The Technium

Inevitable Minds


The rock ant is tiny, even for an ant. Individually each ant is the size of comma on this page. Their colonies are small too. Numbering about 100 workers, plus one queen, they normally nest between slivers of crumbling rock, hence their common name. Their entire society can fit into the glass case of a watch, or between the one-inch covers of a microscope slide, which is where they are usually bred in laboratories. The brain of a rock ant contains less than 100,000 neurons and is so small as to be invisible. Yet an rock ant mind can perform an amazing feat of calculation. To assess the potential of a new nesting site, rock ants will measure the dimensions of the room in total darkness and then calculate – and that is the proper word – the volume and desirability of it. For many millions of years, rock ants have used a mathematical trick that was only discovered by humans in 1733. Rock ants can estimate the volume of a space, even an irregular shaped one, by randomly laying a scent trail across the floor of the space, “recording” the length of that line, and then counting the number of times it encounters that scented line during additional diagonal runs across the floor. The calculated area is inversely proportional to the frequency of intersections times length. In other words, the ants discovered an approximate value for Π derived by intersecting diagonals. Headroom is measured by the ants with their bodies and then “multiplied” with the area to give an approximate volume of their hole. 

But these incredible tiny ant minds do more. They measure the width and number of entrances, the amount of light, the proximity of neighbors, and the degree of hygiene for the room. Then they tally these variables and calculate a desirability score for the potential nest by a process that resembles a “weighted additive” fuzzy logic formula in computer science. All in 100,000 neurons.

The minds of animals are legion, and even fairly dumb ones can yield amazement. Asian elephants will strip away branches to construct a fly-switch to keep pesky flies away from their hind parts. Beavers, mere rodents, have been known to stockpile construction materials before starting to build their dam, thus displaying the ability to anticipate a future intent. Beavers can even outwit humans trying to prevent their dams from flooding fields. Clever Homo sapiens will install ingenious “beaver baffler” drainage tubes upstream to siphon off water lower than the dam level, but the rodents can somehow figure this out and engineer ways to plug the intake – a modern need that would not ordinarily come about in nature.  Squirrels, another rodent, continually outwit very smart human suburbanites over control of their backyard birdfeeders. (I’ve been battling my own black squirrel Einstein.) Biologists have witnessed hawks submerging hard-to-kill prey underwater till they drown. The honeyguide bird in Kenya lures humans to wild bee nests so that the birds can feast on the remaining bee brood after the humans remove the honey; sometimes according to researchers the honeyguide will “deceive” the hunters of the actual distance to a deep forest nest if it is more than 2 kilometers away.

Asked what he could infer about nature’s Creator, biologist J. B. S. Haldane reputedly concluded that He must have “an inordinate fondness for beetles.”  But far more than beetles, nature displays an inordinate fondness for minds. Since every species of insect, and every animal, yields a mind, however limited, there are more minds, and more varieties of minds than beetles. Nature has a fundamental affinity, if not fondness, for intelligence.

Plants, too, posses a decentralized type of intelligence. As Anthony Trewavas argues in his remarkable paper, “Aspects of Plant Intelligence,” plants demonstrate a slow version of problem solving that fit most of our definitions of animal intelligence. They perceive their environment in great detail, they assess threats and competition, then they take action to either adapt or remedy the problems, and anticipate future states. Time-lapse motion pictures that speed up the action of vine tendrils probing their neighborhood make it clear that plants are closer to animals in their behavior than our fast lives permit us to see. Charles Darwin may have been the first to observe this. He wrote in 1822, “It is hardly an exaggeration to say that the tip of the root acts like the brain of one of the lower animals.” Like sensitive fingers, roots will caress the soil, seeking out moisture and nutrients much as a nose or trunk of a herbivore might dig in the earth.

