This has appeared as:
Chapter 3 of YEAR MILLION (ed. by Damien Broderick, Atlas & Co. NY, 2008)
Additional error corrections by the author are included
A Million Years of Evolution:
Wetware to Software to Fashionware
Steven B. Harris
A million years is not much on the scale of astronomy, nor even of most geology. A few rivers and seas change. If continents move a hundred kilometers (about sixty-two miles), the difference is hardly noticeable [as Dougal Dixon notes in Chapter 2]. Discounting human macro-intervention, the Earth will look about the same in the Year Million, give or take a polar icecap or two.
In biology, though, a million years can work wonders—or, sometimes, do nothing. Much can be accomplished over the tens of thousands of generations this span represents for mammals, if the pressure to adapt is extreme. Around two million years ago, we humans were big meerkat-style hominids, either Homo habilis or something similar, standing four feet high on our hind legs and shambling about on the East African savanna, sometimes falling prey to leopards. At this stage we were not an impressive animal, but even with brains half our present mass, we were learning slowly to handle rocks as tools. We had little other choice, exposed on a relatively treeless plain, already too large to find safety in hole-digging. A million years of such pressure, or perhaps even less, turned us into hulking, six-foot tall Homo ergaster, with a brain now expanded to three-quarters its present mass, a nearly modern human body below the neck [1], and the ability and proclivity to chip big rocks into wicked two-sided razors. By this time we were at least no longer cat-food—even if it wasn’t yet apparent that we were destined to be rulers of the planet.
Somewhere during this crucial time, even faster evolutionary mechanisms became available. Along with a bigger brain came several new selection mechanisms, each of them with a drastically reduced time-scale.
Consider artificial selection, an improved type of imposed evolution running at high speed. Darwin knew it well, and knew vaguely that it works especially quickly. Artificial evolution can move much faster than Nature, because the interventive pressure that runs it can be maximized. Nature rarely eliminates all animals that lack a given trait; artificial selective breeding can and does do so, nurturing survivors, protecting them against predators and providing ample nutrition. Most of our modern agricultural crops are scarcely older than ten thousand years, and most of our domestic livestock breeds, too. That is the accelerated pace made possible by ordinary “barnyard” artificial selection. The relatively recent development of toy poodles, bulldogs, modern cultivars of maize and wheat, and other improbable organisms that would not survive in the wild, shows that evolution can be engineered and turbocharged.
But did we artificially select ourselves? It doesn’t take laboratory equipment to carry out applied genetics, and humans have always been at least as interested in our own breeding as in agricultural breeding. This is sexual selection. Other animals use it, to conspicuous effect. Even in the wild, peacocks have ridiculously long and survival-handicapping tails: good for impressing the opposite sex, bad for escaping a predator. Pea-hens evidently use a glance at the peacock’s tail as a simple one-shot look into the general health and genetic fitness of a male, who is bragging that he can support growth of a superior-looking tail despite the pressures of parasitism and predation. Sexual selection is particularly interesting if done by high intelligence, which then has an opportunity to sell itself. In this case, intelligence itself becomes a sexual attractor.
Interestingly, the mushroom-like growth of the human brain over the last 1.5 million years—it has increased a third in mass while the rest of human anatomy changed relatively little—resembles nothing so much as that of the peacock’s tail: the utility of it is not immediately evident. Obviously, however, much evolutionary pressure has been applied differentially, above the neck. The human brain now requires twenty percent of our resting calories, and is the largest for body size of any mammal. How did we get it? In humans, language and fine athletic and other coordination (which takes intelligence) are some sexual selection factors. Such abilities provide a look at the function of the brain, an organ that not so easy for other species to evaluate in their own mate-selection.
In other animals, the brain is not so critical as in humans, who have few other physical gifts. The way humans survived in their savanna niche depended on keen wits and cooperation. Consider the importance (and difficulty) of assessing “team-player intelligence” in a species that has few natural defenses, and relies on the ability to coordinate hunting groups for survival. Perhaps our ancestors selected themselves into being what we would now recognize as “big players” on athletic teams, who found themselves with the most desirable admirers after a successful hunt [2]. That happens in all hunter-gather societies in the present day, and we might suspect that’s been going on for far longer than recorded history. At the individual level of reproductive choice, humans have been at least as choosy about whom they breed with as they are now. We Homo sapiens are not like relatively promiscuous Bonobo chimps. Women have always made men work hard for their favors, because a human infant is a relatively terrible burden, in terms of calories and time. So human males rapidly grew the size of the parts they needed, and that included the organ behind the eyes. In turn, those changes were handed down to both sons and daughters, and so on, in a self-reinforcing cycle.
Cultural Evolution
The subtle brain differences that make Indian elephants potentially rideable but African elephants not, aren’t visible even to a microscopist working with fresh brain tissue. But we know they must be governed by the DNA of both species (even if we don’t know which genes are involved). Even subtler changes, not genetic at all, can take place over time in an individual’s brain—changes in software, so to speak. For example, the behavioral difference between a wild mustang horse and the saddle-broken horse it may be turned into on being caught, is simply programming. For humans, this software, which we call our “culture,” has become ever more important to our behavior—yet culture, too, is subject to evolution. Humans raised without socialization turn out to be barely more than animals, despite their large brains. Those brains, therefore, represent only raw processing power (with perhaps some of the same domestication proclivities as domestic animals), and they need a lot of programming.
Homo sapiens sapiens (the fully modern human) is now largely a software species, perhaps the first. We are governed mainly by epigenetic factors (outside the genome), some of which are extrasomatic (outside the body). In short, much of what makes us special as a species is stored not in genes or brains, but in libraries, laws, traditions, and songs.
Artificial selection (of which sexual selection is just one instinctive part) requires some culture to decide on breeding goals and standards. How can a society decide what characteristics it desires in its mates, children, and livestock without cultural values for such things? Without language, there are limited possibilities for cultural variation and depth. Some culture (extrasomatic knowledge passed between generations) has been observed in higher animals, but it seems limited to a sort of direct learning via direct imitation. Not much knowledge can be stored in this fashion. Language and a good memory allows cultural programming to become predominant over genetic programming. A good brain can store many thousands of lines of oral tradition, particularly if formatted mnemonically as a saga. And even without language, a good brain can also store a large amount of motor learning about how to knap flint, use fire, and weave textiles. We can’t tell when humans began to exchange such complex pieces of cultural programming, but somewhere in the Neolithic era (with active agriculture and large cities, starting around fifteen thousand years ago), people started doing complex tasks that must have taken a lot of coordination—far more than the average wolf-pack or lion pride manages.
Our own high culture came very late in our brain evolution. Something more was needed for the next phase in evolution to happen without any obvious brain growth. Near the end of the last ice age, perhaps ten thousand years ago, bands of people began to settle in the fertile crescent in Asia—to wait for food crops to grow (which they had learned to replant annually) and to hunt the beasts that came to eat the crops. Soon, they began working to tame and remake these beasts as they had the plants, and eventually had domestic animals and were fully farmers. Rooted from year to year for annual cycles of harvests and slaughters, they built the first cities, and in the process they invented specialization. In some cases, markers for stored goods evolved into written symbols.
Writing, when it arrived, changed everything. Now intrinsic memory was not so important. Complex contacts and laws could be impressed even into tablets of clay, and societies could store their cultural programming outside oral tradition. That allowed (finally) for the technological leap into the present day. The leap depended on a completely new form of evolution in culture, one dependent on information storage in which most of the content is stored outside the structure of the brain. Now, operations on this kind of data could be done much more collectively, across both space and time.
