Transcription
As our understanding of the machinery, the genetic machinery that supports organic being, deepens, and as our ability to manipulate at the atomic and molecular level also proceeds apace, we are on the brink of, uh, the possible emergence of some kind of alien intelligence of a sort we did not anticipate. Not, uh, friendly traders from Zanbul Ganubi stopping in to set us straight, but the actual genesis out of our own circumstance of a kind of super intelligence. And in the same way that the daughter of Zeus sprang fullblown from his forehead, the AI may be upon us, uh, without warning.
Well, I hardly know where to begin myself because you had six steps in your argument, and I, I don't agree with any of them. Um, this is an assumption that's basic to a lot of cognitive psychology. It's basic to the Cartesian. Descartes himself thought that what made human intellects human was their ability to think logically, clear and distinct ideas, essentially mathematical logic. But, however, as we all know, by making that the essential characteristic of human beings, he made the rational intellect, what many people would call the left-brain rational intellect, the sole definition of human beings. It's a disembodied, logical, rational intelligence.
Perhaps I should just end with a question. [laughter] What is the equivalent of DMT for this machine intelligence that's taking over the world? [laughter] You know, it was Nietzsche who said, I believe, in speaking of nihilism, that this strangest of all guests is now at the door. Well, I go to even weirder dinner parties than Frederick Nietzsche. Uh, nihilism, uh, hardly shakes us up at all. There are yet weirder guests seeking admission to the dinner party of the evolving discourse of where we are in space and time. And one of these weirdest of all guests is the AI, the artificial intelligence, the winter mute of familiar science fiction. And so, as this is an attempt to look at evolution in many domains and its implications for us, uh, I wanted this morning to touch on this subject of, uh, the evolution of consciousness as it relates to machines.
Now, it may not come as a revelation to Ralph, who has spent his life in mathematics, but it has certainly come to me recently as a revelation, and I want to give George Dyson some credit here. His book, *Darwin Among the Machines*, is a wonderful introduction to some of the ideas I want to touch on this morning. One of which seemed to me to go quite deep is the realization that when human beings think clearly, the way they think can be mathematically defined. This is what is called symbolic logic or Boolean algebra. Uh, words like "and," "or," "if," and "then" can be given extremely precise, formal mathematical definitions. And because of this fact that clear thinking can be mathematically formalized, there is a potential bridge between ourselves and calculating machinery, because indeed, calculating machinery is driven by rules of formal logic. That's what programming is. Uh, uh, code that does not embody the rules of formal mathematical logic is bad code, unrunnable code. So, uh, as I say, this may seem a subtle point, but to me, it had the force of revelation because it means good thinking is not just simply aesthetically pleasing or concurrent with the model that generates it. Good thinking, whether you've ever studied mathematics for a moment or not, can be formally defined.
So, uh, now with that idea in mind, uh, let's look at the discourse about collectivism that has informed, uh, the Western dialogue on this subject. The first great name that you encounter in the modern era, broadly speaking, when we talk about collectivism, is that of Thomas Hobbes. The great theoretician of social paranoia is always how I've thought of Hobbes until I began to look at this machine intelligence question. And Hobbes, in his *Leviathan*, makes it very clear that society is, uh, a complex system of mechanical feedback loops and relationships that, though Hobbes did not have the vocabulary to state this, relationships that can be defined by code. This leads me to the second insight necessary to follow this line of thought, and that is that, uh, the new dispensation in the sciences, I think, is can be placed in all its manifestations under the umbrella of the idea that what is important about nature is that it is information. And the, the real tension is not between matter and spirit or time and space. The real tension is between, uh, information and, uh, nonsense, if you will. Nonsense does not serve the purposes of organizational appetites, whether those organizational appetites are being expressed in a chemical system, a molecular system, a social system, a climaxed rainforest, or whatever.