Plants share with animals an almost mathematical ability to optimize their energy efficiency while gathering the most nutrients for the least effort. Plant and animal “foraging” models are almost identical.  Roots search for fertile areas while avoiding adversarial competitors. A distant rootlet can also recognize another rootlet as its own among many the underground tangle of roots, even when all the neighboring plants are genetic clones. Thus, says Trewavas, “Individual plants are able to distinguish self from non-self.”

Plants are in constant motion because their world is in motion. The microclimates around a plant vary by height above ground. The density of carbon dioxide and sunlight can vary by the minute as wind blows and shade shifts. Available nutrients vary by the day as other plants thrive nearby. Growth factors vary by the seasons and temperature, and they vary by the decade as the ecological progression changes. A plant’s entire shape, metabolism, and behavior will thus change by the minute as well, as if it is being governed by a nervous system. It can quickly fill its green leaves with toxins or anti-fungal pesticides to retaliate if animals or parasites munch on it. As competing plants invade its territory a plant can alter the orientation and the structure of its parts by deflating the stomata on one side of its leaves to bend them in certain directions. A plant can thus “see” its environment and move in response to this sight. Some plants walk across a landscape, branch bent to root raised to branch again, on the scale of decades. The ability of a leaf to follow the sun (heliotropism) to gain optimal light exposure can be replicated in a machine, but only by using a fairly sophisticated computer chip as a brain. A plant thinks without a brain. It uses a vast network of transducing molecular signals instead of electronic nerves to carry and process information.

Remarkably a plant mind also contains a memory. There are numerous examples of plants remembering signals for days, and even years. A memory is a way to move information from the past into the present. Plants can also move information from the present into future, or anticipate, which is a true mark of the most primitive intelligence. When a nearby competing plant dies, opening up previous shade to light, some plants send out “exploratory speculative growth” in order to test the new zone with small investments of chlorophyll before spending big time on expensive branches and leaves. Light reflected from nearby vegetation is richer in far-red wavelengths than unreflected light. Plants can use this information to not only see shade, but to anticipate the likelihood of shading by a competitor in the future. “When a change in the balance of red to far-red radiation is perceived,” says Trewavas, “an integrated adaptive response in phenotype structure [of the plant] results. New branches grow away from the putative competitor, stem growth is increased; the rate of branching diminishes, and such branches assume a more vertical direction: leaf area increases in anticipation of reduced incident flux; and the number of layers of leaf cells containing chlorophyll diminishes.”

Plants exhibit all the characteristics of intelligence, except they do it without a centralized brain, and in slow motion. Decentralized minds and slow minds are actually quite common in nature, and occur at many levels throughout the six kingdoms of life. A slime mold colony can solve the shortest distance to food in a maze, much like a rat. The animal immune system, whose primary purpose is to distinguish between self and non-self, retains a memory of outside antigens it has encountered in the past. It learns in a darwinian process, and in a sense also anticipates future variations of antigens. And throughout the animal kingdom collective intelligence is expressed in hundreds of ways, including the famous hive minds of social insects.

The manipulation, storage, and processing of information is a central theme of life. Learning erupts over and over again in the history of evolution, as if it were a force waiting to be released. A charismatic version of intelligence – the kind of anthropomorphic smartness we associate with apes – evolved not just in primates, but in at least two other unrelated taxon: in the whales, and birds.

Stories of dolphin intelligence are famous. Dolphins and whales not only demonstrate intelligence, they occasionally give hints they share a style of intelligence with us hairless apes. For instance, captive dolphins have been known to train each other. Yet the most recent common ancestor for apes, whales, dolphins was 250 million years ago. In between apes and dolphins are many families of animals without this variety of thought. We can only surmise that this style of intelligence evolved independently.