People who cannot write, also cannot achieve a culture that produces anything so complicated as firearms, much less the rest of an industrial society. An industrial revolution requires printing. Before the movable-type printing press arrived in the fifteenth century, record-keeping for complex collaborations was just too much work. Printing mechanisms allowed spare time for complex coordinated activities, for experiments, and for the easy reproduction and wide dissemination of the results. This included what we know as experimental natural science, and engineering complicated enough to need blueprints and careful specs.
Finally, when electronic computers arrived and shrank enough to fit into homes and then into the hand, suddenly, almost magically, some individuals gained digital access to all of human culture. That assimilation process seems likely to be complete within a generation.
Culture as Artificial Intelligence
Our next limit in a fully digitized culture (now fast approaching) is total processing power, not access to the information to be processed. The next limit therefore involves the connectivity and bandwidth of communication between brains. Most people can talk (and write) faster than they can produce first-rate thinking and conclusions. For this reason, while brain-to-brain information transfer remains limited by the information transfer rates of speaking and writing, that’s not nearly as big a problem as it might seem. Until now, flows of high-quality information in and out of brains have been crimped more on the input-side (such as such as looking for information at the library) than in the output-side (such as writing content worth publishing and looking up in that library). Fortunately, the worst crimp—the search problem—is disappearing first and fastest.
Human culture is mostly software, but a society is more; it’s partly interactive software that does information processing (by both brain and computer), and is often unaware of what it’s doing. Adam Smith realized that economies run that way, using what we’d now call “parallel decentralized processing”—a process where nobody in particular is in charge, and there is no central control. In one sense, that’s some of what our high culture is: a collection of increasingly intelligent “entities” constructed out of multiple brains and data storage centers.
Such collective entities can do things no single human can, and this was true before computers. A modern city or a nuclear aircraft carrier is beyond the mental capacity of any one person or group of unaided persons, but it gets built anyway, though the collaboration of thousands of engineers and whatever devices they use to communicate and draw plans. Such mega-projects already depend on a form of “artificial intelligence” (intelligence greater than that which humans can mount without artificial aids). Today, the augmentation of intelligence needed for such tasks is mediated by data-processing machines. Does it matter whether or not some of the total processing to build an aircraft carrier is still done organically? That fraction done by brains will finally become irrelevant, when the distinction between processing information “organically” and “electronically” disappears.
Moore’s law and Accelerating Change
Electronic data-processing power per unit of money is increasing exponentially, and shows no sign yet of hitting any wall. At some point in this century, information will be processed by manipulating spins of individual electrons on appropriate electronic “devices.” Some number of these (I won’t guess how many, but I think it will be an attainable number like a million or a billion) will equal the processing power of a human neuron. This must be a recognizable number, because a neuron itself contains around five hundred trillion atoms [3] (this figure is calculated easily: most are atoms in water molecules), and most of these atoms aren’t harboring a spinning electron involved in some neuronal calculation. It has been suggested that neurons are quantum-entangled computers, but very few people who know of the jiggling that happens in our warm interior world of cellular proteins, believe it. No matter whether neurons are complex computers or simple ones—analog, digital or something in between—at some point it will be possible to replace any given neuron with its electronic equivalent. Then, more of them. Then finally (why not?), all of them.
Well before then, however, it should be possible to connect neurons more directly to computer memory, and also to neurons which are in other brains. At some point, mechanical telepathy will become possible, with all that this implies. Imagine being able to remember the entire contents of our digitized civilization with the same ease that you remember basic spelling. Now, imagine being able to do that trick with selected memories from another human being. Scary thought, isn’t it? Do you trust anybody enough to give them access to your candid diary, let alone your entire memory files? Possibly not.
But suppose you did, for a while. When you pulled the plug, how much of the other person would remain “in you,” as a part of you, and vice-versa? That would depend on how good your file copy system is. If it were large, you could retain copies of the other person’s memories to access at your convenience. At some point, you could even give these borrowed memories permission to influence your unconscious actions. You would then acquire not only something of another person’s declarative memory, but also the other’s intuition, judgment, anxieties, and passions. We do this slowly, now. Already, people in a long and close relationship often know what the other is going to say. This sort of internal emulation could be amplified to any extent.
Now imagine drawing in other minds, perhaps of anybody willing to participate. The result is the internet on crank: direct brain-to-brain connection, amplified by computerized search and processing help.
The Chain Reaction of Intelligence Explosion
Explosions are typified by runaway positive feedback. A familiar example is a chemical explosion, in which the reaction releases heat, which in turn speeds up the rate of the reaction, liberating more heat. This leads to an exponential growth that continues until it runs out of fuel, or is otherwise stopped.
All the kinds of intelligence increase we have discussed have a built-in positive feedback loop, starting with our first homely example of sexual selection for intelligence. Presumably, more-intelligent animals become better at any kind of sexual or artificial selection—including those that produce more intelligence. But what’s special about selection focused on intelligence itself is the positive feedback loop that makes the process accelerate at it's fastest. This exponential curve probably explains how human high culture arrived on the scene so fast—after those hundreds of millennia of chipping rocks and wearing skins. Only recently did we reach the point where a certain spark in selection for effective intelligence has received the fuel to create a “blowup” fire.
We know it’s possible for humans to vary greatly in intelligence without much difference in brain size. Culture matters. For example, there’s a well-known “secular I.Q. trend” (the Flynn Effect) that extends across many cultures and socioeconomic groups, indicating that as a species we gained more than twenty-five I.Q. points in the previous century. Nobody really knows why, but it doesn’t seem to be better nutrition. Instead, it seems we’ve simply learned to think better, with better educational tools. Thus we see directly that intelligence can improve by strictly cultural means, even if hardware (brain wetware) doesn’t.
The rapid exponential growth of both individual and collective individual human intelligence becomes inevitable once it is freed from the constraint of having to wait, generation by quarter-century generation, for better brain hardware. Cultural evolution works fast. All the positive feedback loops in effective intelligence are present for cultural evolution, but the cycle time shortens and re-shortens, depending on ease of information storage. Our brains probably haven’t improved much in innate capability since the last ice age. But with information technology driving “effective intelligence” progress (measured simply by gross capability), we have moved from writing to the printing press in eight thousand years, then from press to mimeograph in about 430 years, from there to the photocopier in less than seventy-five years, to personal computers and universal word-processing in another thirty years. That takes us to the late 1970s. The modem reached ordinary consumers in the early 1980s, the www software that now runs the internet arrived in 1990, and the first workable web-browsers for the masses, about 1994. Ten years later, these became portable, through Wi-Fi and cell phones.
As I write (2007), a billion of the world’s 6.6 billion humans have regular access to the internet. Another decade or two should see this particular level of interconnectivity—only about thirteen years old now—nearly complete, save for some backwaters. Even these may be surprising: cheap cell phones have penetrated to parts of many countries where people have access to a phone but still not to clean water or enough food. This trend will continue with web-enabled cell phones which are presently expensive, but will not be expensive for long.
When the average person has access to the internet, all culture will have moved accessibly outside our brains, although the ability to process it will mostly still reside within the organic parts of the machine. There we hit a temporary snag, but even after full culture digitization and full downlink connectivity, we’ll still have 997,000 years to go before the Year Million. Patience.
