The Singularity, defined by some scientists as “a future time when societal, scientific, and economic change is so fast we cannot even imagine what will happen from our present perspective,” (1) seems far in the future; past our lifetimes at least. But data from all aspects of life shows that this might not be the case. In fact, Ray Kurzweil predicts in his book The Singularity is Near that this event will happen by 2045. Not surprisingly, this radically different future will have many problems, most of which aren’t discussed by Kurzweil or his followers. There will be external dangers—such as out of control self-replicating nanobots, which I won’t discuss in this paper. There will also be dangers within the human race such as almost a complete destruction of human interaction and motivation. I believe that Kurzweil’s predictions will come true, but that past the inclusion of non-biological intelligence into human brains we should suppress this future technology.
The basis for Kurzweil’s predictions is the law of accelerating returns, which says that technology is improving exponentially over time (p.7). An important example of this is Moore’s law, which states that the number of transistors that can be put on a computer chip will double every two years. While Moore only intended this to be a short-term explosion in computer technology lasting until 1975 or so, the trend continued and still is continuing to this day (p. 111-112). In fact, Paolo Gargini, chairman of the International Technology Roadmap for Semiconductors, said in 2005 that Moore’s law should continue to be correct until 2020 (2). So this law turned out to be even more powerful than its founder predicted. The law of accelerating returns occurs in many other areas with some relation to technology such as DNA sequencing and manufacturing speed (p. 74). Critics of this law believe that it is a self-fulfilling prophecy. Businesses predict that a certain area of technology will improve exponentially and then focus their resources on making it happen. However, exponential improvement has occurred in so many different areas of technology that this cannot be the case.
While there is an enormous accumulation of evidence in support of exponential improvement in technology, people continue to think that technology will improve linearly. People naturally expect most things, including improvement in technology, to happen as it traditionally has. This linear view of the future has resulted in many terribly conservative predictions. Even Popular Mechanics, a magazine that has a more optimistic view of the future than most people, said in 1949, “[While] a calculator on the ENIAC is equipped with 18,000 vacuum tubes and weighs 30 tons, computers in the future may have only 1,000 vacuum tubes and perhaps weigh 1.5 tons" (p. 56). The difference between an exponential and a linear view of the future can be seen with internet usage over the past 30 years. There appears to be a small linear improvement for about 15 years followed by an explosion of use over the next 15 years (3). When the data is shown on a logarithmic graph, however, the true exponential nature of the data is obvious. Thus, people making predictions about the future 15 years ago using a linear view would have grossly underestimated future improvement. All sorts of other technologies and parts of life have similar exponential growth.
The key ingredient of the Singularity, computing power, obeys the law of accelerating returns as well. The number of calculations per second that can be done with $1000 of technology has increased exponentially (p. 67). There is an argument that there are physical limits to technology and so exponential growth will only occur temporarily. However, physical limits have been reached several times already. The method of computing changed from electromechanical to vacuum tubes to integrated circuits as each previous technology approached its limit. So while an individual piece of technology can’t keep being improved over long periods of time, new inventions with more potential will create exponential growth. People using this argument assume that no new inventions will occur, and hence would have expected in the 1950s that vacuum tubes would forever be the basis for computers.
Based on the exponential growth of computing power, personal computers will be as intelligent as human brains by 2025 and superior to human brains soon after. A brain is estimated to be capable of somewhere between 10^14 and 10^17 calculations per second. When Kurzweil published his book in 2005, the best supercomputer in the world Blue Gene was being built to do about 5x10^14 calculations per second, already in the range of possible brain power (4). Less than four years later, the Roadrunner supercomputer can do over quadruple that (4). So supercomputers will reach even the most conservative estimates for human brain power very soon. Thus, having personal computers with this much power by 2025 seems plausible. Of course, increases in computing power won’t stop at 2025 but will continue to accelerate towards the Singularity.