The same can be said for birds. Measured by their intelligence, crows, ravens and parrots are the “primates” of birds. The size of their forebrain is as relatively large as non-human apes, and the ratio of their brain weight to body weight is in the same line as apes. Like primates, crows live long and in complex social groups. New Caledonian crows, like chimpanzees, craft tiny spears to fish for grubs in crevices. Sometimes they save the manufactured spears and carry them around. In experiments with scrub-jays, researchers discovered that that jays would rehide their food later if another bird was watching them when they first hid it, but only if the jays had been robbed before. Naturalist David Quammen suggests that crow and raven behavior is so clever and peculiar that they should be evaluated “not by an ornithologist but by a psychiatrist.”  The famous African grey parrot Alex was taught to name colors, size, and shapes, and to put together simple spoken phrases. When questioned about novel objects he had not seen before, he could give correct answers about 80% of the time. He could also count to six.

The common ancestor for birds, whales and apes was 280 million years ago. The vast majority of smart animals lie between these three very anthropomorphic taxon. Thus, charismatic intelligence evolved independently three times: in birds on wing, in mammals that returned to the sea, and in primates. It may have even evolved a fourth time. Birds are the only dinosaurs that survived the great dinosaur wipeout 65 million years ago. But it is very probable that before they disappeared large dinosaurs were way ahead of archaic mammals (typically no bigger than gophers) on the race to reach complex intelligence. Because birds today with their small brains can surprise us, dinosaurs with much larger brains may have been as smart as apes. Had dinos not vanished under the assault of the heavens, consciousness might have been birthed on earth in a highly evolved reptile, rather than a mammal. We can easily speculate about an alternative world where Saurians ran the place. A few years ago palaeontologist Dale Russell sketched out what Dinoman could have looked like. He would be warm-blooded, 1.33 meters tall, and weigh 32 kilograms.

David3-1

Smartness is a competitive advantage everywhere. We see the widespread recurrence and reinvention of intelligence because the living universe is a place where learning makes a difference.

Organisms are so highly evolved to their respective ecological niches that in terms of variation almost every part of a creature is unique to that species. A good biologist, for instance, can identify an animal from its teeth alone. For the same reasons animal minds also vary by species to fit that animal’s livelihood. There are hundreds of types of learning in the animal kingdom.  Rats excel at the expert spatial intelligence they need to navigate a maze, while their cousins the guinea pigs fail in that department. Humming birds excel at timing, primates in social intelligence. If we knew enough about animal minds a good biologist could identify an animal solely from its species of intelligence.

What we call “intelligence” is in reality a suite of different specialty learning programs, and each species of intelligence is a different mix of those sub programs. One species of intelligence will dial up counting, while another dials up long-tem memory, or neglects social intelligence. Yet, a few animal behaviorists claim that to a rough order of magnitude, all animal intelligence shares a universal core. “There are no qualitative differences in cognition between animal species in the processes of learning and memory,” says Euan Macphail. “Pigeon, rat, monkey… doesn’t matter. Once you have allowed for differences in the ways in which they make contact with the environment…what remains of their behavior shows astonishingly similar properties,” says B.F. Skinner. “There are no consistent differences on such tasks between the performance of goldfish, pigeons, and chimpanzees,” says David McFarland. “…There’s an impressive similarity in basic associative learning among diverse species,” says Leslie Real. Edward Thorndike, one of the first animal behaviorists, was convinced the universal mind extended to invertebrates as well. He saw a common intelligence in honey bees and octopus. He believed that “octopus show precisely the type of progressive improvement in reversal problems [learn one pattern then reverse it] that the rat and monkey show.” In fact there is one report of octopuses learning a task while watching another octopus be trained (and rewarded). A fair number of scientists have tested lowly planarians with a full battery of IQ tests. Planarians are flat worms, tiny slugs. They turn out to be smarter than anyone would have guessed.

The problem in gauging the distribution of intelligence is that we can’t use the same test for every animal. You can’t measure a worm’s intelligence the same way you measure a walrus’s. Their niches, and input/output systems, are so different no uniform test is practical. When you devise a test that is suitable for a particular creature – say a maze for a rat, or levers for a pigeon — what you are really measuring is that animal’s specialty intelligence which has been evolved for its environment. Beneath this mix of specialty intelligences runs a more universal mind. All creatures seem to share a core mind based on elemental learning, memory, and decision-making routines, but with different operational characteristics for different inputs and environment.

