The Singularity Is Near Review

The Singularity is Near By: Ray Kurzweil Chapter 1:The Six Epochs The Singularity is a future period during which the pace of technological change is so rapid that all aspects of human life will be irreversibly changed. The advent of self-improving artificial intelligence will mark the beginning of the Singularity. Kurzweil believes this will happen during the first half of the 21st century. At the heart of Kurzweil’s beliefs is the “Law of Accelerating returns,” which states that technology is growing at an exponential rate. The human brain has the advantage of being massively parallel in design, but is handicapped by very slow speed. The Singularity will mark the end of human history as we know it. A key occurrence will be the merging of human and machine intelligence. There will be no distinction between physical and virtual reality. Kurzweil believes that common predictions of the future fall short because they are based on the belief that technology advances linearly as opposed to exponentially. Kurzweil claims that the 21st century will see a thousand times more advancement than occurred in the 20th. Kurzweil cites the examples of mapping the human genome and building the modern Internet: at the start of both processes, experts in both fields believed the completion of both tasks would take extremely long periods of time because they based their estimates on past trends and failed to take into account the fact that developing technologies would allow the speed of advancement to increase exponentially. Biology and technology have six common Epochs. Epoch 1: Physics and Chemistry -The birth of the Universe and the creation of matter and energy -The Universe is governed by constant forces Epoch 2: Biology and DNA DNA-based life began. Epoch 3: Brains Organisms with sensory organs and brains developed. Epoch 4: Technology -Humans create technology -Kurzweil selectively chooses certain natural and human milestones for his logarithmic plot of progress as proof of his “accelerating returns” theory, but he shores up his claim by offering similar lists created by other academics independently. Epoch 5: The Merger of Human Technology with Human Intelligence Fusion of man and machine to produce a newer, better Posthuman species. Epoch 6: The Universe Wakes Up Posthumanity spreads and makes all the matter in the Universe “intelligent.” The Singularity involves the following principles: -The rate of technological innovation is continually accelerating -The power (speed, capacity, price-performance) of information technologies is increasing supraexponentially.

1

-Human brain scanning techniques, which are dependent on computing power and technological innovation, are also improving at an exponential rate, and a full understanding of how the brain works will exist in the 2020’s. -Computers will be able to “emulate” human brains by this time as well, and they will be able to pass the Turing Test. At this stage, computer intelligence will have achieved human levels of creativity and mental flexibility, but with vastly greater processing speeds and abilities of memory, making computer intelligence superior. -Computers are also able to share information almost instantaneously while human teaching is extremely laborious and slow. -Machines will have access to the Internet and will be able to understand and master all human knowledge. -Machines will have the ability to join into one entity at will, to separate, or to split into new entities. Kurzweil believes that this is essentially the same as “falling in love.” -Access to their own source codes will allow machines to redesign themselves in a way basically similar to human genetic engineering, but with much greater power and speed. -Human beings will re-engineer their bodies to be superior. -The unmodified, biological human brain will become obsolete. -Computers of the future will use three-dimensional microchips that are thousands of times faster than current computers. -Machine intelligence will improve itself in a feedback cycle that ordinary humans will not be able to follow. This cycle will increase in speed exponentially. -Nanomachines will improve human physical and mental abilities and reverse the aging process. -Nanomachines will reverse environmental damage and allow for true virtual reality. -Machines will eventually learn to experience human emotions, though their emotions might evolve and become different. -Human/machine intelligence will eventually become so great that it begins utilizing most of the matter and energy in our Universe for information processing purposes. This will progress in exponential fashion. -Eventually, the Universe will become saturated with Posthuman intelligence, and that intelligence will radiate outwards from the Earth at the maximum possible speed. Kurzweil believes that Posthumanity will still represent the values of humans and that Posthumans will respect and protect unmodified humans. [Why is he so sure?] Chapter 2: A Theory of Technology Evolution One concept of complexity is the minimum amount of information necessary to represent a process. Compression of information can be done by recognizing patterns and redundancies. Complexity is the minimum amount of nonrandom, meaningful, unpredictable information required to characterize a system or process. Paradigm shifts in biology and technology all involve increasing amounts of complexity. Order is information that fits a purpose. The measure of order is the measure of how well the information fits its purpose. A more evolved organism has a more ordered design. Greater order does not necessarily entail greater complexity. Sometimes, the simpler solution proves superior.