In retrospect, it is odd that we didn’t fully realize until about 1965 that the positive feedback loop leading to explosive intelligence increase, was capable of operating overtly. We were looking at intelligence in the wrong way, not realizing that you didn’t need full artificial intelligence (A.I.) to get intelligence augmentation, and consequent exponential evolution in intelligence. In that year, I.J. Good finally addressed the basic question of runaway feedback, starting from a strictly A.I. viewpoint. Good asked what would happen once computers became powerful enough to completely design “smarter” computers (he didn’t ask what would happen once computers got smart enough just to help design computers). Science fiction quickly suggested thinking machines far smarter than humans, mostly malign: the computer in the movie Forbin Project (1970) that we’d like to unplug, but can’t; and Terminator’s cold and lethal Skynet (1984).
None of these prospects, however, are fundamentally different from the feedback loop in intelligence-raising we’ve been engaged in for millions of years. What is taking place now is a new leap-after-leap forward in the pace of the loop cycle. We’re noticing it more and more. The next big meta-jump will take place when human brains talk directly to computers, and to each other.
Direct Connection of Brains
We are not smart enough to do this right now. We’re still at the level of sticking electrodes into brain slices and reading the waveforms. But we can guess what will happen when we connect brains more directly. Composite minds more intelligent than any single person are not only possible, but something we’ve already experienced. They impart to our high-tech culture, from the economy to the internet, some of its odd alien nature, as though it were all controlled by some god-like conspiracy that hasn’t let us in on the big secrets. But of course, that isn’t true. The composite and distributed “mind” that runs our economy and culture is indeed god-like in its abilities, but there’s no conspiracy. It seems secretive because it doesn’t talk, but that’s only because it can’t talk. It’s not self-aware, and nobody with full understanding is “at home” inside it, in the same way nobody who understands chess is “at home” inside the most advanced chess-playing computer. Nevertheless, the end result appears to be creative and possess understanding, even to a grand-master. Such a system does its job in a massively parallel way, without a central government, just as Mother Nature does. It works very cleverly, but it doesn’t explain itself or its thinking (in fact, when parts of it try to do so, we see how ignorant they appear, and wonder if they are part of the conspiracy). We ourselves are pieces of one or more giant “cultural minds” and make up parts of them, but none of us are really aware of the total process, any more than a shopper is aware of participating in a vast distributed economic calculation.
Darwin deliberately read economist Adam Smith for help, while he was developing his theory of organic evolution. He knew he needed a theory in which design happened “all by itself,” without any central planning department. And indeed, all types of evolution, as we now know, are basically equivalent. Living organisms were “designed” in many cases, like Wikipedia articles. The process is gruesome, like what Otto Von Bismarck said about laws and sausages (i.e., those who have any regard for them should not watch closely as they are made). In all evolutionary processes, there are plenty of bizarre and vestigial bits left over, and some malfunctions and vandalisms. But the overall result after enough time, as with the economy and the common law, can be quite magical, and certainly better and more creative than anything top-down or centrally-planned. That is the paradox.
What is the experience of being part of a superintelligence? We already know some of the answer: like being part of a team doing an important task. To everybody on the team, it feels as if godlike people are doing most of the work, someplace else. But if you look around, you can’t find them. Our experience residing in our own brains mimics this experience to a degree, because most of what we do, and think, goes on under the surface of our conscious awareness. We get some hint of it in dreams, when the parts of our brains engaged in modeling the actions of other people suddenly split off and take control, acting as the bizarre stage-builders, playwrights and actors of the little dramas we find our “dream-selves” moving through.
The Borg and Assimilation of the Organics
Some say the next tipping point to the evolution of mind will be I. J.Good’s singularity-like scenario of “ultraintelligent machines”—the point where we can directly redesign the brain (or any brain) to be more intelligent than it was. This enhanced of “brain” is then smart enough to design still better brains, until the whole self-bootstrapping process leaves us behind. This might be too optimistic: the programming and design of such artificial brains is beyond us, and might require evolutionary help from self-evolving software. If so, A.I. specimens of human power may not end up understanding themselves any better than we understand ourselves. Instead, all we will have achieved is substrate and programming that can be improved more directly and rapidly.
The grossest capabilities of a single organic brain can be changed now only by artificial selection, and thus human capacities won’t grow much in the near term—not until we have full nanotechnology, and the ability to build and replicate brains from the cells up. Long before then, however, connectivity between brains and interfaced computers will have vastly increased our effective intelligence. At some point it won’t matter much whether computers are smarter than humans, because what others have called the “Borganism” (the internet of brains and the computers that directly link them) will reach the capacity to do the next step, which is to control and grow computational matter. Then, organic brains as we know them (and grow them organically, by unskilled labor) will no longer be needed. That shouldn’t be too scary a thought, because the “people” who inhabit such collective thinking organizations will still be part of conscious entities. But these people will be programs, “running” on substrates other than gray matter.
With old-style organic brains removed from the Rube Goldberg machine of animals and devices that had been the interconnected network of mankind, at last we will have hit the final evolutionary phase—a pure expression of what we’ve been calling “cultural evolution.” Ultimately, the evolution of software, ideas, and knowledge is the only evolution that counts on short time scales. Any software or hardware design innovation can be translated eventually (if you have the technology) into thinking “hardware,” without waiting for a present-style “organic” brain to grow in the womb, be born, and be programmed. Those scare quotes around “hardware” (or “wetware”) are needed to remind us that computronium (the material used to do the information processing) need not be silicon chips, and almost certainly won’t be. How mechanically “hard” computronium need be, is also open to question. Perhaps it will be diamond compiled from carbon atoms, or a diamondoid sponge. Certainly the mobile units that connect with it, when physical movement is required, won’t be metal robots that clatter around. Why build such awkward things, when we don’t need to?
While we don’t yet know what the mobile units will be made of, we can imagine that the distinction between living and dead matter will lose its meaning. Any technology capable of copying living organisms is capable of copying them with better and tougher components (diamond-fabric woven bones, for example), and any of this material can justly be called “alive.”
Probably the new “stuff” that results, won’t (or will rarely) consist of anything so delicate as twenty-micron sized bubbles of water, full of protein and DNA, scaffolded and bound to each other with stringy polymers, as though somebody were packing water balloons individually with fishing line and duct tape, for long distance shipping. That's how we’re made now, but that’s an accident: life started as little balloons that floated around in water and didn’t need much support. When finally these squishy cells needed support to resist the gnashing of predatory teeth and the vagaries of life on land, they had to develop it by bricolage, and couldn’t redo any of the basics en route. Won’t it be nice to get rid of all of that fragile stuff? The present human body is like a bicycle made of porcelain, which mostly can’t be repaired; all evolution has done is add more packing.
Bodies of the future will not only be much tougher, but will also feature backups of brain information in case of total destruction. In fact, the primary thinking centers probably won’t even be located in the mobile-unit/body, so backups of memory and thinking will be automatic. There is no physical reason any of this change would need to feel to the user any different from the body being “worn” now. In fact, it might well feel like an improvement, with enhanced senses and abilities. What feels “natural” is, when examined closely, strange enough already. Your present brain sits, shock-insulated in the dark, and gets only a view or feel of the world by means of chains of digital signals coming in from sense organs. All this can be manipulated—in theory—at will.
Computronium: Software in Action
The actual structure of computronium is unimportant. Mobile units might be soft and warm and, if desired, have a metabolism. One might even try for the complete “look and feel” of the human body, but without delicate cells full of water, making the unit extremely durable. There’s no reason such a mobile unit shouldn’t have a processor capable of individual thought (perhaps of a better grade than we currently enjoy), even when disconnected from everybody else (the net). The shape could be tailored to any occasion. For exploring the ocean, we might want something like a squid; for the air, something like a bird. On land, for a change, one might choose to be an efficiently-running four-footed body modified with extra arms and hands, like a mythical centaur. The same applies to exploring other worlds: if the mobile fits, wear it. Whatever body can be designed, it will deserve the title and privileges of “human being.” When we look out of the “eyes” of such bodies, we’ll certainly still think of ourselves as “human”—if we remember our heritage.