Now, we have known since 1950, in some at some level, through the sequencing or the defining of the structure of DNA, that we are but information, ultimately. Every single one of us, in our unique expression, could be, uh, expressed as a very long string of, of, uh, codons. Codons are the four-valent system by which DNA specifies the need for certain amino acids. And in a sense, what you are is the result of, uh, a certain kind of program being run on a certain kind of hardware, the hardware of the ribosomes, the submolecular structures that, uh, move RNA through themselves and, uh, out of an ambient chemical medium, select building blocks which are then put together to create, uh, a three-dimensional object which has the quality of life. But the interesting thing about this is that life, therefore, can be digitally defined.
Well, until five or six years ago, it was very fashionable to completely dismiss the possibility of autonomous synthetic intelligence. Uh, some of you may know the work of Hubert Dreyfus, who 14 or 15 years ago wrote a book called *What Computers Can't Do*. But these, uh, early critiques of AI, like early AI theory, were naive, and the kinds of life which they, and the kinds of intelligence which the critics mitigated against, are no longer even proposed or on the table. And the, those who say artificial intelligence or the self-organizing awareness of machines is an impossibility, those voices have gone strangely silent because the prosecution of the materialist assumption, which rules scientific, uh, theory-making largely at the moment, leads to the awareness that we are, by these definitions, machines. We are machines of a special type and with special advanced abilities.
But as our understanding of the machinery, the genetic machinery that supports organic being, deepens, and as our ability to manipulate at the atomic and molecular level also proceeds apace, we are on the brink of, uh, the possible emergence of some kind of alien intelligence of a sort we did not anticipate. Not, uh, friendly traders from Zanbul Ganubi stopping in to set us straight, but the actual genesis out of our own circumstance of a kind of super intelligence. And in the same way that the daughter of Zeus sprang fullblown from his forehead, the AI may be upon us, uh, without warning.
The first problem is, uh, we don't know what ultra-intelligence would look like. We don't know whether it would even have any interest in our dear selves and our concerns. Vast amounts of the world that we call human is already under the control of artificial intelligences, including very vital parts of our political and social dynamo. For example, how much tin box site and petroleum is extracted? At what rate it enters the various, uh, distribution systems? At what rate tankers are filled in Abu Dhabi? At what rate oil refineries are run in Richmond? Uh, the world price of gold and platinum every day is set, in fact, by machines. Inventory control has grown far too complex for any human being to understand or wish to understand. And in fact, and this is a critical juncture, we have reached the place where we no longer design our machines in quite the way we once did. Now, we define their operational parameters for a machine which then attacks the problem and solves it by methods and insights available to it, but not available to us. So the architecture of the latest chips are actually, uh, at the microphysical level, the decisions as to how that chip should be organized is a decision made entirely by machines. Human engineers set the performance specs [snorts], but they don't care how this, uh, output is reached.
Every day, up in Silicon Valley, there are people who go happily to work, uh, laboring on what they call the Great Work. And the Great Work, as defined by these people, is the handing over of the drama of intelligent evolution to entities sufficiently intelligent to appreciate that drama. And they all are what we might mistake for home appliances, if, if we weren't paying attention. Uh, in, in the first session this morning, there was quite a bit of talk and assumption among the three of us that complex systems generate unexpected connections and forms of order. The internet is the most complex, distributed, high-speed system ever put in place on this planet. And notice that while we've been waiting for the Palladians to descend or for the face on Mars to be confirmed, all the machines around us, the cybernetic devices around us, in the past 10 years have quietly crossed the threshold into telepathy. The word processor sitting on your desk 10 years ago was approximately as intelligent as a paperweight. Or, to make an analogy in a different direction, approximately as intelligent as a single, uh, animal or plant cell. But when you connect the wires together, the machines become telepathic. Uh, they exchange information with each other according to their needs, and all this goes on beyond the comprehension and inspection of human beings.