2

Each stage of evolution and technological innovation increases order and sets the stage for the next. Principles of the Law of Accelerating Returns -Evolution applies positive feedback, with each improvement building upon the last and enabling the next. -The returns of technological evolution increase exponentially. -A specific paradigm generates exponential growth until its potential is exhausted. The Life Cycle of a paradigm -Sigmoid shaped curve -Slow growth -Rapid growth -Leveling off -Reflective of organismal population trends when introduced into a new environment. Evolutionary theory of punctuated equilibrium describes evolution as consisting of periods of short, intense change separated by periods of tranquility. The creation of DNA accelerated the rate at which evolution could take place. [Kurzweil compares the time it took for people to adopt traditional telephones to the time required for adoption of cell phones and the Internet to show proof that society’s willingness to adopt new technologies is increasing. But is this a sufficient study, and does it support his idea that people will accept more radical technologies?] Kurzweil demonstrates that a number of factors relating to microprocessor cost and performance are changing in an exponential fashion. Moore’s Law does not apply just to integrated circuits—it applied to the price-performance of the previous paradigms of computers (electromechanical, relays, vacuum tubes, and discrete transistors). “Moore's Law is the empirical observation that the transistor density of integrated circuits, with respect to minimum component cost, doubles every 24 months[1]. It is attributed to Gordon E. Moore[2], a co-founder of Intel.” “A similar law has held for hard disk storage cost per unit of information. The rate of progression in disk storage over the past decades has actually sped up more than once, corresponding to the utilization of error correcting codes, the magnetoresistive effect and the giant magnetoresistive effect. The current rate of increase in hard drive capacity is roughly similar to the rate of increase in transistor count. However, recent trends show that this rate is dropping, and has not been met for the last three years.” “The exponential increase in frequency of operation as the only method of increasing computation speed is misleading. What matters is the exponential increase in useful work (or instructions) executed per unit time. In fact, newer processors are actually being made at lower clock speeds, with focus on larger caches and multiple computing cores. The reason for this is that higher clock speeds correspond to exponential increases in temperature, such that it becomes almost impossible to produce a CPU that runs reliably at speeds higher than 4.3 GHz or so.” “Not all aspects of computing technology develop in capacities and speed according to Moore's Law. Random Access Memory (RAM) speeds and hard drive seek times improve at best a few