For those of us who want to play at being totally old-style human in form, why not? The Amish of the future can choose to use the outhouse if they like, and can choose to need to use the outhouse. If that makes an organism feel organic and “creaturely,” well, it’s available. Remember, however, that if you have a fully human brain in your chosen “body”, there will be limitations in thinking, perhaps in connectivity, and also a certain physical frailty—so don’t complain if there are games you can’t play in that mode.
The interconnected mechanical/organic composite Borg (short for cyborg or cybernetic organism) of Star Trek is horrifying for several reasons, but one is that they seem like badly patched-together and ugly blends of organic beings and hard computerized robot gear [4]. In any case, we can dismiss the television horror view—that’s what things might look like at worst, when brains are still part of the equation, with no way to replace them. Even then, non-organic or biological prosthetics for brains are likely to be considerably less repulsive than a deliberate monster-show. By the time nanotechnology is capable of building complete bodies, using computronium for thinking and other morphable substances for moving, nobody will want to look like the Borg—except for a Halloween party. Instead, a body will look and feel however you want it to, sensitive but tougher than flesh.
And of course, since such stuff should be infinitely repairable and connectable, it will be effectively immortal. No material object is indestructible, but aging as a problem simply goes away automatically, once brains are fully inter-connectable, because whenever a part (even a thinking node) is damaged or wears out, consciousness moves to another location until the damaged body or part is replaced. Repair of bodies (and “brains” or processing gear) should be like replacing any backed-up computer node in a networked system. Aging, as we know it, is only a problem when you’re stuck in a single brain.
Individuality, or the choice of solitude in such a world will continue to be an issue that people will need to solve in their own ways, determined then (as now) by available bandwidth of communication, with all that this implies. Choose a wide bandwidth, and you will find your consciousness completely immersed and diffused in whatever number of others with whom you’ve decided to merge. Chose a bandwidth of zero (or have it forced upon you by distance or equipment failure), and you will become once more, in that sense, an individual. By the Year Million, you might not like such isolation. Or you might be the sort to do it deliberately, as a survival challenge. No doubt the future will contain many such games.
The Singularity
At some point, organic brain material will be replaced finally with “computronium.” As noted, the historical juncture at which designed mental substrate becomes as complex and powerful as the human brain has been termed the “singularity” by mathematician and writer Dr. Vernor Vinge, who coined the phrase in 1982, following the 1965 lead of Dr. Good. Intelligence increase at the singularity will remain exponential and finite, but it will greatly change its cycle time for improvement—to what cycle time, no one knows. The point of the term singularity is that all intelligences past this point are both superhuman in scope and capability (as in fact is any working group of humans now) and, in addition, such intelligences presumably will have the capacity to improve their own effective intelligence more directly than we enjoy now. Their own operating code, unlike ours, might be an open book to them.
By extrapolating the complexity of computer systems in comparison with the estimated complexity of the human brain, many futurists have estimated the timing of the singularity to arrive before the middle of the 21st century—far short of the Year Million. But are there likely to be any fundamental growth limits that might effectively stop super-intelligences from simply taking over, after say, 2050, and then doing anything they want to with the universe, shortly after that?
Limitations of Energy, Time, and Space
There must be limits to intelligence—because time, space, and matter limits apply to the substrates used to create intelligent matter.
Let us assume modestly that in the future there will be no “magic physics.” The usual conservation laws of physics will continue to hold, thermodynamic limits of entropy are real, and causality continues to be inviolable, so that reverse time machines are impossible and the speed of light remains a limit. [Some other chapters in this book drop that constraint.] If we abandon those rules, prediction becomes uninteresting, for then anything becomes possible. I also assume for our purposes that dark energy and dark matter will turn out to have strange properties and strange uses, but none that violate any of the aforementioned laws and properties of the universe.
What, then, are our limits, as we move toward the Year Million? No matter what body forms we use, we will need energy and matter. Where are our largest sources for these? We might be able to partially take apart the Sun, using its own power, but that’s a losing game, since if too much mass is removed, the power fails. Matter obtained this way is costly in energy, since it is deep in a gravity well. Removing even the outer one-tenth percent of the Sun (assuming perfect efficiency) requires thirty thousand years of solar output energy. This mass is about the mass and composition of Jupiter, and can be used for the same purposes at only one percent of the energy cost as using Jupiter. So the outer solar system will be mined for material before the outer Sun is.
Jupiter represents more than seventy percent of the solar system’s non-Sun mass, so the entire system can be considered, for our crude purposes, to consist only of Jupiter and some other debris (like our present habitation) which isn’t worth bothering with in our calculations. For example, excluding the Sun, most of the solar system’s energy of motion is in Jupiter, as is most of the gravitational energy of self-binding. So the other debris will presumably be used for practice when we need material, or else (as in the case of Earth) set aside as memorial or park if we choose.
An intriguing part of Jupiter is what we believe to be its supercompressed core of carbon and heavier elements, probably about six times the mass of the Earth. We probably need elements heavier than hydrogen and helium to make computronium. Hydrogen itself as a “thinking material” looks a little iffy, because it doesn’t seem complicated enough to make anything useful with it. Ultra-compressed hydrogen probably turns metallic, but that doesn’t help us much, either: electronics as we understand it requires electrons to be confined and isolated, and in metallic conductors they are too free.
In any case, we’ll build clouds of solar-orbiting collectors out of the heavier elements, in order to catch all of the Sun’s light we can. Powered by this energy, our information processors will placed at nodes somewhere within the cloud. Such a cloud is called a Dyson Swarm, after Freeman Dyson, who first envisioned this engineering feat. For computation, we need nodes with dense processor connection, because otherwise we have speed-of-light connectivity problems. However, we can’t put the entire computronium collection in one place, because its self-gravity will cause it to collapse into uselessness; computronium cannot be a simple material, so it will have unavoidable pressure limits. Also, we have a lot of heat to get rid of, because computronium won’t be perfectly efficient, and will require a complex infrastructure for forced convective cooling, as any large engine does. Both of these basic constraints will dictate a mix of thinking nodes and connections. [Robert Bradbury’s chapter addresses the tradeoffs of problems in making Matrioshka brains or M-Brains (planet mass onion shells of distributed processing nodes, associated with clouds of solar collectors around light sources such as stars and fusion reactors)]. But imagine an orbiting cloud of collectors and processor planetoids, all bound together by coherent beams of connecting communications radiation. Such a cloud would orbit a star or any smaller fusion source that puts out light.
To make the largest of these, Earth would be just the beginning (or Earth may be preserved as a park, while we begin with asteroids and the smaller planets). Here’s a motive for pulling Jupiter apart (provided we don’t find exotic life there, floating in the clouds). How much energy of self-binding do we have to overcome to do it? It’s convenient to calculate it in terms of solar output. It would take twelve years of solar output to move Jupiter out of the solar system, but about three hundred years of output to disassemble Jupiter completely into usable gas (other authors, like Freeman Dyson, have figured this at a few times higher, but we need only order of magnitude estimates here). The figure for disassembly of Earth is about six days, indicating why we can ignore it, and even the other gas giants that are only a tiny fraction of Jupiter’s usable mass. “Jupiter” stands for everything that is outside the Sun’s gravity well.