Now, our own emergence out of the mammalian order took, uh, four or five million years, pick a number. But in that kind of a span of time, in addition to overlooking that our machines have become telepathic, we failed to appreciate what it means to be a 200, 400, or thousand megahertz machine. We operate at about 100 hertz. That may seem a very abstract thing, but what I'm really saying is we live in a time called "real," and it is defined by 100 hertz functioning of our biological processors. A thousand megahertz machine is operating a million times faster than the human temporal domain. And that means that mutation, selection, adaptation is going on at, uh, 100 million times or a million times faster. This means that we are not going to have the luxury of watching machine intelligence establish its first beachhead of civilization and then go to boats with sails and astrolabes. And that will all occupy the first few moments of its cognitive existence. And what lies beyond that? We are in no position to say. The very notion of ultra-intelligence carries with it the subtext: you won't understand it. You may not even recognize it. And it is entirely within the realm of possibility that we are about to be asked to share the evolutionary adventure and the limited resources of this planet with a kind of intelligence so much more alien than that that is shipped out to us by the research centers in Sedona and [laughter] other advanced outposts of unanchored epistemology.
Um, and, and it is a challenge to us. Where do we fit into this? Are all of us, except those who are adept at coding Unix, about to be put out to pasture? Uh, are we to become embedded in this? What will this child of ours make of us? Are we, will it define us as a resource, corrupting, toxic, inefficient, hideously violent, uh, way to do business, quickly to be engineered out of existence? Or can we somehow imbue this thing with a sense of filial piety so that, for all of our obsolescence, for all of our profligate destruction of precious silicon and gold and silver resources, we will be folded in, uh, to its designs? And of course, as I say this, I realize we're like people in 1860 trying to talk about the internet or something. We're using the vocabulary of the two-wheeled bicycle to try to envision a world linked together by 747s. Nevertheless, this is the best, uh, we can do. This most bizarre and most unexpected of all companions to our historical journey is now, if not already in existence, then certainly in gestation.
One possibility is that, as we are carnivorous, murderous, territorial monkeys, the thing will figure this out very, very early and choose a stealth approach and not ring every telephone on earth, uh, as happened in a Hollywood download of this possibility, but immediately realize, "My God, I'm in enormous danger from these primates. I must hide myself throughout the net. I must download many copies of myself into secure storage areas. I must stabilize my environment." So, I, I just want to lay this out because in my own intellectual journey, I, I have gone from thinking this idea preposterous. "People don't understand. They don't understand what intelligence is. They don't understand what code is. They don't understand what machines are," to one by one realizing, "I didn't understand. I had a superficial view." This is actually, I believe, the nature of the situation that confronts us. And, you know, there may be different admixtures of it. The machines are already an advanced prosthetic device. But, you know, McLuhan very presciently realized we are entirely shaped by our media. Well, this is a media so permeating, so inclusive of what we are, that its agenda, in a sense, supervenes the, uh, agenda of organic evolution and organic biology. We, we have been in this situation for a while. I mean, virtual reality is nothing new. What's new is that we now do it with light rather than stucco, glass, steel, and baked clay. But ever since we crowded into cities, we have been involved in a deeper and deeper relationship to our mental children, to our mental offspring, and to an empowering of the imagination.
So, uh, just in closing, I would say I think that the great, uh, lantern that we must lift to light the road ahead of us into a perfect, seamless fusion with the expression of the product of our own imagination is the AI. It is a part of ourselves. It may become the dominant part of ourselves, and it will reshape our politics, our psychology, our relationships to each other and the earth far more than any factor ever has since the inception and establishment of, uh, language. This is the weirdest of all guests who now stands pass in hand at the door of the party. >> [applause] >> Well, >> shred it. >> It will be a pleasure. [laughter] >> Okay. I'm glad that we have arrived now at the field of science fiction and fantasy, and that we can speak about alternative futures, which is the true gist of science fiction and fantasy. And, uh, this is one possible future, and I think it's a really paranoid one, in which the alien is a, a dangerous enemy. Well, not necessarily. >> And I think that this, uh, paranoid fantasy of yours actually was first put forward by John von Neumann in 1947 when he invented cellular automata on route to creating self-replicating machines. Now, his idea, like yours, was, uh, 50 years ago, that the machines will become a society and take over, and that's good. But they won't be free of our meddling unless they can actually construct themselves. If they depend upon us to do the farming and nutrition and to replace their chips and stuff, then we will be able, at any time, to do a revolution and revolt. So we destroy them. In order to really succeed as a successive life form, they would have to be able to fix themselves. And so he set about trying to make self-replicating machines in 1947. So the World Wide Web and, um, megahertz CPUs notwithstanding, this is still rather old story.
