensely conscious than a dog, and a dog is more intensely conscious than a flea, which is more intensely conscious than a virus, which is more intensely conscious than a molecule One question is: Why is this? But I'm not even sure of the standpoint from which this question Why? is asked. Another question is: What are the specifics of this law connecting Firstness with degree-of-integrated-complexity (an aspect of Fourthness)? This is something that interests me greatly Along these lines, I believe that if one constructs an AGI with a high degree of general intelligence, ensuing from a high degree of synergetic integrated complexity [the only way to get significant general intelligence, I think], this AGI system *will have* a highly intense degree of consciousness, analogous to (though with a different subjective quality from) that of humans. But I don't have a justification for this belief of mine, because I don't have a solution to the so-called hard problem of consciousness. All I have is an analysis of the hard problem of consciousness, which suggests that it may be possible to create artificial consciousness by creating AGI and watching the intense consciousness come along for free. I suspect that in fact the hard problem will remain in some sense unsolved. That is, the qualitative nature of the connection between intensity of consciousness and degree of general intelligence [for lack of a better phrase] may remain mysterious. Yet, by experimenting with artificial minds, we may learn to quantify this relationship. Quantify it how? Various sorts of artificial minds may report their subjective experiences -- in all sorts of subtle realms of awareness, as well as their own variety of everyday ordinary consciousness -- and we and they may learn rules relating their subjective experiences with known aspects of their physical implementations. Yet, even when such laws are known -- laws relating aspects of the conscious experience of a mind to aspects of its brain -- this will still not resolve the core mystery of how consciousness-mind relates to pattern-mind. But this mystery is ultimately not a question for science, perhaps... and we don't need to articulate its solution in order to build and study minds that are genuinely conscious... we can *feel* its solution sometimes, but that's another story... -- Ben G -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of maitri Sent: Tuesday, December 03, 2002 8:04 PM To: [EMAIL PROTECTED] Subject: Re: [agi] How wrong are these numbers? Ben, I think I followed most of your analysis :) I agree with most of what you stated so well. The only difficulty for me is that the patterns, whether emergent in the individual or the group, still pertain to the gross level of mind and not the subtle levels of consciousness. It is quite OK, IMO, to disregard this subtle aspect of mind in your design for AGI, Strong AI or the Singularity. But it should be noted that this is disregarding what I would consider the predominant capabilities of the human mind. For instance, in relation to memory capacity. let's say I could live for the age of the universe, roughly 15 billion years. I believe the human mind(without enhancement of any kind) is capable of remembering every detail of every day for that entire lifespan. A person can only understand this if they understand the non-gray matter portion of the Mind. The mind you describe I would call mind, small m. The Mind I am referring to is capitol M. I believe it is an error to reduce memory and thought to the calculations that Kurzweil and Alan put forth. Clearly we have had incredibly fast processors, yet we can't even create something that can effectively navigate a room, or talk to me, or reason or completely simulate an ant. How can they reconcile that?? If they sy we don't know how to program that yet. then I say well then stop saying that the singularity is near striclty because of processor speed\memory projections. Processor speed is irrelevant when you have no idea how to use them! It is true that few humans reach this capacity i describe above. I would call them human singularities. Therer have only been a handful in history. But it's important to note that these capabilities are within each of us. I will go as far to say that any computer system we develop, even one that realizes all the promises of the singularity
Re: [agi] How wrong are these numbers?
