First let us postulate that the computer scientists succeed in developing intelligent machines that can do all things better than human beings can do them. In that case presumably all work will be done by vast, highly organized systems of machines and no human effort will be necessary. Either of two cases might occur. The machines might be permitted to make all of their own decisions without human oversight, or else human control over the machines might be retained. -Theodore Kaczynski – the Unabomber
In Wired Magazine, April 2000, Bill Joy wrote an article (1) that argued humanity needed to abandon some of our most powerful technologies because they threatened our very existence. Included in these technologies are bio-tech, nanotechnology, and robotics.
To begin with, I take Bill Joy’s warning seriously. Joy, no bomb-throwing Luddite, was a co-founder of Sun Microsystems and a core developer on some of the key technologies which have driven software development over the last twenty years. Bill Joy made his career creating technology – specifically computer hardware and software. Most people deeply involved with technology tend to be huge fans of it. They tend to be ever optimistic about its prospects and ever eager to adopt it and promote it. We do not expect a technologist to turn against technology and urge restraint in its adoption.
A famous Yogi Berra quote: “Prediction is very hard, especially about the future.” If you search the Internet for “bad predictions”, you will find many hilarious ones. Most of them are about the impossibility or uselessness of something. Other predictions are failures of imagination. Popular Mechanics predicted in 1949 that computers would eventually only have 1000 vacuum tubes and weigh 1.5 tons compared to 18,000 vacuum tube 30 ton computer of the time. Other predictions fail because they are simple extrapolations of current trends and trends seldom continue in a linear fashion. Thomas Malthus in 1798: “The power of population is so superior to the power of the earth to produce subsistence for man, that premature death must in some shape or other visit the human race.” Paul Erlich repeated the same mistake in The Population Bomb in 1968 says: “The battle to feed all of humanity is over. In the 1970s hundreds of millions of people will starve to death in spite of any crash programs embarked upon now.” Moore’s Law about the doubling of computing capacity every 18 months or two years has continued to be fairly accurate but I expect it to break down eventually. Other predictions have been correct or may be correct but have the time frames wrong. Arthur C. Clarke envisioned a world in 2001 with moon colonies and artificial intelligence (AI). We will probably will have the moon colonies eventually and probably some form of artificial intelligence as capable as HAL but it certainly is going to be a few more years away.
Many of the examples I cited above relate directly to computer technology, I think there is little doubt that AI will come sooner or later. That doesn’t mean I think machines will be conscious. That is a more complex topic that I will address in a later post and a topic mostly irrelevant to this discussion. Intelligence does not require consciousness. Computers can make decisions and take actions without consciousness. They do so today. I think there is probably little doubt Moore’s Law will break down but, by that time, we may very well be close to quantum computing and we will have machines that so outstrip human capabilities that it will make little difference to the overall outcome in movement towards either the dystopia envisioned by Billy Joy or something else. Even though Arthur C. Clarke’s prediction about AI might have been optimistic, his vision of HAL actually looked in the movie more like computer technology from twenty years ago. The real HAL when it arrives may be able to be held in the palm of a hand and it may only need to be that large to allow human interaction with it. As a brain implant, it might be very well be able to slipped behind the ear as part of an outpatient procedure.
When I first began my career in Information Technology, I worked on large mainframe computers. At the time, I thought these systems to be modern-day equivalents of the pyramids of ancient Egypt. They were large, complex, and mostly hidden in cooled rooms away from the eyes of ordinary people. They were presided over by a cognoscenti of programmers, system analysts, and operators who worked tending to daily rituals to keep the computer operating much like the rituals of the ancient priests were thought to bring back the annual Nile floods and cause Orion to appear in the East at the appropriate time. Shortly after beginning this career, the IBM personal computer came out. Year after year brought more storage and more processing speed. The computer came out from the cooled room and moved to the desktop. Then came the Internet and everything got connected. Suddenly computing was everywhere. The achievement of true AI still struggles with the problem of creating a device/program which can actually think of something beyond which it has been programmed to think. For the most part, we still have nothing more than deterministic systems – what goes in determines what comes out in a fairly predictable manner. Nevertheless, I am convinced the time will come when non-deterministic elements get introduced either by deliberate design or through some accidental interaction of the increasing complex elements of the entire system.