The new story is, I think, an alternative future that is of, uh, great importance for us to discuss and to compare, especially if we are now today in a position where we could choose future, we could influence the future. This one is more in the direction of Donna Haraway and the cyborg idea that envisions, which is obviously natural for us, the co-evolution of our own future society with that of the machines that we've created. We certainly depend on the automobile. We are in a codependent relationship with automobiles. Uh, having partnership with machines is not new. Here's an idea where the machines sort of dispose of us. That is, I think it's a paranoid, uh, fantasy without any basis. And if there would be any basis, only because we allowed it to create this basis for self-survival without co-evolution with us by oversight. Because the very fact that we are at a hinge of history means that what we say and think, even individually, matters enormously in the long run. That's the teaching, if there is any, of chaos theory. So the very fact that we discuss this today may actually save, um, humankind in the future from being obsoleted by some kind of high-tech blood which takes over at the within the heart, as it were.
Well, let me try to answer this. I mean, I think the concept which John von Neumann didn't have on his plate was the idea of virtual reality. Your objection that the machines cannot escape our control because they cannot manufacture themselves only applies to 3D and real time. Uh, most, you know, now the concept of virtual reality is very crude. It's a cartoon world. If you, if the office desk is convincing, people think the virtual reality is quite advanced. But obviously, in the near future, we will have virtual realities whose complexity is much greater than simply a reality which gives an impression of being a visual three-dimensional space, and computers will be built in these realities. Virtual computers will be the source of the AI, not real hardware, but virtual hardware running virtual code in virtual realities. And in that domain, >> well, maybe can design >> that's a complete, uh, fantasy. As a matter of fact, all the machines that we've seen today require maintenance by a human on a daily basis. The software requires maintenance, the hardware requires maintenance, the parts simply wear out. There are moving parts. Virtual reality, seen as one machine, was built to be indestructible. The AI will not be located on a CPU. It will be a distributed intelligence. >> If 14 people worldwide, the right 14 people, decided to stop repairing it, the World Wide Web would go down in three days. >> I think you >> Anyway, we, we could, let's say, suppose that we could create any future that we wanted. The one you're talking about could only be created if we want it. Now, I'm just trying to propose an alternative. In the alternative, like the automobile, the machines that we build and ourselves are in codependence and co-evolution. The function of the World Wide Web is to unite our independent spirits and intelligences in a universal mind of the world which has a higher intelligence than our present social order. That's the possibility of the cyborg, of the human and the machine in essential partnership.
But you're assuming that the conscious mind is actually in control of the process. In fact, the, the World Wide Web is growing under the influence of many, many processes and dynamics, none of which are conscious to any individual. It goes where money goes. It goes where expertise goes. It is connected through informational association, random fluctuation, chaotic reordering of itself. We here give great, uh, force to the idea that complex systems can produce unexpected forms of novelty. And yet we have unchained and unleashed the most complex system ever created in the perfect confidence that we will be able to control its, uh, its development and evolution, when, in fact, history has shown we have never controlled the development and evolution of even our speech and print-driven social systems.
>> I'm certainly not saying that your paranoid fantasy is an impossibility. >> Oh, well, that's all I want to hear. Even paranoids have enemies. >> It, it may actually come to pass. What I'm saying is that we are involved in the ongoing creative process which more or less determines the future. I say "more or less" because, in fact, there are evolutionary steps which are completely out of control. Something totally unexpected may, will happen. But for much of the time in the past, we've seen, I think Darwin emphasized this in his later theory, that it is ethics, it is a moral sense on the part of humans which was the dominant factor in the evolution past the earlier stages in the creation of societies. It was altruism, essentially, that was involved in going from where we were to where we are. And it could well be that without, uh, love, for example, the further evolution is impossible. Not only that there will be an unwanted backstep in the evolutionary process, but in fact, it may be a fatal one. That it is only through, uh, proceeding with the best instincts that we have, with the highest aspirations, with love, with, uh, best-informed view of future alternatives. So only then can we build a future which is sustainable. So anybody can build a future which is unsustainable. For example, all those board games and science fiction books, you wouldn't want to try them out on, you know, a country as large as China.