Up and down the six kingdoms of life, minds have evolved many times. So many times, in fact, that minds seem inevitable. Yet, as inordinately fond as nature is of minds, the technium, or the seventh kingdom of life, is even more so. The technium is biased to birth minds. All the inventions we have constructed to assist our own minds – our many storage devices, signal processing, flows of information, and  distributed communication networks, – all these are also the essential ingredients for producing new minds. And  so new minds spawn in the technium in inordinate degrees.

Technology is anything a mind makes. Built by minds, the technium is primed to make more minds. These mind children will be small, dim, and dumb at first, but tiny minds keep getting better. And more abundant. Last year there were 1 billion electronic brains etched into silicon. Many contained a billion transistors each but the smallest had a minimum of 100,000 transistors, about as many neurons as the brain of the rock ant. They, too, can do surprising feats. Tiny synthetic ant-minds know where on earth they are (GPS), and how to get back to your home, and remember the names of your friends, and translate foreign languages. These dim minds are finding their way into everything: shoes, door knobs, books, lamps, pets, beds, clothes, cars, light switches, kitchen appliances and toys.

We are blind to this massive eruption of minds in the technium because humans have a chauvinistic bias towards any kind of intelligence that does not precisely mirror a human’s. Unless an artificial mind behaves exactly like a human, we don’t count it as intelligent. Sometimes we dismiss it by calling it “machine learning.” So while we aren’t looking, billions of tiny minds, on the scale of biology, have blossomed in the technium.

One of the smartest artificial minds going is a website that plays the classic game of twenty questions. Nobody, including the inventors, calls it an artificial mind. In this game you think of something and the software mind in the website’s servers, called 20Q, will guess what you are thinking of after 20 yes-no questions.  20Q will guess right 80% of the time, and if you let it go on to ask 25 questions, it will nail what you are thinking 98% of the time.  With 10 million synaptic connections, 20Q is about as smart as a house fly – if a fly spent its brain power on answering questions instead of flying. 20Q has been trained to think by people playing the game. It has been played, or trained, over 70 million times so far. It has been learning since 1988. No other AI has lived so long. 20Q’s bottom-up training produces answers that reflect what people in general think. If most people believe a book has a spine, it does too. If most people believe “blue” is something you can wear, it is. What 20Q believes can change day to day. Creator Robin Burgener says, “It has a personality. Some days it does well, and some days, it’s just off.” There are a dozen different language-specific 20Qs, so the questions play out differently in different languages.  A smaller version of the American/English mind has been extracted and squeezed into a tiny chip inside a plastic toy orb about the size of a snow globe. Some 35 million of these handheld micro-minds have been sold. This autonomous 20Q knows only 2,000 objects, but its still a marvel. Most children can’t stump it. Because it is so small, so self-contained (it sits in your hand like a sea urchin), its intelligence is shocking to the unprepared. For about ten dollars you can get a tiny insect mind which, in one narrow specialty, is smarter than you are.

The cramped, autistic mind of a $10 digital calculator is another species of intelligence, currently produced in the hundreds of millions. It also is an alien intelligence smarter than you — in mathematics.  Billions of insect-like artificial minds have spawned deep into the technium doing invisible, low-profile chores like reliably detecting credit-card fraud, or filtering email spam, or reading text from documents. These proliferating micro-minds run speech recognition on the phone, assist in crucial medical diagnosis, and aid stock market analysis, power fuzzy logic appliances, and guide automatic gearshifts and brakes in cars. A few experimental minds can even drive a car autonomously for a hundred miles.

Two arenas where artificial minds are flourishing without disguise is in multi-player online games, and in web search. In game’s artificial intelligences (AI) can inhabit non-human players in order to make the game more exciting or challenging. Other types of game AIs (such as those in Left 4 Dead) work behind the scenes creating dynamic narratives for human players. The artificial tiny minds generate new monsters, sounds, or actions based on a player’s previous game. Or they place enemies based on each player’s location, status, and skill. Some game AIs (as in Spore) will complete your actions in a smart way.