The Sun provides energy at larger rates than most other processes we can think of, so its energy will likely be tapped first. That rate is 4 x 1026 watts, and given the thermal gradient from the 5500 oK (degrees Kelvin) sunlight to the 2.7 oK of the microwaves in interstellar space, nearly all of this energy is available to do useful work, in theory, depending on how clever we are in building solar collectors. And of course we will be very clever. We have only about eight to ten Earths’ worth of non-hydrogen/helium building materials in our entire non-Sun solar system (most of it inside Jupiter, as noted), and this may not be enough to collect most of the energy of our Sun. For example, at the radius of Earth’s orbit, a shell of this matter would be only a tens of meters thick. If it were to be converted to orbiting clouds of solar-power satellites (even disregarding what matter we’ll have left over to use for thinking), we end up with effectively a thin layer of matter to stop and process the Sun’s light. We may not have enough for really good efficiency, but it will be a good start. To get a limit of ninety percent efficiency we need only radiate waste heat to space at a temperature of one tenth that of the Sun (at 550 oK), and the collector orbital radius to do that is well inside Earth’s orbit—something like the orbit of Venus. Going to this distance instead of Earth’s cuts the collector material requirement in half.
What do we do if we need more energy than this? Or if we need the energy from more solar collectors, but cannot afford material to build them, because we need it for computronium? (It would be nice if certain chemical elements work for collectors and not for brains, though we can’t guarantee this.) The Sun only produces energy only at a certain rate, and if we want energy faster—perhaps for building space probes to shoot at the billion-plus stars within ten thousand light-years, or merely to power faster thinking in our own solar system—we’ll need to extract it faster. Remember, by this time we will think really fast, and we will be ready to use energy much faster than our puny star releases it. We can always go to other solar systems for materials to build more collectors for our own star, but that will raise the output a fraction, because we already expect reasonable efficiency in our “Dyson-swarm” collectors. If we must go space-faring to get multiples of the solar output power, we need look to alternatives other than the Sun.
If we want to disassemble the planets for their elemental content, we’re going to look closely first at their fusion fuel content. This represents the primary energy resource of our solar system and presumably of other systems as well.
Mining Deuterium
We want Jupiter and the other gas giants not only for the heavy raw materials at their core, but also for their deuterium. Deuterium is the heavy stable isotope of hydrogen. It is absent from the interior of the Sun, because it’s destroyed there. In fact, no natural processes create deuterium in significant amounts today, but it was produced in the Big Bang and so remains in the Sun’s outer layers, and in the gas clouds that form planets in our solar system and other star systems. However, it’s a diminishing, non-renewable resource. All the time a star burns, or when it goes supernova, scattering the ash that makes new star systems, the process destroys some of the deuterium. As time goes on, less and less deuterium is left in the universe.
Deuterium is very much easier to fuse to get energy than is ordinary hydrogen—so easy that the first true U.S. hydrogen bomb was little more than a big uranium-covered thermos bottle of liquid deuterium, squeezed by an ordinary fission bomb. At ten megatons yield, this worked fine; about half of this energy was from the deuterium.
On the materials side, about ninety-eight percent of Jupiter is hydrogen and helium, which we can perhaps find some use for, as computational material. But even if we cannot find any practical way to use hydrogen and helium as substrate computational materials, we can possibly use them as reaction mass for our fusion rockets (Orion, etc). And there is always the remaining two percent of Jupiter, which is still six
times the mass of the Earth, and still represents much of the usable heavier elements available in the solar system. (We don’t yet know the exact mass of the Kuiper belt and Oort clouds, but there is reason to believe it is small enough to ignore.)
As for using Jupiter’s ordinary light-hydrogen (not deuterium) for nuclear fusion power, we’re far past the edge of what is technically possible, there. So far as we know, ordinary matter and electromagnetic fields cannot confine this light hydrogen isotope well enough to get much energy efficiently out of it, so only gravity works for the process. We know light-hydrogen can be burned in fusion, because stars do it—but on the other hand, nature may require a star for it. A clue to the difficulty is that, even at the heart of the Sun where temperatures are fifteen million degrees Kelvin and densities are far higher than can be sustained indefinitely by any other means, the energy output is a modest 15 watts for a volume the size of the adult human body-- a reptilian metabolic rate about a fifth of the energy production of the human body itself [5]. The Sun therefore manages its prodigious energy output with the poor fuel it uses, only by having a vast volume and by working continuously-- two things that cannot be simultaneously done by any imaginable device.
So we are stuck with light hydrogen. Humans at present cannot fuse light hydrogen, in either bombs or machines, and we have no idea how to do so artificially with any efficiency that will give us net energy. It’s just too difficult (so difficult that nobody has ever provably even done it). It’s one of those things that might be possible in the future (like making small black holes), but it’s also one of those things that are possible in theory, but might well never work out. But it’s a worthy goal: if it can be done, it’s worth doing, because Jupiter has about one one-thousandth the mass of the Sun, and thus holds enough hydrogen to equal at least eighty million years of solar output, in fusion (compared to the eighty billion we could in theory get from the Sun’s hydrogen; though even the Sun is expected to get only fifteen percent of this, before it goes haywire).
We know Jupiter contains deuterium, and how much (fifty parts in a million of the hydrogen by weight is deuterium). Dismantling the giant planet for energy is worth doing even if the deuterium is the only usable thing we find (again, we could fuse deuterium even with 1950s’ technology). Jupiter’s total deuterium supply is equal to a bit more than one percent of the mass of the Earth. We can get about 3400 years of solar-equivalent output from that amount [6]. Still, that’s more than enough to pay for mining it out of Jupiter’s gravity, because by using solar power we can take Jupiter apart in a few hundred years, and burning its deuterium saves us having to wait 3400 years for the Sun to generate as much energy as we get from the deuterium. The ratio of energy obtained from Jupiter’s deuterium, to the energy needed to pull Jupiter apart to get it, is at least ten to one.
Deuterium energy, as noted, is easy to harness. For bombs, the fission trigger that does the initial heating and squeezing can be made of minimal size (a few kilogram of fissionable material), after which successively large deuterium bomb stages may be used as igniters for larger ones, ad infinitum (three-stage bombs have been made already, and there appears no limit to how many stages a bomb may have). Thus, we won’t run out of uranium before using all our deuterium.
The energy in deuterium can be used directly by surrounding a bomb with solar type energy gatherers (at a distance), or else a series of small bombs can be used as a pusher in an Orion-style craft. Or presumably our super-intelligent progeny will figure out how to make “slow burning” deuterium-based thermonuclear power plants. This is something we haven’t quite done (at least efficiently), but being well on the way to it, we have every reason to think it’s possible (as might not be the case with light hydrogen, even in the Year Million).
What will we want all this energy for? Pure computation in theory costs zero energy, but storing the results and memories does, and we may need the energy from deuterium fusion for that. This may allow us to use the eight or ten Earths’ worth of non-gas material in the solar system, to build things other than solar collectors.
If it’s possible to fuse light hydrogen, deuterium collection won’t be worth doing—we’ll have a surfeit of energy just from Jupiter’s stock. But if we need deuterium, it should even be worthwhile to visit local star systems to harvest deuterium from the many gas giants that must surely be circling nearby stars. For example, some sixty-six stars reside within ten light years of Earth. We believe gas giants are common in star systems, since astronomers see a lot of them. Even if the sixty-six star systems at near distance hold only ten percent of the gas-giant planetary mass that our system does, each of them would still have (on average) 340 years’ worth of our total solar output in deuterium stocks, or 22,000 solar-output years, total.
