ALT! WHO GOES THERE? PART 6E.
It has been said that questions regarding the nature of posthumanity are unanswerable. But while some aspects of posthuman existence may well be beyond lesser beings such as ourselves, the fact remains that a post-human is as bounded by physical law as any inhabitant of the Universe. Assuming no radical re-write of these laws, we can use the ultimate limits to draw up some sketches of the post-human world.
So, let’s start with bodies. Some people in the transhumanist community have been pushing to rename ‘mind uploading’ ‘substrate-independent minds’, the point being that a mind needs a brain and a body (or at least a perception of being embodied) but a software emulation of a mind can be copied and transferred to any substrate capable of supporting it. What is the smallest space you could pack a human-scale intelligence into? Hans Joachim Bremerman worked out the maximum computational speed of a self-contained system, a calculation he derived from Heisenberg’s uncertainty principle and Einstein’s principle of mass-energy equivalency. ‘Bremerman’s limit’ allows for approximately 2.56 times 10^47 bits per second per gram. According to Ray Kurzweil, mind uploading requires 10^18 bits, so every gram of matter potentially exceeds the capacity needed to run an upload by 29 orders of magnitude. So it seems fairly conservative to say that a 0.0026 gram grain of sand could (if its atoms are organized to perform useful computations) provide enough bits to accommodate an upload. At the other end of the scale, the speed of light imposes a limit on how large a body can be before signal delays slow reaction time to an extent that each unit has to act as a separate decision-making individual. According to Hans Moravec, the light speed issue, plus other factors such as the need to radiate waste heat, makes a posthuman ‘massing more than a 100km asteroid implausible’.
A limb or a sense organ need not be physically attached to the body in order to feel like a part of the individual. All that is required is a feedback loop whereby an act of will corresponds with a change in an object. In other words, if I consciously decide that a remote robotic manipulator will grasp an object, and it does so, my body map will expand to incorporate that remote device as part of my body. This would be especially true if the remote manipulator fed information back to the brain (maybe its gripper feeds tactile information to the brain, so that the mind perceives the weight and texture of anything it moves).
Remember, that many tests have revealed that, as far as the mind is concerned, there is no fundamental boundary separating the body from the rest of the world. The perception of the body is fluid and can expand to accommodate anything that comes under mental control. Currently, much of the world does not directly respond to one’s will and so is not incorporated into the body map. But some nanotechnologists believe we may one day build incredibly tiny robots that can interact with each other to form physical 3D objects. One such group pursuing this idea is Canergie Mellon-intel’s ‘Claytronics Research lab’. The goal here is to create a millimetre-sized spherical robot called a ‘Catom’ (short for Claytronic Atom), capable of emitting variable color and intensity of light, and which can move in three dimensions relative to other catoms. Ultimately, with sufficient numbers of catoms and the ability to map a 3D object (maybe using techniques similar to those that scan physical objects in order to recreate them as virtual copies) the catoms would arrange themselves into a copy of that object. Among other things, this could be used for tele-presence. Your catomic double interacts with a friend who is in another place, copying your every gesture, expression and movement. Meanwhile, of course, your friend’s catomic double is in your living room, copying her movements.
Homes are already seen as an extension of the self, and this perception would be emphasised by claytronic furnishings, decorations and appliances that appear as if by magic in accordance with the occupier’s wishes. If towns and cities were made out of catoms, the resulting experience might be akin to a lucid dream where the very environment continually adapts and reshapes itself to your heart’s desire. Of course, aesthetic taste differs from person to person, so a claytronic city configured to match one person’s idea of heaven would no doubt correspond to another’s vision of hell. Territorial battles over the right to rez whatever takes your fancy, versus the neighbour’s right to a harmonious environment, may therefore move out of Second Life to become an ongoing issue in C-city.
Talking of people, who will they be? It is possible that one uploaded mind could become many individuals. The reason why this is so, is because software can be copied. Once a whole-brain emulation is created, you can run as many copies as your hardware has capacity for. And, as we have seen, a gram of computronium (matter organized to be as efficient as possible at computation) has the capacity to run trillions of uploads. One thing that a human mind is capable of is running internal models of the world that includes a model of oneself. In other words, we can imagine a scenario, place ourselves in it, and predict the outcome of a choice. It seems very doubtful that more than a scant few imaginary selves have the full mental capacity of a person, simply because the brain of that person has to divide its processing power to run these ‘partials’. But, a post-human, running on hardware that vastly exceeds the capacity of a human brain, could run vast numbers of ‘partials’, each one of which equals the original human mind that provided the scan data from which the whole-brain emulation was created.