While enormous increases in computing power don’t seem physically possible, there are a variety of unconventional methods that are starting to be used for computing. For example, a DNA computer with speed rivaling that of supercomputers has already been made. It uses a single DNA molecule to solve simple problems and give a yes or no answer. While it seems ridiculous that this is possible, DNA after all can store large amounts of information and enzymes can manipulate this information accurately. DNA computers are smaller and use less energy than any previous computing device so they have enormous potential. A method with even greater potential uses the spin of electrons found in all matter. While Kurzweil predicted in 2005 that this method will be used in the future, in November of 2007 researchers at the Delft University of Technology were able to use an electric field to control the position of a single electron(5). Whether the electron is “up” or “down” is akin to a 0 versus a 1 in a traditional computer. Using enormous numbers of electrons allows the computer to test every possible answer to a problem at once. These quantum computers, however, use virtually no energy and can be built from anything that contains electrons, so any piece of matter can be used.
These interesting predictions for future computing seem non-threatening and theoretical until their potential to profoundly change humans and the universe is considered. A quantum computer built from a one kilogram rock, for instance, has the potential to do 10^42 calculations per second if its electrons can be controlled (p. 131). That is the equivalent of at least 10^25 human brains. Furthermore, a rock doesn’t need any energy to “run”; the electrons move on their own. This means that every kilogram of matter on earth can be built to have more intelligence than the combined intelligence of every person that has ever lived on earth. People can use a kilogram of this non-biological intelligence along with their roughly kilogram of biologic intelligence to become enormously intelligent. Of course, a single brain alongside the equivalent of 10^25 brains will be meaningless, so biological brains will pretty much become obsolete. I think this is ok, however. People will still have differences in every way that they have differences today. They will still be able to learn and work; albeit at an extremely fast pace. Contrary to the genetic engineering we discussed in class, however, these upgrades will be available to everyone regardless of income or age. The cost of non-biological intelligence decreases exponentially, so it will be affordable to everyone by the Singularity. This intelligence doesn’t require a permanent alteration of genes or the brain like genetic engineering does, so people can upgrade their intelligence throughout their life rather than being born with a set amount of intelligence that will soon become obsolete.
The Singularity will enable humans to go much farther than adding a small portion of non-biological intelligence to every human. However, I believe that humans should stop here. Two potential uses of the Singularity, brain uploading and virtual reality, show why. When humans are able to understand the brain enough to create non-biological intelligence, they will know enough to upload a brain onto a computer. It will work in exactly the same way as the original brain, so if it would be considered human if it had a body attached to it. This would allow humans to be immortal; they could simply change bodies every once in a while and keep their brain. Also, people could be in multiple places at once using multiple copies of themselves(6). While this might have some benefits, the ethical problems and possible abuses of this technology make it too dangerous to use. Another technology with the potential to transform our lives is virtual reality (VR). VR that is 100% real in our minds will be possible using nanobots in our blood vessels. The nanobots will simply be tiny parts of our non-biological intelligence. They will be able to make a neuron (nerve cell) fire or suppress a neuron from firing whenever they are programmed to do so (p. 313). Thus, they can suppress every signal that comes from our senses to make us sense nothing naturally. Then they will create sensations of their own to simulate any conceivable action and environment. We can get pleasure just as humans did in Lilith’s Brood or be in any imaginable environment, including physically impossible environments (p.314). Reality will seem boring and painful compared to this, so humans would spend every possible second of their time in VR. Therefore, human interaction would be brought to a complete end. We would also have no motivation to take care of our bodies, since we can escape them whenever we want. The real world would be almost completely replaced by our fantasies.
While some extremely influential technology resulting from the Singularity should be suppressed, humanity will still be profoundly changed for the better in the near future. The Law of Accelerating Returns will make computing power decrease exponentially until it can be done extremely cheaply and efficiently using quantum computers. These computers will have enormous intelligence that can be incorporated into humans. Simply adding a one kilogram non-biological “brain” with a traditional brain will enable humans to have near-perfect memories and to learn and work at speeds incomprehensible to us now. These predictions seem very unlikely to occur anytime soon, but that is because our linear view of the future doesn’t well approximate the true nature of technological improvement.
Sources (not cited/in bibliography correctly yet)
1. http://www.kurzweilai.net/meme/frame.html?main=memelist.html?m=1%23696
2. http://news.cnet.com/New-life-for-Moores-Law/2009-1006_3-5672485.html
3. https://www.isc.org/solutions/survey/history
4. http://www.top500.org/
5. http://www.sciencedaily.com/releases/2007/11/071101144942.htm
6. http://www.transhumanism.org/index.php/th/more/339/
7. The Singularity is Near
No comments:
Post a Comment