So AI will come. Machines will become smart and more human-like. What happens next? Do we control the technology? Does it control us? Do we lose control over it as slowly step by step we turn over control of our lives to machines until we reach a point of no-return?
Let me quote Bill Joy: “I think it is no exaggeration to say we are on the cusp of the further perfection of extreme evil, an evil whose possibility spreads well beyond that which weapons of mass destruction bequeathed to the nation-states, on to a surprising and terrible empowerment of extreme individuals.”(1)
When we look at our evolution as a species, we see humanity has gone through several stages:
Evolution drives the anatomical changes to the upright walking position and put ourselves on the path to larger brain size.
We augment our biological capabilities with technological ones. We create primitive tools for digging, cracking open food, and eventually for weapons.
We develop soft technologies – language, myth, story, leading to religious practices and shamanism. We begin our first efforts at understanding the universe and our place in it.
We create mathematics which drives the further development of hard technologies. We develop the wheel, machines, and agriculture – we begin the control of other life. We begin to develop science.
The soft and the hard technologies begin to feedback upon each other. This is where are at today. Our technological change is at an almost exponential pace. With biotechnology, we directly alter our physical evolution. With computers, we emulate and enhance our own thought processes.
At each stage, the evolution of humanity has been driven by a complex interaction of biology, soft technology (programming?), and physical tools, devices, or hard technology. Intelligence has come from the biological body, manifested itself in the products of thought, such as mathematics and science, and now is on the threshold of externalizing itself in intelligent machines and modifying the biology that gave rise to it in the first place.
The underlying psychic force behind this evolution is the ongoing effort of humanity to deal with the problem of mortality. It is no surprise that Raymond Kurtzweil has written not only The Age of Spiritual Machines in which the hails the coming age of AI but also has written Fantastic Voyage: Live Long Enough to Live Forever in which he hopes through biotechnology and eventual downloading of his mind to a machine to achieve immortality. The project to create intelligent machines and the project for immortality are two sides of the same coin. Both of these projects will likely come to a critical point in next hundred years, what I call Stage Six. This arises from a fundamental conflict between these two goals. Increasing human lifespan with the goal of immortality will occur at the same time intelligent machines increasingly make human beings more useless and redundant.
The extension of life span will happen in three phases which will likely overlap. In the first phase, science will develop general purpose anti-aging medicines and vaccines. Next, aging major organs will be able to be regenerated using stem cells. Finally, major repairs and cancer elimination will be achieved using nanotechnology.
We have known for years that animals placed on caloric restriction diet live longer than animals allowed to eat freely. What’s more, animals on the diet also suffer less from degenerative diseases and maintain their vitality until greater ages. So it is already clear that the rate of aging is not predetermined. Just slowing the aging rate would probably on average extend the human lifespan by twenty to thirty years and slow the onset of diabetes, arthritis, and cardiovascular disease. Probably all that is required to accomplish this is the activation of various genes. The Sir2 gene has already been discovered as one of the prime candidate for this and resveratrol is already known to increase the activity of this gene (although the usage of resveratrol in humans is complicated by difficulties in getting sufficient concentrations into the body to produce the effect). Another anti-aging gene found as a mutation in a village in Ecuador seems to prevent diabetes and cancer, but also only allows the body to grow to about three and half feet tall. Perhaps activating that gene after normal growth has taken place would extend life. Another gene, the KL gene, encodes for the Klotho enzyme and Klotho deficient mice show accelerated aging so a manipulation of this gene might slow aging. Within ten to twenty years general purpose aging medications or vaccines will become available that will slow the rate of aging. These will become increasingly potent over the following decades.
Stem cells are biological cells that can divide and differentiate into diverse specialized cell types. The division of embryonic stem cells, of course, are what create the body to start with in the womb, but stem cells continue to exist in the body to repair and regenerate tissue. Once this process is understood and able to be enhanced and controlled, we will be able to regenerate damaged organs and tissues or perhaps even perform an entire body regeneration. We already have cardiac patients who have undergone stem cell therapy that replaced dead heart tissue with live tissue. Regeneration of heart tissue will be common within five years. The regeneration of kidneys, liver, bladder, lungs, and other internal organs will be possible in twenty years, and the regeneration of spinal cord and brain tissue will be possible in forty years. The regeneration of whole limbs may be possible in fifty years.