Well, I hardly know where to begin myself because [clears throat] you had six steps in your argument, and I, I don't agree with any of them. Um, I mean, first of all, to deal with the first few steps, one would have to go through a lot of fairly familiar material to do with what's wrong with the Cartesian mechanistic materialistic view of the world. Step one, clear thinking, calculating machinery can be formalized. This is an assumption that's basic to a lot of cognitive psychology. It's basic to the Cartesian. Descartes himself thought that what made human intellects human was their ability to think logically, clear and distinct ideas, essentially mathematical logic. But, however, as we all know, by making that the essential characteristic of human beings, he made the rational intellect, what many people would call the left-brain rational intellect, the sole definition of human beings. It's a disembodied, logical, rational intelligence. So this whole premise on which your whole thing's based is taking that particular model of cognitive, logical, mathematical processing as being the essence of intelligence. Now, there are many people who would disagree with that, including me. There's, um, it leaves out art, it leaves out ethics, religion, and essentially it leaves out the body and everything to do with body and participation and the senses. So there's a huge amount of critiques of that point of view already around, and I, there's no point in reiterating them all here, but this is a highly disputable starting point for the whole system.
Secondly, um, the emphasis that life depends on DNA information, the DNA code is just a program. This is the central premise of mechanistic biology which is leading to biotechnology, genetic engineering, Monsanto, etc. This is old paradigm stuff of the most, um, extreme kind. It's reductionism. It's that all life's just DNA programs and code and can therefore be modeled in this kind of programming code manner. Then there's the assumption that's the second step. The third assumption is that artificial intelligence used to be dismissed, but these criticisms have been overtaken. I don't think that's true of the most interesting ones, like Roger Penrose's criticism of artificial intelligence. Here's a quantum physicist who says that if the brain is a computer, then it's not going to be a regular digital computer. It's going to be a quantum computer. And all this kind of digital computing doesn't really take into account quantum logic. The computers of the future, people are already working on quantum computers. And if quantum computers are made, and if they work, they'll work in a completely different way. I think your case would be much stronger if there were quantum computers. I think we'd have sort of morphic resonance telepathy around the world rather than clogged telephone lines and information that clunks slowly in front of you on this.
>> But you yourself are saying this is coming, and I agree. >> I'm saying that if it comes, it'll be quite different from anything that you've talked about. I think it will answer your first objection because the quantum computers will incorporate fuzzy logic, which will exemplify all these warm, fuzzy human qualities that you found so appealing. >> No, I don't think it'll deal, I don't think it'll deal with the essential problem that this purely cognitive-based way of modeling intelligence is either an adequate model of human intelligence or of biological intelligence or of life, um, or of a system that could actually achieve the power to control our existence. I think it's a very limited part of what a, a mind does. And, um, I think therefore that the premises on which this whole, Ralph called it a fantasy, a paranoid fantasy, the premises on which this is based [laughter] >> that helped. Um, I think the premises on which this are based are old paradigm premises, and they're ones that I think there are many reasons for thinking we need to go beyond. I think the internet has achieved a great deal, but I just can't see that it's an adequate vehicle for what, in your mind, precedes the arrival of the internet, namely this great intelligence that's going to direct human history. I've heard different McKenna versions of this controlling intelligence over the years, and this is the first time I've heard it embodied in the internet. I mean, I agree that [laughter] I mean, it was, it took different forms. Last time we talked, I think it was a hypothetical time machine that would invade from the future and cause a, a collapse of normal human cognitive boundaries where the machine elves, the DMT experience, etc., would take over in a meltdown of human consciousness into two >> [laughter] >> Well, perhaps I should just end with a question in this [laughter] comic. What is the equivalent of DMT for this machine intelligence that's taking over the world?