3

percentages per year. Since the capacity of RAM and hard drives is increasing much faster than is their access speed, intelligent use of their capacity becomes more and more important. It now makes sense in many cases to trade space for time, such as by precomputing indexes and storing them in ways that facilitate rapid access, at the cost of using more disk and memory space: space is getting cheaper relative to time. Another, sometimes misunderstood, point is that exponentially improved hardware does not necessarily imply exponentially improved software to go with it. The productivity of software developers most assuredly does not increase exponentially with the improvement in hardware, but by most measures has increased only slowly and fitfully over the decades. Software tends to get larger and more complicated over time, and Wirth's law even states that "Software gets slower faster than hardware gets faster". Moreover, there is popular misconception that the clock speed of a processor determines its speed, also known as the Megahertz Myth. This actually also depends on the number of instructions per tick which can be executed (as well as the complexity of each instruction, see MIPS, RISC and CISC), and so the clock speed can only be used for comparison between two identical circuits. Of course, other factors must be taken into consideration such as the bus size and speed of the peripherals. Therefore, most popular evaluations of "computer speed" are inherently biased, without an understanding of the underlying technology. This was especially true during the Pentium era when popular manufacturers played with public perceptions of speed, focusing on advertising the clock rate of new products. [8] It is also important to note that transistor density in multi-core CPUs does not necessarily reflect a similar increase in practical computing power, due to the unparallelized nature of most applications.” Kurzweil believes that three-dimensional molecular computers will form the sixth paradigm. Kurzweil’s supraexponential growth of computing power means that supercomputers will match human brain capabilities by 2010, and personal computers will be just as advanced by 2020. The human brain’s most advanced feature is its three dimensional, parallel design. Paradigm shift = innovation GNR = Genetics, nanotechnology, robotics The coming GNR Age will result from the advance of multiple technologies. Fredkin and others have postulated that all matter and energy in the Universe is really the same thing, but in different forms. Simple programs can produce systems of great complexity. The blueprints for the human brain are dwarfed by the complexity of the finished product. Kurzweil believes that the key to creating A.I. is the reverse engineering of the human brain. Kurzweil believes that modern economists focus too much on capital costs of physical infrastructure and fail to pay enough attention to the impact of information technologies, patents, etc. Information will eventually dominate the economy. Kurzweil believes that technology advances make widely accepted projections of the future useless. Actual increases in longevity will far surpass what is currently expected, people won’t

4

retire at 65 when medical technology can greatly improve their quality of life, and economic growth resulting from information technologies will be vastly greater than we expect. Kurzweil shows that national and per capita GDP growth has been exponentially increasing in real terms. Productivity figures fail to take into account the increasing personalization and hence worth of consumer goods (i.e.—cheap clothing tailored for individuals). Products are improving qualitatively. Private manufacturing output is also increasing exponentially. Kurzweil is not concerned about deflation because it is overridden by the insatiable consumerism of human beings. Kurzweil still believes that “computers will disappear as distinct physical objects” by 2010, and that mini computers will be built into eyeglasses, clothing, and furniture, and that full-immersion virtual reality will exist. [I’d bet my last dollar that the most powerful and useful computers will still be in the familiar form of laptops and desktops] The advent of cheap, high-quality VR will allow workplaces to be totally decentralized, wreaking havoc on the real estate industry. [Again, this may be another misanalysis of the market —just because a technology exists doesn’t mean that average people will use it] Kurzweil’s trackings of exponential economic growth show that the normal boom-bust cycles have no significant effect on overall economic growth. [Correct] Chapter 3: Achieving the Computational Capacity of the Human Brain Gordon Moore understood the importance of integrated circuits and formulated his famous law (“Moore’s Law”) in 1965. Price performance of computer chips has increased exponentially due to the shrinking of transistors. Three-dimensional molecular computing will reduce the cost and size of microchips. Prototypes have already been built. Kurzweil believes that carbon nanotubules are the materials of choice for 3D computer semiconductors. Recent advances have made nanotube synthesis easier and more precise. [But still a long way from perfect or even economical] Carbon nanotubules are grown from a substrate. Advances have also been made in using individual atoms and small molecules for computation. Self-assembly of nanoscale components would be much cheaper, easier, and space-efficient than current methods, and defective creations could be disposed of immediately. Again, prototype nanostructure assemblers have been recently created. The idea of building self-replicating and self-organizing computers is inspired from biology. Prions are self-replicating proteins. DNA computing is highly useful for limited applications (SIMD—Single Instruction Multiple Data). Computers can also utilize the two different possible “spin” properties of electrons—“up” or “down”—for memory storage and logic. This can be done at room temperature. Lasers can also be used to perform logic functions. Quantum computers also are SIMD, though are of a more radical design than DNA based computers. The design is based on manipulation of “qubits”—fundamental ambiguities inherent in quantum mechanics such as the spin state of individual electrons. Each atom acts as a 1 or 0.