Boosting this much deuterium to one percent c velocity, for transport between star systems, so we can get it within one thousand years, is actually not very expensive energy-wise. At good efficiency, perhaps using linear accelerators, all the gas-giant deuterium orbiting a sun-like star (more than 3000 solar-years output worth) would cost less than a 30 years of stellar output at the “mining star,” to send our way. It should be worth doing, even with the inefficiencies involved (remember, if we use a star to power a laser for a solar sail launcher, then we’re effectively using a working light hydrogen reactor to ship heavy hydrogen to us). We can slow it down at this end via solar sail, by using solar-powered coherent beams of radiation, at the same cost. Think of the future as involving streams of deuterium coming into our system from other nearby star systems, something like water converging from all over California into Los Angeles. Let us hope there are no aliens to object to this redistribution of galactic resources.
These issues are framed in terms of bringing resources home, because there’s no point in going to other star systems if we’re going to stay there and use the star's resources that way (see colonization issues below). The whole point of collecting energy from other stars and sending it back to central place for use, is that it can be used to power a massive computational system which is compact and not limited by connection delays at the speed of light. So long as light-speed remains a barrier, people in other star systems will forever remain individuated by their distance: they won’t be “us.” If we want to use other stars’ energy for our thinking, that energy will have to be sent to our solar system.
As our probes travel outward, converting the gas giants of other star systems into deuterium mines which are sending a stream of energy home, the automatic process becomes more lucrative, because the shell of expanding automated mining probes moves outward like a fire, encountering more and more stars with each passing year. The number of stars per year and the power we receive from mining their gas giants increases with the square of time, and as the cube of the speed at which the probe wall proceeds, and deuterium returns back. Deuterium mining pays off almost immediately, as the nearest star is four light years away, which is 400 years away at 1% c, and so in 800 years (if the star has gas giants) we will get back anywhere from 300 to 3000 solar-years of energy output from its deuterium. Thus, the process has a good chance of already outpacing our sun’s output in the first thousand years, and will only improve after that, as we reach more stars. By the Year Million, our expanding wall of mining probes ten thousand light years away should be evaluating around three thousand new stars per year and (assuming life is scarce, so ecological plunder is not an issue) we’ll be getting back 255,000 years of total-solar-output energy from these, every year [7]. Of course this assumes that it will all come to us; instead it will more likely be divided between our colonies and us. Struggle for natural resources won't end when humans leave our solar system.
Finally, deuterium is a convenient storage option which allows use of far-away resources: deuterium can be manufactured at great energy cost, by making neutrons and combining them with light hydrogen. No one would want to make deuterium this way unless they had large supplies of otherwise unusable energy to “burn,” but that may be the case with energy radiated from lifeless stars which have no gas giants and thus no native deuterium supplies. How can we best condense the output of such stars and get it back to us? Beaming it for long distances is not a viable option. Antimatter energy storage has been suggested, but storing large collections of antimatter (which blows up on contact with any normal matter) presents possibly insurmountable problems, which are outside our scope here (however, they are not outside the laws of physics). However, the much more tractable deuterium is likely to remain the future energy currency of choice. So deuterium may one day be produced inefficiently as a synthetic fuel at stars which have no gas giants or planets ever likely to support life, simply because the energy of such stars goes to waste otherwise.
Limitations in Manufacture, and Thus in Time and Space
We have a lot of exploring we can do, if we choose to do it. In a million years, even probes traveling at one percent c can reach stars ten thousand light years out. This is a bubble ten percent of the diameter of our own Milky Way galaxy (which is about half that thick), and one percent of the disk’s area. If we cover one percent of the galaxy’s three hundred billion stars, that is some three billion stars. Perhaps the number will be less, since we’re not close to the star-rich galactic core, but at least a billion stars to explore in that time and volume seems likely.
So far as we can tell, self-assembling nanotechnology appears physically possible. This means that any physical assembly that’s conceivable, stable, and made of atoms, can probably be manufactured to order, as many times as one would like. This includes biological organisms, which should be as copy-able (and transmittable) as photographs are now [as discussed by Wil McCarthy in Chapter 6]. Self-replicating factories should be able to take hold in any solar system, building whatever can be compiled out of ordinary matter, using matter and free-energy available locally. These factories will be limited by physical laws, but such limitations (even by the laws we know) are wide.
Alien cultures (if any exist) and certainly our own culture (if it survives) are thus likely to develop general atom-by-atom constructor and duplicator machines. These should be able to remake themselves on the basis of a seed amount of information, wherever time, energy, and suitable matter is available. The seed amount of information for a universal constructor appears not to be large. Ribosomes in living cells come close to being universal constructors of biology as we know it, and they (and the DNA/RNA information that drives them) are fairly simple and small. Doubtless, constructors of greater ability are possible and will need more information, but information storage can be much more compact than the 165 daltons per bit that DNA uses in Earth organisms. A packaged fertilized seed is a good metaphor for any Von Neumann-type self-replicating machine.
One of the corollaries of these ideas is that, if we meet space-faring aliens from other stars, they are extremely unlikely to arrive Star Trek style, as some kind of “meat in tin cans,” traveling between stars in faster-than-light vehicles. Instead, sub-light velocities will be used, and matter sent at these relativistic speeds will be kept to a minimum, in order to conserve energy, and allow for decent speeds which maintain the ability to stop at the end of the journey. The problem of slowing matter down at the opposite end of exploratory star-flight is far worse than the problems of accelerating it at the beginning (which can be done electromagnetically, or by laser). A probe at one to ten percent of the speed of light cannot be stopped directly by a solar sail. Schemes that use giant mirrors to decelerate other sails or mirrors, powered by coherent electromagnetic beams from another star, are not convincing. Solar sails work well for star-flight only with lasers or masers at both ends. Landing at the “other end” is hard if there’s no technology there.
A better idea seems to be “Orion” type craft (powered by those mini-nuclear bomb blasts) which possibly can be constructed to travel at 0.01 c. However, the material costs required are gigantic. Another alternative for hardy seed probes is “photosphere aero-capture,” a one-time pass through a target star’s atmosphere, allowing gravitational capture of the craft. This is difficult: only the outer photosphere (visible to us as a star’s “surface” and typically about ten percent Earth sea level gas density), is suitable for the job, but so thinly-layered that a probe will have to skip through it like a stone off water. Too shallow a dip will give too little energy loss; too deep and the frictional temperatures and g’s will destroy any object.
Probably only diamonds and carbides, cooled by ablation, will survive a one to one hundred second photosphere pass. Electronics now survive tens of thousands of g’s in artillery shells, and the Galileo Jupiter-entry probe reached 15,000 oK (far hotter than the photosphere) in a two-minute deceleration from 0.00016 c. Both parameters are far from ultimate limits. Photosphere aerocapture requires a minimal perihelion velocity of 0.002 c (escape velocity from a sun-like star at that distance) but allowed excess interstellar energy above this, will be set by the toughness of the seed [8].
Such star-diving probes can be small (the size of a grapefruit or basketball), because they don’t need reaction drives. After slowing to capture velocity, a probe must pass though the target system looking for suitable rocks to impact, then turn into welcome stations. These factories can then be used to build first habitats for visitors, and then finally compile the visitors themselves. Think of Sir Arthur C. Clarke’s Rendezvous with Rama, only on a much smaller scale. Whether the information for doing all this comes with the probe, or is sent later by laser, is a detail. (Probably the laser will be needed in any case, to send updates for those users of Windows still struggling to eliminate worms and viruses.)