The artificial intelligence behind internet search does not pretend – at the moment – to be human. In fact, web AI’s main attraction is its ability to find material in a way no human could –- by remembering everything.  However, in the near future services which can understand queries asked in ordinary language will make the web’s non-human intelligence seem more human. Then the AI will feel less alien, and more familiar.

Unlike the billions of minds in the wild, the minds of the technium are  getting smarter by the year. There seems to be no place a mind can’t be born, or inserted. Technology wants to produce minds.

The daily grinding of evolution, as accelerated by technology, churns out more and more complex organisms, with higher rates of energy use, and with increasing specialization. Minds are the ideal way to express complexity, energy density, increasing specialization, expanding diversity — all in one system. Mindedness is what evolution produces. Mindedness is what technology wants, too.




Comments
  • Bob Welch

    As David R. Hawkins might say; “Gloria In Excelsis Deo” ;-)

  • Bill Hunt

    Another fantastic production, Kevin. Thanks!

  • Mark Essel

    the Mind: the seat of the faculty of reason

    How curious that you can perceive it in so much of life, and now the extraordinary environment of our technology. Is the mind the conscious base of perception, or is it characterized by it’s ability to process information.

    AI as I imagine it, will be fully aware of it’s existence. Although it may take us some time to share in that “awareness”.

    Since life has struggled to optimize efficient survival no matter what it’s environment over billions of years we can expect artificial life within the technium to optimize it’s survival on Earth as well.

  • Nova Spivack

    One of your more beautiful, and important, articles, Kevin.

  • Doug King

    Kevin, you are hitting your stride.

    Glad to see you acknowledging the importance of the 20 Questions game. I have been pointing this out as the first example of the future AI of the web to whomever would listen.

    Now, take a look at GWAP, Games With a Purpose http://tinyurl.com/dlelsk. This will be the next wave, and the trainers will be anyone that enjoys the interaction. It will be up to the ‘game’ designer to implement the layer that entices humans into contributing their ‘work’, which will consist of teaching the AI.

    What we will need ASAP is a type of game that encapsulates the rules for the best human values, so we will have some sort of Asimov’s laws of robotics that will be necessary as a governor for future strong AI.

  • Ralph Weidner

    Kevin,

    You bring a fascinating new perspective to your subject I’ve never seen before. Right now I’m struggling with how to accommodate it and my own, in many ways more conventional perspective.

    For example, in this blog I understand you to be arguing not only that mind is common to both the born (life) and the made (the technium), but all minds are essentially on the same plane: we can’t say that one is inherently better than any other, in contradiction, for example, to the common assumption, which I hold, that the human mind stands at the apex of evolution.

    I like something Ken Wilber has said in this respect (as I understand him): in comparing various (sentient) beings, there are three types of values to consider: ground value, extrinsic value and intrinsic value. All beings are of value in themselves–they have equal ground value, are on the same plane in this respect. However, they typically have different extrinsic and intrinsic value. These two, though, tend to balance each other out, i.e. a being with less extrinsic value than another is apt to have more intrinsic value.

    These last two types of values are the result of evolution. If one thing arises from another in evolution, for instance molecules from atoms, then it is said to be more significant, and to have more intrinsic value, while its predecessor is said to be more fundamental, and to have more extrinsic value. In this case, atoms are said to be more fundamental and to have more extrinsic value than molecules, which are said to be more significant and to have more intrinsic value. We can say here that molecules exist on a higher plane than atoms, because they not only can they do essentially all the things an atom can do (e.g. absorb and emit light) because they include atoms as their parts, but more significant actions such as a myriad of chemical reactions.

    But if a molecule, say, strays too close to the sun, the extraordinary heat will cause it to come apart and to die, but its constituent atoms will do just fine. In contrast, when U235 is put into a nuclear reactor and its atoms disintegrate, the molecular form of U235 (I don’t know what it is, but it must have some molecular form) necessarily bites the dust, too.