For instance, in relation to memory capacity. let's say I could live for the age of the universe, roughly 15 billion years. I believe the human mind(without enhancement of any kind) is capable of remembering every detail of every day for that entire lifespan. That is contrary to actual experience. Many of the elderly complain of difficulty in forming new memories. I believe it is an error to reduce memory and thought to the calculations that Kurzweil and Alan put forth. Egad! I'm being compared to Kurzweil the Weenie... =\ The entire point of the entire AGI enterprise is to reduce memory and thought to calculations. Clearly we have had incredibly fast processors, yet we can't even create something that can effectively navigate a room, or talk to me, or reason or completely simulate an ant. All of those are software problems. How can they reconcile that?? If they say we don't know how to program that yet. then I say well then stop saying that the singularity is near striclty because of processor speed\memory projections. Processor speed is irrelevant when you have no idea how to use them! Okay, I have some theories... Unfortunately I'm only a theorist so I'll need some code-slaves to make any progress but I think that's doable. The research machine that I tried to build a few months ago (and is still sitting in pieces) will only be a high end PC. It should be enough to make excelant progress even though it only uses 1.2 ghz processors... It is true that few humans reach this capacity i describe above. I would call them human singularities. Therer have only been a handful in history. Then we'll worry about dealing with the mein intelligence first. ;) But it's important to note that these capabilities are within each of us. As you said, only savants. I am surely not one of them. I will go as far to say that any computer system we develop, even one that realizes all the promises of the singularity, can only match the capacity of the human Mind. Why? Because the universe is the Mind itself, and the computational capacity of the universe is rather immense and cannot be exceeded by something created within its own domain. This is almost theistic... You should check your endorphine levels. -- pain (n): see Linux. http://users.rcn.com/alangrimes/ --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
Re: [agi] How wrong are these numbers?
Ben, As always, thanks for the well thought out reply...I am glad you could make some sense of my ramblings... Just a couple thoughts.. In relation to the subtle consciousness, or store consciousness, I believe it interpenetrates all things equally. So to speak of more or less conscious, from this vantage point, is incorrect. A wooden doll is interpenetrated as well, but is not conscious of it because it lacks the causes and conditions for thought to arise. This does not negate the presence of the store consciousness within it. As far as the connection between the first and the fourth. I think of the store consciousness as a sea of potentialities, When the appropriate causes and conditions are in place, something will become manifest. Your writing a response to me springs from the store consciousness being stimulated by the higher level consciousness. Anger, lust, love and compassion, etc all are potentialities within the SC. It is useful to think of them as seeds. whatever seed is watered, that is what will grow. This is also how species seemingly in disparate locations can seem to operate as a unit. A bird in France figures out how to open milk jugs after the milk man delivers them, and soon after the birds in Kansas are doing it as well... In the case of humans, it can be said that even the simple task of buying a tie is not made without the influence of the collective.. I should state that I do not hold the store consciousness as the absolute substance underlying all things. In fact, it cannot be so, because the store consciousness, although extremely subtle, is itself conditioned and arises only dependently. As such, it is not self existent and impermanent and cannot be the Ultimate suchness of the Universe. And this reaches the limits of my knowledge on the subject... Thanks again for your thoughtful dialog. Here in PA there's no one to talk to about such things. I'm really a marginal character in society for sure :) Kevin - Original Message - From: Ben Goertzel [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, December 03, 2002 8:26 PM Subject: RE: [agi] How wrong are these numbers? Kevin, You raise a critical point, and my thinking on this point is a bit unorthodox, as well as incomplete... There is a big unanswered question close to the heart of my theory of mind, and this is the connection between Firstness and Fourthness. I sum up this question with the Orwell paraphrase All things are conscious, but some things are more conscious than others. I'm a Peircean animist, in the sense that I believe consciousness is everywhere. Yet, I believe that I'm more intensely conscious than a dog, and a dog is more intensely conscious than a flea, which is more intensely conscious than a virus, which is more intensely conscious than a molecule One question is: Why is this? But I'm not even sure of the standpoint from which this question Why? is asked. Another question is: What are the specifics of this law connecting Firstness with degree-of-integrated-complexity (an aspect of Fourthness)? This is something that interests me greatly Along these lines, I believe that if one constructs an AGI with a high degree of general intelligence, ensuing from a high degree of synergetic integrated complexity [the only way to get significant general intelligence, I think], this AGI system *will have* a highly intense degree of consciousness, analogous to (though with a different subjective quality from) that of humans. But I don't have a justification for this belief of mine, because I don't have a solution to the so-called hard problem of consciousness. All I have is an analysis of the hard problem of consciousness, which suggests that it may be possible to create artificial consciousness by creating AGI and watching the intense consciousness come along for free. I suspect that in fact the hard problem will remain in some sense unsolved. That is, the qualitative nature of the connection between intensity of consciousness and degree of general intelligence [for lack of a better phrase] may remain mysterious. Yet, by experimenting with artificial minds, we may learn to quantify this relationship. Quantify it how? Various sorts of artificial minds may report their subjective experiences -- in all sorts of subtle realms of awareness, as well as their own variety of everyday ordinary consciousness -- and we and they may learn rules relating their subjective experiences with known aspects of their physical implementations. Yet, even when such laws are known -- laws relating aspects of the conscious experience of a mind to aspects of its brain -- this will still not resolve the core mystery of how consciousness-mind relates to pattern-mind. But this mystery is ultimately not a question for science, perhaps... and we don't need to articulate its solution in order to build and study minds that are genuinely conscious
Re: [agi] How wrong are these numbers?