THE ECONOMICS OF UPLOADS
Why run copies of yourself? At the very least, for the same reason some people run multiple avatars: So as to participate in multiple events simultaneously; to explore alternative identities. As well as these possibilities, the economist Robin Hanson argued in favour of an economic incentive for copyable minds. This argument rests on the assumption that it will prove easier to scan a brain and run a copy than it would be to build a human-level general intelligence from scratch. “Imagine that we learn how to take apart a real brain and to build a model of that brain by identifying each unit, its internal state, and the connections between units…Software with human-level intelligence, yet created using little understanding of how the brain works on anything but the lowest levels of organization”. One of the hopes of the H+ movement is that uploads will live in a world where money is not necessary. A combination of molecular nanotech (capable of copying any material object by arranging atoms and molecules to produce exact duplicates) and a vast organization of robotic workers happy to run the world economy free of charge, will leave people free to pursue an endless, hobby-filled life of leisure and luxury. But, what if building such intelligent robots proves more difficult than copying human brains? What if the hardware that runs these uploads and the tele-operated bodies they use to interact with the world are not given away free to anyone who asks for one? Or, worse still, are actually expensive? In that case, while the robotic workforce may well have automated much of production and overall wealth may have increased, there would still be jobs that require people and an economic incentive to pursue them. You need money to buy or rent the hardware you run on.
Some people take more than one job, and an upload could take on dozens, hundreds, thousands. One worker could potentially fill every vacancy that matches his or her skillset. No need to quit one job in order to begin another. Instead, you duplicate yourself as many times as capacity allows and send the copies to drive the tele-operated bodies or avatars ready and waiting in the other workplaces. It gives a whole new meaning to the term ‘family business’ and copies of yourself have a distinct advantage over raising children in the hopes that they will one day take over the business. It takes decades to raise a child, whereas a duplicate begins life knowing exactly what you know. As Damien Broderick put it, “from day one they are as adept as you, and can earn their own keep”. Perhaps companies would pay for the cost of uploading and embodiment, provided the employee agrees to be duplicated, thereby maximizing his productivity.
AN OLD DEBATE
It must be admitted that this scenario stretches the concept of ‘me’ to breaking point. Surely, what we have here are people who only think they are me? But then, why am I anything more than a person who believes herself to be a continuity of someone, an assumption based on the great store of MEMself available for access? What makes me special, compared to any other Extropia with access to the same MEMself as me?
This fission scenario, in which one person becomes two, has been debated in philosophical circles for a long time, and in fact it pre-dates mind uploading itself. In some schools of thought, the prefission person is not identical with either of her post fission descendents. In other words, if I were duplicated the result would not be a copy of Extropia emerging from booth B while the original emerges from booth A. It would be a copy that emerges from both booths. The original is lost. I suppose you can compare this to an ameoba that starts off life as a single-celled creature but then fissions, so that where there was one now there are two. The two cells are not the original cell and a copy, but two descendents of the original- two copies- the original being destroyed in the process. I should point out that this thought experiment is not intended only for those uploading processes where the original brain is destroyed but also non-destructive uploading where the original brain is unharmed while a copy is produced. Whatever the procedure, be it destructive or non-destructive, this school of thought contends that each fission descendent is different from the prefission person.
Another school of thought argues that we should adopt a four-dimensional view of identity. The idea here is that what exists are ‘time slices’ or ‘stages’, and an individual is an aggregate of momentary person stages existing through time. This view does not contend that a person is lost and replaced with two other people after duplication has occurred, because there is never any such thing as a person. All that really exists are aggregates of person stages (each ‘time slice’ existing for short intervals of time). According to Raymond Martin and John Barrasi, “according to a four-dimensionalist, in a fission example the prefission person does not cease. Rather, what happens is that the prefission person stage becomes a shared-person stage…in a four-dimensional model, no-one ceases in a fission example (and, hence, identity is never traded for other benefits). Rather, what happens before fission is that the “prefission person” is really a shared stage of two persons. After fission, these two persons no longer share any present or future stages”.