    So, while we humans may pride ourselves on how significant we are, we are at the same time the least fundamental. We fear that our excessive polluting of Gaia (the procaryotes) might cause it to die, and it’s true that if Gaia were to die, since it is the most fundamental life form on Earth, all other life, including ourselves, would vanish. But we might just kill off all multi-cellular, nonhuman life, say, which would spell our demise since it is more fundamental than us, but Gaia (as well as heterotrophic ecosystems of eucaryotes, perhaps), with a little adjusting, would do just fine, just as it did before multi-cellularity arose, because it is yet more fundamental–gut bacteria probably wouldn’t make it, but there are many other types of bacteria that would.

    (I’m sorry this argument is becoming so long-winded, but I’d hate to have to stop now.) Of course, humanity is not in a direct line of descent with rock ants, so their comparison becomes more complicated, so it’s best, I guess, not to go into this and, instead, limit myself to more circumstantial evidence for humans being more significant than other organisms, or groups of organisms such as a hive of bees.

    The examples you give of animal and machine intelligence strike me as being the result of specialization. humanity specializes, too, but never (I guess this is open to debate) as an entire species, i.e. not at the genomic (biological) level, rather at the more significant level of culture, i.e. the technium, which is much more flexible: I can over-specialize in some direction, but if so, the next generation will probably be smart enough not to follow my path, whereas if a colony of rock ants ends up where there are no rocks, they simply won’t be able to adjust.

    You’ve argued for our contributing to the technium so that future generations will have yet greater opportunities than we do. This simply isn’t possible for rock ants–they’ve so specialized their limited capabilities that they can no longer evolve in any significant way, just like crystals were an evolutionary dead end for molecules.

  • Oto kiralama

    The examples you give of animal and machine intelligence strike me as being the result of specialization. humanity specializes, too, but never (I guess this is open to debate) as an entire species, i.e. not at the genomic (biological) level, rather at the more significant level of culture, i.e. the technium, which is much more flexible: I can over-specialize in some direction, but if so, the next generation will probably be smart enough not to follow my path, whereas if a colony of rock ants ends up where there are no rocks, they simply won’t be able to adjust.

  • Orpheline

    Hi Kevin,

    Pick up a copy of “Flesh and Machines: How Robots will Change Us” by Rodney Brooks. Brooks is the former Director (1997 – 2007) of the MIT Artificial Intelligence Laboratory, and many of his robots (particularly Genghis) are modeled on insect behavior. The book provides many interesting insights into human assumptions of intelligence: what it is, how you measure it, and whether machines can ever achieve it. A fascinating read.

    • http://www.kk.org Kevin Kelly

      @Orpheline. Yes, Brooks is superb. You can read more about him in my first book (Out of Control), second chapter, called Machines With Attitude. Available on my site.

  • deston

    I took the 20q classic test, my “answer” was sponge.Using your criteria for intelligence and my knowledge of biology, etc., I answered the questions:

    Is it larger than a microwave oven(or bread box)? You said Sometimes, 20Q was taught by other players that the answer is No.
    Can it do tricks? 20Q was taught by other players that the answer is No.
    Is it considered intelligent? You said Sometimes, 20Q was taught by other players that the answer is No.
    Is it smaller than a golf ball? You said Sometimes, 20Q was taught by other players that the answer is No.
    Is it a scavenger? 20Q was taught by other players that the answer is No.
    Is it white? You said Sometimes, 20Q was taught by other players that the answer is No.
    Does it usually hunt for food at night? 20Q was taught by other players that the answer is No.
    Do you use it in cooking? You said Yes, 20Q was taught by other players that the answer is No.
    Does it eat fish? 20Q was taught by other players that the answer is No.
    Is it a carnivore? 20Q was taught by other players that the answer is No.
    Does it weigh more than a duck? You said Sometimes, 20Q was taught by other players that the answer is No.
    Is it black? 20Q was taught by other players that the answer is No.
    Does it live in groups(gregarious)? You said Probably, 20Q was taught by other players that the answer is No.
    Does it squirm? 20Q was taught by other players that the answer is No.
    Would you pay to use it? You said Yes, 20Q was taught by other players that the answer is No.
    Is it round? You said Sometimes, 20Q was taught by other players that the answer is No.
    Does it move? You said Yes, 20Q was taught by other players that the answer is No.
    Is it pleasurable? You said Sometimes, 20Q was taught by other players that the answer is No.