For instance, in relation to memory capacity. let's say I could live for the age of the universe, roughly 15 billion years. I believe the human mind(without enhancement of any kind) is capable of remembering every detail of every day for that entire lifespan. That is contrary to actual experience. Many of the elderly complain of difficulty in forming new memories. That is because of a defect in the brain, not the Mind. I believe it is an error to reduce memory and thought to the calculations that Kurzweil and Alan put forth. Egad! I'm being compared to Kurzweil the Weenie... =\ Sorry, but your analysis smacked of his... The entire point of the entire AGI enterprise is to reduce memory and thought to calculations. That's fine, I wish you luck, but I still have that sawbuck in my pocket... Clearly we have had incredibly fast processors, yet we can't even create something that can effectively navigate a room, or talk to me, or reason or completely simulate an ant. All of those are software problems. That's the argument we've heard for some time. I think Ben is closer to anyone in having a true mapping of the brain and its capabilities. As to whether it ultimately develops the emergent qualities we speak of..time will tell...even if it falls short of singularity type hype, i believe it can provide tremendous benefits to humanity, and that's what I care about. How can they reconcile that?? If they say we don't know how to program that yet. then I say well then stop saying that the singularity is near striclty because of processor speed\memory projections. Processor speed is irrelevant when you have no idea how to use them! Okay, I have some theories... Unfortunately I'm only a theorist so I'll need some code-slaves to make any progress but I think that's doable. The research machine that I tried to build a few months ago (and is still sitting in pieces) will only be a high end PC. It should be enough to make excelant progress even though it only uses 1.2 ghz processors... I'm new to AI, but I am reading Norvigs book and one of the first things he says is that what's important is not what you can theorize, its what you can actually **DO**. If you can't encode it...Its just an unproven theory It is true that few humans reach this capacity i describe above. I would call them human singularities. Therer have only been a handful in history. Then we'll worry about dealing with the mein intelligence first. ;) I would suggest this is negligent on your part, but that's your choice.. But it's important to note that these capabilities are within each of us. As you said, only savants. I am surely not one of them. I never said savants. the only reason you and I haven't become a singularity is because we are steeped in delusion and somewhat lazy. I will go as far to say that any computer system we develop, even one that realizes all the promises of the singularity, can only match the capacity of the human Mind. Why? Because the universe is the Mind itself, and the computational capacity of the universe is rather immense and cannot be exceeded by something created within its own domain. This is almost theistic... You should check your endorphine levels. If you read any discussion of the Singularity, its hard to separate what is being said from theism. These machines are given God like qualities and powers. It smacks almost of a new religion with its dogma 1's and 0's, but reeks of the same old idea that i am flawed and weak and small and mortal and I want to be super and sumpremely smart and immortal! I can't blame people for looking for such things as the human condition is a rather sad one Good luck Alan! Kevin -- pain (n): see Linux. http://users.rcn.com/alangrimes/ --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
RE: [agi] How wrong are these numbers?