YOU ARE YOUR COMPETITORS
We saw earlier how the notion of a job for life is expected to become obsolete, and work patterns may even become so flexible that companies bid for workers on a day-to-day or hour-to-hour basis. If uploads come first, this could be problematic. You might very well find yourself in employment one day, and unemployed the next. But, when you try to get another job in your areas of expertise, you could well discover they are already taken…by you! You would be in competition with yourself to secure that contract so what advantage can you offer? Skills and knowledge are unlikely to offer much of a distinction between yourselves. Perhaps the only option would be to offer your services for lower wages. According to Hanson, “upload population growth would be highly selective, selecting capable people willing to work for lower wages, who value life even when life is hard. Soon the dominant upload values would be those of the few initial uploads with the most extreme values, willing to work for the lowest wages”.
So, the upload job market could become extremely competitive, encouraging every employee to become as productive as possible. Those who could afford the most powerful and speedy processors would have an advantage. If your brain processes information a million times faster than your rivals, you could take a few moments to think up business plans that require centuries or millennia of thinking time by your competition. “Strong social hierarchies might develop; some might even be ‘gods’ in comparison to others”, reckoned Hanson.
God, it is said, made man in his image. But the fastest uploads may find a physical humanlike body to be a disadvantage. Their minds would process information at speeds orders of magnitude faster than biological human brains. From their subjective point of view, the world of humans would be one in which everything happens at a crawl. Decades would seem to go by without anything much changing at all (I am talking about changes wrought by biological humans, not that caused by uploads, cyborgs and other trans or post human beings). Also, if their minds were teleoperating humanlike robots, they would find such limbs physically incapable of moving fast enough to match their reaction times. Desiring bodies that can perform motions hundreds of times faster, Hanson reasoned that such uploads may inhabit tiny 7mm bodies, with arms and legs moving at 260 times the rate of human limbs, and running on 16 watts of power. “Such uploads would glow like Tinkerbell in the air, or might live underwater to keep cool”, thinks Hanson.
Hanson anticipates that these bodies will still be humanoid, only very tiny. There is a problem with this notion, because the laws of geometry would require a modification to body plans. Those B-Movies in which ants become the size of elephants or people shrink to the size of insects ignore the fact that such a massive ant would break its legs unless they were redesigned to be more like, well, an elephant’s. Similarly, a human body scaled down to insect size would have to be modified into a body shape best suited for life at the millimetre scale. These uploads would inhabit insect-like bodies. Of course, they could also live in VR where the laws of physics can be set to allow for ant-sized elephants and elephant-sized ants. But if these uploads were engaged in work that would be useful to the real world, fantastical virtual realities would be disadvantageous. Much better to have a world that reflects the rules of physical reality. “Since, compared with the software we write, the human brain seems especially good at dealing with the physical world…many uploads should remain familiar with the physical world for a long time to come”, reckoned Hanson.
THE NEW ERA
As with anything, the upload era would end eventually. We would one day progress from understanding just enough about the lower levels of brain organization in order to copy brains, and move onto the next stage, in which we develop an understanding of the higher levels of organization and operation- when we understand how brains work. Then, we will be in a position to design smart software and robots from scratch. Moreover, the robots, uploads etc could also understand the workings of their own brains and use their superior mental powers to design even better brains, and so on in a positive feedback loop. The era of the superhuman artificial intelligence would begin.
Some transhumanist thinkers like Hans Moravec and Ian Pearson see the rise of robots as providing the biggest incentive to upload. Simply put, if you can’t beat ‘em, join ‘em. In other words, if your shitty biological brain is so sluggish in comparison to the radically-improved brains AIs and uploads possess, if you are a moron in comparison to the fundamentally smarter people around you, you would be crazy not to junk your obsolete mind for something more capable of useful activity in this brave new world.
At the same time, as uploads and SAIs proliferate and outnumber biologicals, they would want to alter the world around them to their own advantage, just as humans have shaped landscapes to serve their needs (often at the expense of other species). And what do software-based people need? Computation, the more powerful the better. Post-humans would realise that there is a gargantuan amount of computation going on around them, because within all apparently solid and immobile objects, atoms are in motion, rapidly-moving electromagnetic fields are being generated, and particles are changing spin. According to Ray Kurzweil, electromagnetic interactions alone amount to “at least 10^15 changes in state per bit per second going on in a 2.2 pound rock, which effectively represents about 10^42 calculations per second”. To put it in perspective, consider that 10^42 cps is ten trillion times more computation than that performed by the entire human race.