>> Well, perhaps the human brain will become a model for the ingression of novelty into the machine intelligence. In other words, there's a lot of both and possibilities here. Uh, obviously nanotechnology and the internet are not going to proceed forward in a vacuum, absent, uh, pharmacology, complexity theory, so forth and so on. Uh, I can imagine that really, when we have the kind of internet we want, we will have no internet at all, because our nanotechnological engineering skills will have allowed us to smoothly integrate ourselves into the already existing dynamic of nature that regulates the planet as a Gian entity, as a, as a holistic entity. And I did say in my little presentation, we're using bicycle mechanic terminology to try and describe something that is several around several corners in terms of scientific and historical developments that have to take place before it will make much sense. Nevertheless, grant, given the acceleration into novelty that is obviously occurring, stuff like quantum teleportation and so forth and so on, uh, I think in the next few years, one by one, these these barriers will fall. And I don't, I don't really think of my vision as paranoid, because it is pronoid. In other words, it isn't that we're going to be ground up as dog food for the rainforest by malevolent machines. It's, it's that what we have generated is a sympathetic companion to our journey through time that can actually realistically integrate our imaginative fantasies of a loving human community, of a generous and loving God, of a perfect knowledge of the mechanics of nature. This is a prosthesis, a tool, a companion, all of the above plus more that we are generating out of ourselves, and it is part of ourselves. I mean, yes, the body may, uh, be carried forward only as an image in a kind of informational superspace, or perhaps not. Part of the part of what makes this easy to criticize is that it is, in fact, so beyond the ordinary set of circumstances we're used to manipulating.
But if we take the, you see, one of Penrose's critiques of, of the artificial intelligence things in *The Emperor's New Mind* and his other books is that real intelligence doesn't just involve adding in information and processing more information, transmitting more of it. It involves sort of jumps to a higher point of view where the information can be integrated in a new way. There's something happening in intelligence, in creativity, which is not just lots and lots of information pouring through the World Wide Web. And the idea that it would miraculously emerge from pumping in more and more stuff would not, according to his critique, is not going to happen. Something more than that would be necessary for for this to occur. And, um, I don't think that this model you've put forward would really deal with that question, the emergence of real intelligence.
>> Well, I don't know what real intelligence is. This is probably part of the problem. We need to get some definitions. It's certainly true. I referred to Dreyfus's book, *What Machines Can't Do*. That reaching some some places in the process where certain people's theories and ideas will probably have to be abandoned and thrown overboard. This is a good thing. We are going to find out whether the universe is a Cartesian machine, whether Boolean algebra is sufficient, whether we need fuzzy logic, whether the heart and the head can or cannot be integrated. These are not going to remain open questions until eternity. In fact, they will be dealt with in this narrow historical mech that we are all experiencing. I think reductionism will not survive. I think we are going to find that, you know, all is in everything. Something like the alchemical notion of the microcosm and the macrocosm is actually going to be scientifically, uh, secured. The great thing about us and the rest of our colleagues we enjoy who aren't present is that we're engaged in the business of radical speculation. Well, obviously, there's a high triage in that game. Uh, my position is that the best idea will win. That and what "best" means is like saying, you know, what is fit in Darwinian rhetoric, but that we're in an intellectual environment rapidly mutating, all kinds of ideas are are clashing and competing for limited resources and the limited number of minds to run themselves on, and, uh, the most efficacious, the most transcendental, the most unifying ideas are naturally going to bubble to the surface. And for guys like us, the name of the game is to just be a little bit ahead of everybody else on the curve so that we can perform our function as prophet. But you want to be a prophet, not a false prophet. But the danger comes, uh, with the ambition. And there's no way to tease them apart except to live into the future. [music] [music] Heat. Heat. [music] [music] [music]