    many of my answers diverged from the common player answers. the danger here is that my knowledge of sponges, such as it is- that they move- albeit very slowly, that they vary in size, that they can be round and – oh, that sponges in the ocean and sponges on the shelf in a store are two very different things, one potentially not even biological- all these answers not being ‘cut and dried’- ie, tangible specimens, but the result of my fool intelligence, which is very different from the next fools and specialized in ways that his isn’t, vice versa- the 20q cant account for this and couldn’t guess sponge in 30 questions. a human, sharing my language, might have been better equipped, a human sharing my language and specialized knowledge or bias, would have been better equipped. so if a person standing trial offers answers which fall outside of the norm, and the 20q cant tell s’he is being honest, are they a liar and thus perjured? we cant rely on this intelligence for emotional or social knowledge yet, no matter how many trained transistor they have: even “Yes, no, sometimes, doubtful, possibly, and unknown” is not enough to mimic the intuition of the human- or even plant- mind.

  • jomega

    If “The common ancestor for birds, whales and apes was 280 million years ago”, and “Birds are the only dinosaurs that survived the great dinosaur wipeout 350 million years ago”, does that then mean that whales and apes are descended from birds, and that Humans, being descended from apes, are also Dinosaurs?!? WOW! I had never heard such a thing before. What are your sources for this info?

    • http://www.kk.org Kevin Kelly

      @jomega: You are right, 280 million years ago and 350 mya don’t match up. For some reason I didn’t type in the correct date for the great dinosaur extinction event which even most school boys know is 65 mya. I’ve corrected the text. Thanks for noticing that.

      @Dileep: It’s not very easy to transfer my Endnote citations to the web, and I am in hurry to keep writing rather than mess with html, so yes, there are references and yes, you’ll have to wait for the book.

      I appreciate the many careful eyes.

  • Ryan Lanham

    Just beautiful. A work of art. Thanks, really made my day to read it. I would love to hear your thoughts extrapalated to human organizations and their evolution.

  • Jshot

    Kevin,

    Artificial/digital mind(s) of the Technium seem prefer to socialize with other like-minded digital/artificial minds, since incompatibility issues continue arise when syncing/interacting digital minds with biological human ones. While digital minds of the Technium still require biological ones for creation and implementation, their dependence on human biological minds will become smaller and smaller as digital minds start to detach themselves from human intelligence and develop their own preferred method of communication and reasoning (and tell us what to do). Why have a human (biological mind) do a job in which the technological (artificial digital mind) can do more efficiently, more precisely, and more affordably?

    In my opinion the digital mind is the creation of the autistic human mind and therefore the digital mind is inherently autistic in a nature. Even non-autistic human minds will adopt more autistic behavior if they frequently interact with artificial/digital minds and technology.

  • Dileep

    While reading the post, I kept wondering after every claim was made, “How on Earth did someone determine this? What kind of test was designed? How long did it take? Surely it would have spanned several grad-student lifetimes!” It would be very helpful kk, if you’d provide references at the end. Or are you saving that for the book :P

  • Elsah Cort

    Martin Prechtel’s story of the “Never-Before-Seen-Bird comes to mind reading this (as well as an avalanche of other thoughts, including the memory of the experience today of getting caught in the new leafed branch edges, of a 200 year old blue oak, that caress the sloped roof of my house almost at ground level, where, suspended in my tracks in the tangled hug of new leaves and branchlets, I “heard” the tree saying something to me in a language I could not translate.) Prechtel’s story is in the back of his book called “The Disobedience of the Daughter of the Sun.”
    This is my first reading of your “book.” Now that I am hooked, I guess I will read it backward and forward at the same time.