The catch lies in the fact that these computations are random and not doing anything useful. So, the post humans would endeavour to organize the rock’s atoms and particles in a more purposeful manner, ultimately creating a 2.2 pound computer that could perform the equivalent of all human thought in ten thousand years every ten microseconds. If the era of copyable people resulted in family business run by duplicates of an original human, the post-human era might be one in which entire world populations exist in specks of dust, all descended from one upload or AI.
Eventually, the post-humans would reach fundamental limits to how small computational elements can be built. Perhaps, at this point, computational growth will spread outwards. Of course, this happens already, because the number of chips we produce is currently expanding at a rate of about 8.3% per year. Once fundamental limits to miniaturization are reached, and assuming post-humans are not content with the amount of computation at their disposal, outward expansion of computronium would be the primary form of growth.
How would this alter the environment? Some thinkers like Anders Sandberg and Ray Bradbury have outlined mega scale construction projects that would maximize computation in the post-human’s local environment. Bare in mind that we are talking about post-humans here, weakly godlike intelligences, so by ‘local environment’ we do not mean a city, country, or even the planet. We mean the solar system.
One such mega scale engineering project is “Zeus”, a ‘Jupiter Brain’ outlined by Anders Sandberg. Zeus is a sphere of nearly solid diamondoid, with a radius of 9000 km and with a total mass 1.8 times that of the Earth. Its computational elements consist mainly of reversible quantum dot circuits and molecular storage systems. Each processing element acts as a semi-independent unit, communicating with other nodes using either fibre optics/waveguides or directional signals sent through vacuum. According to Sandberg, “since the many processing/memory nodes need to be close to each other due to the many short-range connections, the possible distributions are either a central core surrounded by connections, a cortex with connections through the interior, or distributed clusters of nodes in the interior. Of these the cortex model is most volume-efficient, and it will be assumed that Zeus is organized in the same way”. A concentric shield surrounds the central sphere, both to protect the processing/memory nodes from the hazards of space, and to dissipate heat via radiators. The number of operations per second Zeus can perform is between 10^49 and 10^61 ops. Lower estimates assume nodes acting as single processors, whereas the high estimates assume clusters of parallel processors.
It can be difficult to grasp the size of a number like 10^61, but Hans Moravec made some calculations that help put it in perspective. Suppose it takes one hundred million megabytes or 10^15 bits to run a human-equivalent mind. Furthermore, suppose you need a thousand times more storage- 10^18 bits- to encode a body and its environment. A large city with a million inhabitants might require 10^24 bits. Notice that we are still nowhere near 10^61 bits. In fact, if Moravec’s estimates are correct even the efficiently encoded biospheres of a thousand galaxies (assuming one planet with a population of 6 billion per star per galaxy) would require a ‘mere’ 10^45 bits! It therefore seems possible that a Jupiter brain could simulate whole universes with planetary systems populated by hundreds of billions of simulated people, running through hundreds of millions of years of evolutionary and cultural history during a mere ten microseconds of thinking time.
This might all sound crazy, but we already use agent-based modelling in which adaptive rules permit behaviour to change in response to previous interactions, and we also run simulations of astrophysical events like planetary formation or supernovae. There are plans to create a model with 10 billion agents- the first simulation of the development of an entire planetary population. Of course, each agent is not as complex as a human- I doubt they are as complex as microbes- but you only need to extrapolate out to the feats of computing as permitted by physical laws, as well as other areas like whole-brain emulation and planetary modelling, to see that this is all within the realms of possibility.
RELATIVITY OF REALITY
Some thinkers (most notably, the physicist Bin-Guang Ma) have proposed a ‘relativity of reality’. In relativity theory, ‘motion’ has no absolute meaning. You can only say if something is at rest or in motion if you have a frame of reference. ‘Relative to the station, the train is in motion; relative to the train, the passenger is at rest’. Applied to reality itself, relativity theory would say there is no absolute meaning to reality. You can only tell if a world is simulated when there is a reference world it can be compared to. In the case of online worlds like Blue Mars, we can use the physical world as the frame of reference and say for sure that Blue Mars is a simulation. But if there were intelligent software beings within Blue Mars, who had no contact with physical reality and knew only their own little ‘universe’, Blue Mars would be the primary reality for them and they could not know their world was a simulation running on machines in a more expansive and ancient reality. Similarly, we cannot know for sure that the physical world is not, itself, a simulation. Perhaps it is one of many billions of simulated worlds existing in a Jupiter brain? In fact, any reality, even one running a simulation as sophisticated as our Universe, has no better or worse chance of being a simulation than any other.