So forget multigenerational ships full of organisms like humans, or even cryonically preserved organic bodies awaiting medical resurrection. We shall long since have redefined “human” and become able to make intelligent matter and its mobile representatives, as needed.
We Meet The Aliens: Hey, They Look Just Like Us!
Which is to say, they look like everything, but not particularly like anything. “Humans” exploring another star system will be able to take on any material form (a range of them) needed for the environment to be explored. One might expect space-faring aliens to do the same. Indeed, given the “culturally driven” evolutionary changes that might happen over the millennia needed for star travel (remember the exceedingly fast thought process that might proceed even aboard spacecraft, and certainly near the departure star), we might find that “meeting ourselves” at another star is just as odd and exciting as meeting aliens who have evolved independently, somewhere else. After all, once technology allows full control of body design, then designs will be based on the physics of the universe plus input from imagination and art. Probably those results will both converge (on some issues) and diverge (on others), to the point that when it’s all over, it might not matter how any given intelligent race started, or what it looked like originally. By that time we’ll all be our own artwork and engineering, with no traces left from previous slow evolutionary segments. This may be a little disappointing, but worse news is to come.
Navel-Gazing and the Fermi Paradox
You might ask yourself: if we don’t expect to meet anybody more interesting in other star systems than whatever we ourselves can become in a fraction of the time it takes to go there (and, what’s more, we will be able to assess aliens remotely, at light speed, before committing ourselves to the teleportation journey)—then why go all the way out there in the first place?
In 1950, nuclear physicist Enrico Fermi realized that space-faring aliens should be quite common, and many of them should have evolved ahead of us, and some of these evolved to high-technology while located inside of our light-cone. It would thus be quite possible for them to have visited us by now (even without warp drive). So, Fermi famously wondered, where are they? Why have they not even sent radio signals or shot seeds at us?
A popular answer, in brief: the average high technology phase of an intelligent civilization, counted from their first radio wave emission to the time when they stop emitting, is quite short—perhaps through technical regression, or destruction of their species. Perhaps, as Carl Sagan surmised, aliens always blow themselves up in nuclear wars shortly after discovery of radio, and that’s why we never hear them [9].
Today, we see that self-reproducing assembler-type nanotechnology is a far greater danger than mere nuclear war; in fact, simple nuclear device manufacture from ordinary uranium-bearing minerals (or even the uranium in seawater), is just one subset of nanotechnological prowess. Possibly high-tech aliens have all long since dissolved each other in assembler-wars or accidents, and are now melted into what futurists call “gray goo.”
Sagan would never have considered a gentler, more renunciatory alternative, in which aliens are able to explore, but choose not to. Star travel using faster-than-light drive involves paradoxes, and star travel at sub-light speeds is boring. The faster your thought processes, the longer and more boring a typical trip is. One can slow the rate of thinking, so it feels as if you’re moving faster, or even suspend it by turning off consciousness (or by riding on a light beam as information). But what’s the point? The civilization you left behind will be roaring ahead, and you’ll want to talk to somebody when you get home—unless you’re prepared to stay away for good. To our descendants, already trying to squeeze the last yottaflop [10] of performance out of their M-Brains while orbiting the fires produced by stripping nearby lifeless solar systems of their deuterium, real star travel with its attendant time lag and low energy availability for thinking on the journey, is going to be monstrously un-fun. Especially if it is effectively a one-way proposition.
Consider the alternatives. For a hive-mind of huge mental capacity (such as the future human race), any type of virtual reality (VR) experience can be had for the asking, as in a dream, but with better control and richer content. Why bother constructing a thing or a scene in the “real universe” (waiting out the constraints of material inertia, while the atoms get to where they need to be), when you can have the same experience, or a better one, in virtual reality? As thinking rates more and more outpace atom-moving rates, the quality disparity between the two types of experiences is bound to diverge more and more. Real world exploration is bound to suffer—and if residual curiosity about the real world bothers you, it can be turned off.
The Problem of Motivation.
The philosopher Schopenhauer first noted that “A man can surely do what he wills to do, but he cannot determine what he wills.” In the film Lawrence of Arabia, the character of Lawrence makes a similar observation (”A man can do whatever he wants… but he can’t want what he wants”), then plucks at his flesh and adds: “This is the stuff that decides what he wants.”
In the future, all this will change once the flesh can be engineered. At present, most of the time, humans are driven by our old monkey/mammal motivations, even if we’re not always aware of how much we are. We want recognition (as stand-in for better reproductive choices, or a shortcut to it); we want money (ditto); and we want education and entertainment. All of this will shortly be irrelevant. When our drives are no longer controlled by chemistry, or we can modify any drive, feeling, or desire by act of will, then what will happen to us? What determines our will, our desires, when we can determine what our will shall be? It’s a conundrum without bottom. Instead of exploring, you can stay home and watch Seinfeld (or the future equivalent) with your pals in the Matrix. Except that this fine Matrix won’t be a product of evil computers. Instead, it’s one you control, so you can be your own Neo or any other Matrix character, without guilt (which can also be turned off if you like, along with the memory of having had it and why). This is choosing the red pill with a vengeance.
Versions of the answer to Fermi’s paradox thus involve the aliens all staying home in their orgasmatrons, not caring to go out across those cold light years simply to probe funny-looking primates like us. And what happens when a brain stimulation fully emulates what certain drugs are already reputed to do: obtain the complete satisfaction of achievement, without the user having done anything to “earn” it? Even the disheartening feedback of non-accomplishment is in theory controllable. One can always think of things just as important as exploration and being an astronaut. And when you think about them, you can feel just as good as if you had been an astronaut.
Perhaps the world of reality will end up populated by beings who for some reason made the choice—or have been coerced—to spend some or all of their time “here” in the world of time and space, instead of in the virtual world which runs much faster and promises ecstasy at command. Once it seemed to me unlikely that one hundred percent of minds could be susceptible to any “drug” or “infection” or “fad.” But that might be wrong. The world of “reality” in the not-too-distant future might end up as nothing more than the province of the “drug-resistant” (VR-resistant). Ultimately these people might not be there of their own choosing; the idea of choice might not even apply. Intelligence and sapience can exist without self-awareness, a fact we all know from losing ourselves completely in a good film or a hard math problem. This state can be made permanent. It’s a good one to provide slaves with.
“Reality-Slaves” may be kicked out of “Heaven,” in a sense, and sent back as probes and servants into the external universe by Master lotus-eaters who have nothing better to do. This idea resembles Gnosticism; as an explanation of our present reality, it’s a bit perverse. These exiled minds (self aware or not) might perform needed real-world maintenance for the Lotus-class and their computronium and power-sources.
The Deuterium Mines—You’re Guaranteed to Like It, If You “Like” Anything
Perhaps navel-gazing really does finally capture all tech-savvy races, and is the answer to the Fermi paradox. That wouldn’t be very interesting from our present point of view. But it’s an uncaring universe—one that probably allows for all its cruel, heedless aspects to be dialed away (at the cost of living in Disneyland), so long as you can program a few slave entities to be happy to go out and get your deuterium for you, or else not think about it at all, any more than a fish presumably thinks about water. So long as the VR habitués feel entitled enough make deuterium-collection the focus of fulfilled existence for some other reality-stuck types, those slaves will be able to do most things. Except change that one little coding feature that makes them phobic about going into virtual reality themselves.