Kevin wrote: I will go as far to say that any computer system we develop, even one that realizes all the promises of the singularity, can only match the capacity of the human Mind. Why? Because the universe is the Mind itself, and the computational capacity of the universe is rather immense and cannot be exceeded by something created within its own domain. Well... I empathize with your experiential intuition, but this doesn't quite feel right to me. Why doesn't your argument lead also to the conclusion that no computer system can exceed the capacity of the dog Mind? Why is the human Mind special? If you're going to say that the human and dog minds have the same capacity, then I'm going to respond that your definition of capacity is interesting, but misses some aspects of the commonsense notion of the capacity of a mind... I turn again to the Peircean levels. For Mind as First, there is one mind and only one mind, and all minds have the same capacity. For Mind as Third, some minds are more intelligent than others, they hold and can deploy more relationships than others, and this is a meaningful distinction. This is the level on which we are operating as AGI engineers. -- Ben G --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
RE: [agi] How wrong are these numbers?
Kevin wrote: I think Ben is closer to anyone in having a true mapping of the brain and its capabilities. As to whether it ultimately develops the emergent qualities we speak of..time will tell...even if it falls short of singularity type hype, i believe it can provide tremendous benefits to humanity, and that's what I care about. I appreciate your enthusiasm support. I'd like to clarify, however, that I'm not actually trying to map or model the human brain, but ONLY to emulate (and eventually exceed) its capabilities. I have studied neuroscience fairly extensively, but have chosen to make Novamente very unbrainlike in many ways, in order to adapt it better to the available hardware platform. Alan Grimes wrote: Okay, I have some theories... Unfortunately I'm only a theorist so I'll need some code-slaves to make any progress but I think that's doable. Code-slaves, huh? I suggest that you're unlikely to make much progress with this management philosophy ;-) Kevin wrote: If you read any discussion of the Singularity, its hard to separate what is being said from theism. These machines are given God like qualities and powers. It smacks almost of a new religion with its dogma 1's and 0's, but reeks of the same old idea that i am flawed and weak and small and mortal and I want to be super and sumpremely smart and immortal! I can't blame people for looking for such things as the human condition is a rather sad one Well, yeah. I am flawed and weak and small and mortal, and I want to be super and supremely smart and immortal. The interesting thing to come to terms with, is that transforming oneself into a superbeing is in effect a form of *death*. One's current self is really disappearing, if one transforms oneself that completely. What is the thread of identify/awareness that is left, surviving such a transition ??? -- Ben G --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
Re: [agi] How wrong are these numbers?
Boy, I opened a can of worms.. here goes... Kevin wrote: I will go as far to say that any computer system we develop, even one that realizes all the promises of the singularity, can only match the capacity of the human Mind. Why? Because the universe is the Mind itself, and the computational capacity of the universe is rather immense and cannot be exceeded by something created within its own domain. Well... I empathize with your experiential intuition, but this doesn't quite feel right to me. Why doesn't your argument lead also to the conclusion that no computer system can exceed the capacity of the dog Mind? In terms of the Mind, all dualiuties fall away, so dog, human, computer are irrelevant and nothings is bigger or smaller than anything else.. Why is the human Mind special? I don't recall saying it was..But amongst animals, the human, although intrinsically identical with the dog, is capable of directly realizing the Mind. If you're going to say that the human and dog minds have the same capacity, then I'm going to respond that your definition of capacity is interesting, but misses some aspects of the commonsense notion of the capacity of a mind... All things around arise from the Mind including phenomenon, thoughts and other layers of reality and including the subtle consciousness(in my Buddhist lingo: the alaya Vijnana). But the arising and falling is only apparent and like a dream, leaves no stain or trace on Mind itself. I turn again to the Peircean levels. For Mind as First, there is one mind and only one mind, and all minds have the same capacity. For Mind as Third, some minds are more intelligent than others, they hold and can deploy more relationships than others, and this is a meaningful distinction. This is the level on which we are operating as AGI engineers. OK. I am certainly not discouraging that on any level... After all, I may just be confused anyway :) -- Ben G --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED] --- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]