Medical nanotechnology will eventually create nanobots and nanobot-organism hybrids that will be able to do work at a cellular level. This technology will be able to target specific cells, such as cancer cells or other damaged tissue, for destruction or repair. In the more near term, nanobots will likely be used in diagnostic agents. Instead of invasive surgeries and biopsies, a simple injection of nanobots might be able to diagnose most conditions. The targeted drug delivery for cancer will be possible in ten years, but more sophisticated repair operations may be fifty years away. Eventually almost all damage to the body from injury that does not result in almost immediate death will be able to be repaired .
All of these things by themselves do not allow us to achieve immortality. It only gets us perhaps to an extended and relatively long life. The final stage to immortality as envisioned by Kurtzweil is the complete encapsulation of human consciousness in a machine that is self-repairing and self-regenerating. For reasons I will discuss more in Part II of this post, I do not think this is possible with anything we might envision as machines today, but it may be possible with a new (or perhaps really old) sort of machine. However, for the next hundred years, whether it is possible or not is probably not pertinent to the sort of societal dynamics we can expect to occur.
Since industrialization numerous jobs have gone out of existence as new technology came into being. Ditch diggers have been replaced by backhoes. Elevator operators by automatic elevators. Secretaries by personal computers. Other jobs have had dramatic increases in productivity. A farmer at the turn of the century might have been only able to farm fifty acres whereas today he might manage several thousand acres. It might have required a team of mechanics a few months to build a single car in a garage but the same team might have output a car per day per person on the early assembly line. With today’s robotic assembly lines, the output is even greater. With technology we gain increasing improvement in productivity through the elimination or transformation of jobs.
Although we have experienced momentary downturns in the economy which has resulted in large numbers of people out of work, for the most part, new jobs have appeared to take the place of the ones that have become obsolescent. This has occurred largely because new industries and new types of jobs have come into existence. If all we mostly did was farm as we did several hundred years ago, then now almost every one would be out of work because we now only need a small fraction of the farmers we needed then to produce much more food than we did at that time. So as long as when new industries appeared, new jobs have appeared then we have mostly maintained a balance.
I do not think likely that some dramatic reduction in the creation of new industries will happen. What is more likely is at some point in the next hundred years new industries will stop creating significant numbers of jobs, particularly jobs that can be performed by average people of average intelligence.
We will see this first in industries involving the creation of physical goods. We already see tendencies toward this in the automobile industry where more and more mechanization means few and few workers. What happens when the assembly is completely robotic and the workers are nothing more than repairmen and caretakers. But that is old industry, you say. Okay. Take this more recent example. Foxconn factories in Taipei that make the iPad are bringing in robots. As Martin Ford write on his blog: “Automation is not just about increasing efficiency. There’s some evidence to suggest that workers are simply being driven beyond their limits. As production speeds continue to increase, there has to come a point where the only option is to get the humans out of the loop. In many industries, automation may penetrate more rapidly than we expect simply because a threshold is reached where people can no longer keep up.”(2) We are probably not far from the day when a new industries will be created with completely robotic workers. The investors will raise money. Robots will build the plant and install the manufacturing equipment with a few human supervisors. Somebody, perhaps the CEO, will throw the switch and product will come out of the back of the plant. But what about the people who designed and built the equipment that went into the plant? In fact, there is no reason why they too could not be made obsolescent with the exception of a few highly skilled individuals.
The core point is there is almost nothing in the creation of material goods that could not be almost completely robotized. The next step in this process is the complete fabrication of machines themselves from design with few or maybe no workers in the intermediary process. We already see the beginnings of this in three-dimensional printing where items are fabricated much like words or images are printed by laying down successive layers of material. We might eventually have mufti-purpose factories that might be to produce any of a number of goods on demand and might be able to be converted from one output to a new one with a change in the input to the printing process.