It’s rather hard to imagine resisters of the VR opium voluntarily going back to the Hell of reality, just for the hell of it. Or even voluntarily doing so in small bites, because (like the old joke) it feels so good when they stop. Miltonic arguments aside, obviously someone will need to “force” them to the deuterium mines and the solar radiator maintenance and whatever else needs to be done for the Dyson swarm. Although, as we noted, even the ideas of “force” and “free will” become a bit mixed if the only sense of purpose of a designed creature is the satisfaction of your desire. Yet without such meddling, the pleasures of traditional attractions in “reality” (including those of power, sex, childcare, etc) can never compete with unadulterated pleasure itself, available in virtual reality. We organic beings have found something like that already, in crack-cocaine and “crystal meth”—for some, these brain-modulators simply supersede all evolutionary drives.
Looking for High-Tech Aliens
If they aren’t looking for us, then we can recognize them only by how they make their living, like recognizing the lair of a hidden octopus by the litter of crab and clam shells that surround it. Advanced aliens presumably aren’t radiating radio, nor are they looking for us—long ago, they’ve imagined far more entertaining stuff than they’ll ever get from watching us or talking to us (given our current slow mental processes, it’d be like conversing with a tree). Probably they’re busy entertaining themselves, at a unimaginable multiple of our clock-rate.
Can we recognize them from afar? Theoretically, energy can be extracted by an advanced civilization from rotating black holes, but at present we have no good way of figuring out how fast black holes rotate even when we spot them (the closest we’ve found is sixteen hundred light years away). If our Sun is not near enough to advanced aliens to represent a handy deuterium source, then they (or their agents) probably aren’t even interested in our solar system. Presumably all we’d have to worry about is the theft of a few outer planets—but we can hope they’d choose to talk to us before doing anything so crass as engineering an already-inhabited solar system. Natives who are low-tech today will be upset about loss of their non-sustainable resources tomorrow. (If aliens do show up, expressing a diffident interest in our spare gas giants, let us not sell mining rights for a few glass beads).
Any far-off alien presence in our slow reality probably doesn’t look special. At home, they probably appear as cool dust-clouds cloaking good high-output stars. If the manufactured dust-clouds are effective thermodynamically, then we won’t see the star, just the local Dyson-cloud glowing uniformly from within, at a much lower black body temperature. Interestingly, some nebulae and galactic cores do look a bit like this. A “gas” halo glowing at blackbody temperatures of five hundred to a thousand degrees Kelvin will appear as an unnatural “giant brown dwarf”—a star far too large and energetic to have the low temperature it shows. Unfortunately, we cannot directly assess sizes or “luminosities” (energy outputs) of any but very near stars, so any “brown giants” that exist may fool us by appearing falsely to be ordinary (natural) brown dwarves that are taken by us to be more distant than they really are. Without a way to resolve them or see parallax, we cannot know. Similarly we might hope to intercept communications beams from such immense habitats but, alas, the more efficiently-compressed a communication is, the more it looks like natural electromagnetic noise. We can expect superintelligences to compress their inter-mind and intra-mind chatter thoroughly indeed.
But there is something else that’s hard not to reveal—something our own heroically modified solar system will show by the Year Million. In the neighborhoods of so-called Kardashev Type II civilizations—those powerful enough to use their whole solar output—some planetary systems will simply be missing their gas giants. This alone will be hard to infer, but in these systems, clouds of hydrogen gas could be spreading that are suspiciously free of the deuterium originally present since the Big Bang. Has anybody looked for missing deuterium in molecular gas clouds near other stars? The spectra of the two types of hydrogen are subtly different, so our present telescopes may not be capable of seeing that a trace is not there. Eventually, however, our vision will improve. If we see light hydrogen without any deuterium, in a cloud that seems never to have been part of a star, that can mean only one thing. No natural process could do this, as we can see from studying our own solar system. A star didn’t eat all that deuterium. Somebody ate it.
Notes:
[1] The best exemplar for the first early humans who were nearly modern in their post-cranial skeletons, is the so-called Turkana Boy (KNM-WT 15000), a teenaged male who died about 1.5 million years ago, and whose skeleton was found almost completely intact. He has also been classed with Homo erectus, and is at least a close precuror of this ancestor of ours. Debate continues on whether H. habilis is the ancestor of H. ergaster, but even if not, the ancestors of H. ergaster must have been something roughly similar.
[2] Frank W. Marlowe, “Hunting and Gathering: The Human Sexual Division of Foraging Labor.” Cross-Cultural Research, Vol. 41, No. 2, 170-195 (2007). DOI: 10.1177/1069397106297529. SAGE Publications.
[3] This assumes a neuron with volume of 5000 cubic microns, which is in the middle of a widely variable range.
[4] Such creatures are not at all what Norbert Weiner had in mind when he coined the word cybernetic, but we’re stuck with the alternate meaning of the word as referring to a human-machine meld, and there’s little we can do about it.
[5] See http://fusedweb.pppl.gov/CPEP/Chart_Pages/5.Plasmas/SunLayers.html. Accessed October 25, 2007. For volumetric power output densities for regions in the Sun.
[6] 3400 solar-output equivalent years = 0.00005 x (0.84) x [80 billion years/1000]. The 0.00005 is the mass fraction of deuterium in Jupiter’s hydrogen (25 deuterium atoms per million hydrogens), the 0.84 factor is due to deuterium fusion being only 84% as efficient as light hydrogen fusion on a per-mass basis, 80 billion years is the theoretical time to burn all of the Sun’s hydrogen, and the factor of 1000 corrects for Jupiter’s smaller hydrogen mass as compared with the Sun.
[7] The number of stars explored per year dN/dt = 4 π ρ r2 dr/dt = 4 π ρ r2 v = 4 π ρ v3 t2 [eq. 1] by a probe shell moving outward speed dr/dt = v (assumed the same return speed as the probes move), into a galaxy with stars at a volumetric density ρ, If A is given in solar-output-years of energy, ρ in stars per cubic light-year, v as a fraction of c, the speed of light, and t is in total years of probe outward exploration.
Power P, the rate at which energy is returned from deuterium mining as a function of the time t at velocity (assumed the same return speed as the probes move), where a mean deuterium energy supply at each star is A, is given by is given by: P = π ρ A v3 t2. This equation for P results from inserting t/2 for t in eq 1. (since energy returned is at a time only half of the star discovery time) and setting P = dE/dt = A [dN/dt]. The figure given in the article assumes a star density of 0.00024 stars/cubic light year (this will be hurt by the flattened distribution of stars in the galactic plane), and a probe and deuterium return speed of 0.01 c, and an estimated average of 280 solar-output years of deuterium in the gas giants of each target star (as compared with an estimate of 3400 solar-output years contained in Jupiter).
[8] For slower speeds, there is a bit of help, in that aerocapture is a gravity-well maneuver, so that speed loss needed in a photosphere is not full interstellar speed, but only the speed associated with energy in excess of solar escape energy at photosphere distance. For example, a probe crossing between Sun-like systems at 0.002 c will arrive at the target photosphere at 0.00283 c (not .004 c), and thus needs to shed only 0.00083 c for capture, which is less than 42% of its interstellar speed.
[9] Or (Sagan was an optimist), perhaps some aliens learn to get along, and these are the ones we’ll hear. When we tune into what they’re saying—sometimes Sagan’s optimism could be wild—their communications might contain helpful tips on how avoid nuclear annihilation (see Chapter 22, The Quest for Extraterrestrial Intelligence, pg. 324,
Broca’s Brain, Ballantine Books, NY. 1979).
[10] Computer power is expressed in FLOPS (FLoating point Operations Per Second).
The largest computers today are in the petaflop range (1015 operations per second). A yottaflop is a billion times faster, and is the fastest operational power we presently have a defined word for.