  • Jerry Daniels

    Kevin,

    I see you used the phrase “charismatic intelligence” and I’ve seen it used only a couple of other times. What exactly does it mean? I like the phrase, but am curious about how and why you’re using it. I want to understand it.

    Thanks!

    • http://www.kk.org Kevin Kelly

      @Jerry Daniels: I have never seen anyone else use the term “charismatic intelligence” so I don’t know what they mean, but here is what I meant: I meant a type of sentience that we find attractive. It may or may not be ‘superior’, but it feels that way. There are charismatic species which are usually mammals that are like us in some way. So I am extending the concept to apply only to intelligence. When we finally make super AIs some will be charismatic (attractive to our way of thinking) and some won’t.

  • Thell

    I’ve been a subscriber to The Technium feed since early ’08 and wanted to let you know that this article is one of my favorites. It is the rare article that I forward to others, and this is one.

    Thank you!

    Thell

  • John

    This is an amazing article. Truly stunning in it’s content and style. Thank you for taking the time to compose it and share. I learned, I was amazed, and I was made hungry for more.

  • Chris Weekly

    Bravo. Thank you for a fantastic, very interesting piece of original thinking.

  • CanadianAlien

    I have to agree with others, even before reading comments, this is a brief, but sweeping and insightful summary of manifestations of intelligence.

    It may be a matter of perspective but calling the phenomena ‘intelligence’ could be misleading to some. I discern the use here describes a phenomena. Others may perceive to refer to higher cognitive function, or manifestation of will, for example, on part of plants to ‘do’ certain things, motivations which most reserve for humans.

    I regard the phenomena of ‘intelligence’ as an outcome of automation derivative of an elemental level which is universal to all self-organizing systems, only different in its expression via evolutionary process. The automata are neural structures or their analogues that facilitate environmental awareness, memory, pattern recognition so elegantly summarized by KK, but are things such as fractal development, simple cause-effect and equilibrium. Moravecs’s curve of processing power vs biological level, and Kurzweil’s extrapolation of this to technological rate of growth overlap here too.

    Look at a mirror into your pupil, consider it is just a clear window to allow light to hit your retina, in turn be interpreted by your brain. Consider that ‘you’ looking into ‘your’ eye and ‘your’ comprehension of this, is ‘just’ the result of an elegant, complex, self-organizing system – not unlike those rock ants.

    Kevin, you elevated those Rock Ants to human level in many respects, but you could’ve also dropped us down to their level! Either way though thanks for the piece!

  • CanadianAlien

    I suspect too, that when you say “the minds of the technium are getting smarter by the year” many will interpret this as their calculator is going to achieve sentience. I understand it is as meaning that the aggregate of the technium, and specifically, it is subsequent iterations that are ‘getting smarter by the year’.

    If ‘intelligence’ is ‘just’ automata then it is ‘just’ an engineering and evolutionary iterative process, a la Moravec and Kurzweil.

    I also suspect that many otherwise capable people dismiss AI because they may not believe artificial systems can replicate the natural state (eg Jaron Lanier), or for whatever reason are unable to understand or comprehend the ‘mindedness’ as you call it, or the fundamental automata that all self-organizing systems possess.

    This is not a trivial issue, because we are astonishingly in the thrall of our technology, and increasingly in accelerating way, to our information systems. If, when, the automation you describe above can be replicated in the technium, it can be spread like copying a CD, or evolved in virtuo in iterations much shorter than 3.5 billion years of evolution.

  • Daveygriff

    I was amazed to see my casiio calculator using its own screen to construct verticle lines of code vaguly similar to bar codes but pulsing similar to a heart beat.