HOW TO BUILD A BUDDHA
Uploading is often seen as a form of immortality. The body may succumb to decay, but the mind- now independent of any one substrate- can copy/transfer itself to a replacement body. However, Vernor Vinge’s speculation regarding the nature of post human existence questions whether this is life-everlasting or death redefined: “A mind that stays the same cannot live forever; after a few thousand years it would look more like a repeating tape-loop than a person…To live indefinitely long, the mind must grow…and when it becomes great enough and looks back…what fellow-feeling can it have for the soul it was originally?”.
How might such a mind grow? One way would be to incorporate aspects of others’ identity and experiences into its own MEMself. As we have seen, people already do this to a limited extent, with mirror neuron activity enabling one individual to model the mind of another. Recall also how personalities are ’constructed’ by incorporating aspects of others’ behaviours and mannerisms etc into a growing web of micros, minors and majors. It should be possible, in principle, for virtualized minds to blend together. This act of fusion throws up questions of identity, just as the fissioning of one mind into several does. But the resolution to the paradox is much the same: Just as substantial loss of MEMself results in a POVself that perceives no connection to its past self, so a substantial blend of MEMselves from persons A and B would result in a person believing itself to be neither. Or, to put it in four-dimensionalist terms, a pre-fusion stage becomes a post-fusion stage.
One might question why two souls would choose to effectively destroy themselves and be replaced by somebody else. Surely, each mind would take care not to incorporate too much of another’s MEMself, thereby retaining a sense of its own identity? However, any one mind must be limited in what it can experience. Eventually, to avoid the ’repeating tape-loop’ effect, the desire to experience things outside of the boundary of one’s identity would become overwhelming. The mind would seek to dissolve the boundaries that both define and confine it, gradually shedding its identity in order to encompass experiences outside the scope of its prior self.
Recall also how neuroscience research into happiness and meditation reveals an increase in synchrony in neural activity. Because pattern recognition requires one set of neurons rather than another being active, large scale synchrony and reduced frontal activity (also observed during functional scans of meditating subjects) leads to a diminishing sense of identity. So, we can either say that a drive toward reduced neocortical activity relative to the baseline state and higher than average synchrony among brain regions is a pursuit of pleasure, or see it as a drive towards self-annihilation. If pleasure is desirable, then so is a transition to a state where the self no longer stands out as a subjectively unique and distinct identity. Bruce Katz argued: “the real opposition is not between the pleasure principle and the death instinct (as Freud supposed) as the survival instinct and both of these taken together. Survival requires boundaries between concepts and especially the self and other; pleasure delights in the breaking down of barriers, and the reduction of specific thoughts in favour of a more distributed buzz of activation”.
Perhaps, then, minds that seek to grow beyond the confines of the self will seek each other out, merging together to form something greater. To a limited extent, individuals do this already. Emile Durkheim spoke of a collective effervesence that emerges when large groups of people congregate at a concert, sports even or some such collective experience. “This energy holds the group together; it makes each individual feel as though he or she is an element of something greater than the sum of its parts”.
As post-human selves seek ever-more sublime states of collective effervesence, Katz argued that the result would ultimately be to “manufacture Buddha. A being without division…It would not seek to look beyond itself for intellectual satisfaction, for it is truly self-contained…Perhaps, though, in its inner core, among the many thoughts it has amalgamated, would be the smallest realization that without differences, its perfection is marred. Perhaps it would choose to fracture. Perhaps we ourselves are the product of an earlier such fracturing”.
Our whole reality is a post-fission stage of a supreme collective effervesence derived from an earlier epoch in which post-biological civilizations in another dimension coalesced into some kind of technological godhead ? Our universe is an alt on its way to coalescing once more into ‘the godhead state? I think we have gone as far as speculations into the nature of alts can take us.