After material goods will come intellectual goods. Computers already do most of the drudgery of computation. They are increasingly relied on for image recognition and complex data analysis. For the most part, all of this is done with programs still created by humans. Within a hundred years, computers will not just be able to run programs but able to create them too perhaps with just minimal direction from human beings. We might be to ask our computers (possibly something the size of a ring on our finger) not just to play music but to create it for us on the fly or to design for us a next generation of some super iPad-like device of the future.
What about service jobs? Bank tellers have been progressively replaced by ATMs. Cooks by mass-produced fast foods. Doctors might be reduced to doing little more than supervisory duties for armies of robot surgeons and diagnosticians. Lawyers, alas, may be with us forever.
Of course, niches will remain in the service industries. Particularly talented writers, artists, chefs, and athletes may be paid premiums but those jobs are few. Custom produced items – furniture, cars, and homes – will still have a market. Beyond that, what is left for the average worker?
More importantly, what is the basis for a rational economy under these circumstances?
Imagine this. Medical advances dramatically increase the possible human lifespan. Robotic technologies increase the amount of goods able to be produced without human workers. The free market system is based on each worker being able to sell in the market place his services and abilities, but if the production of high value products does not require humans than the worker has nothing to offer. We are left with large number of workers competing for the niche service jobs, driving down those wages, while the wealth is maintained by owners of the machines. Who will buy the products of the machines? And will the workers be able to afford the medical advances to extend their lifespan? The situation is perhaps even more dire outside the industrial world. Just as jobs in the industrialized world have been shipped off to the Third World, the workers there may within a few decades see their jobs disappear to robotic factories leaving those countries to return to subsistence.
What alternatives are there for a world like this?
Let’s look at some muddling through types of scenarios with little significant change except extension of the current direction. We might end with a wealthy elite owning the machines, trading goods among themselves, and living for 150 years while the remainder of humanity live short lives of poverty, perhaps on the government dole to prevent civil unrest. Or, perhaps material goods become so plentiful and so cheap to produce that no one can become wealthy owning them. The machines may become like public lands, held by all, and for the benefit of all.
I don’t see either of those futures happening like that exactly. That would be a mistake similar to that of Malthus or Erlich – the mistake of simply extrapolating trends. But I do see these themes playing out sharply in the next decades. There will need to be some kind of fundamental reorganization of our economy and a reevaluation of the basis of economy when people begin to live routinely for over a hundred years and most goods and many services are provided free of human labor.
What about the Bill Joy dystopia? The machines become so intelligent and we so dependent on them that eventually the machines decide they no longer need us – the Skynet Terminator scenario. The machines decide to destroy humans. It could be, in fact, a supremely rational and intelligent choice. Humans have mostly wrecked havoc upon the Earth’s ecosystem and we undoubtedly consume far more resources than we should for long-term stability of Earth’s systems. Machines seeing and understanding our insatiable demand for goods and energy could make a cold calculation that we should go, perhaps leaving a few of us alive for their amusement.
Perhaps there is another way. We could abandon the projects of immortality and abandon the machine in one way or another. Henry Miller in Sunday After the War:
“The cultural era is past. The new civilization, which may take centuries or a few thousand years to usher in, will not be another civilization. It will be the open stretch of realization which all the past civilizations have pointed to. The city, which was the birthplace of civilization, such as we know it to be, will exist no more. There will be nuclei, of course, but they will be mobile and fluid. The peoples of the earth will no longer be shut off from one another within states but will flow freely over the surface of the earth and intermingle. There will be no fixed constellations of human aggregates. Governments will give way to management, using the word in a broad sense. The politician will become as superannuated as the dodo bird. The machine will never be dominated, as some imagine; it will be scrapped, eventually, but not before men have understood the nature of the mystery which binds them to their creation. The worship, investigation and subjugation of the machine will give way to the lure of all that is truly occult. This problem is bound up with the larger one of power—and of possession. Man will be forced to realize that power must be kept open, fluid and free. His aim will be not to possess power but to radiate it.”.
1- Why the future doesn’t need us by Bill Joy Wired Magazine http://www.wired.com/wired/archive/8.04/joy.html