5

The Quantum Computer has the advantage of being able to solve problems at extremely fast speeds. However, as with all SIMD computers, its applications are narrow. There are a number of major design challenges that need to be resolved to make a useful Quantum Computer. By Kurzweil’s calculations of the exponential increases in computing, even misestimates of the computational capacity of the human brain by factors of a thousand or more would delay the Singularity by only a few years. Based on studies of nerve tissue in the retina and extrapolation to the rest of the brain, Hans Moravec estimates the computational capacity of the human brain to be 1014 instructions per second. Lloyd Watts et al. achieved the same estimate by analyzing human hearing. Researchers at the University of Texas estimate 1015. Kurzweil decides to be conservative and use 1016 instructions per second as his estimate of human brain processing. [Cerebellum functions are included in this?] Supercomputers will be able to process information at the same speed as the human brain by early next decade. Kurzweil believes that the arrival of human-level computer processing speeds could be achieved by: -Building computers with Application Specific Integrated Circuits (ASIC) to perform rote, repetitive functions at much higher speeds and efficiencies than General Purpose Processors. With the two types of microprocessor, the computer would be much faster. The cerebellum’s functions are identical to ASICs. -Fully utilizing idle computational devices via the Internet. Supposedly, 99% of potential is wasted. -Use of “analog” processors for certain, specialized functions. Kurzweil estimates the human memory to have 1013 bits. In 2018, 1013 bits of computer memory will cost $1000. [Most of the human brain’s energies are devoted to maintenance of homeostasis, not to problemsolving or constructive thought. Then why does Kurzweil include all of the human brain’s potential in calculations of the required level of computer performance necessary to match human cognitive abilities? ] The processing of information (and hence the power of computers) has limits imposed by the laws of physics. Energy requirements are a major limiting factor. The reduction in microprocessor energy demands has been counterbalanced by the squeezing of more transistors onto ever-smaller microchips. The result is that proper heat dissipation is still a major concern in computer design. Parallel processors could operate at cooler temperatures than really powerful single processors. Intel has recognized this and is creating multiple processors on single chips. Parallel processing will only do so much to control overheating. Much heat results from the computer discarding unneeded electronic information (by the conservation of energy law)— specifically, the intermediate steps of a calculation to save memory space. In Reversible Computing, the intermediate steps are kept in memory, causing no heat release and allowing the algorithm to be run in reverse since no information is lost. Running the program normally and then backwards will supposedly require no energy and generate no heat.

6

[Rolf Landauer did pioneering research on logic gate reversibility and the effect of the second law of thermodynamics on computing in the 1960’s. The topic of reversible computing is not Kurzweil’s invention and is under serious study.] Information is stored both in discrete objects (as small as subatomic particles) and in the connections between objects. The connections provide the exponentially large storage. Ed Fredkin created a theoretical computer utilizing reversible logic gates that was 99% as efficient as irreversible logic gate computers. Supposedly, this computer does not need to dissipate any energy. The output of the calculated information, however, requires energy for every bit transmitted. However, this is of little importance since the algorithms themselves generate most of the heat. Even computers make mistakes and have error rates. Correction of errors is nonreversible and thus requires energy, so in reality, even the most advanced computer with no output device will still generate some heat and require some energy. Hyperfast, reversible computers will eventually start to encounter problems with heat dissipation due to this. Kurzweil believes that nanoscale fuel cells spread out across the computer processors will provide energy. The potential amount of computation is positively correlated with available energy. Energy is associated with each atom and subatomic particle. More atoms mean more computation. Atomic energy increases with increased movement. The theoretical limit of an object’s ability to perform computation is the total energy divided by Planck’s constant (6.6 x 10-34 J/s). A kilogram of “matter” (ambiguous) therefore has the theoretical potential to perform 5 x 1050 operations per second, compared to 1016 operations for a human brain. But this conversion of matter to energy might produce a thermonuclear explosion. A more stable computer operating at normal temperatures could perform 1042 operations per second, weaker but still vastly more powerful than every human mind that ever lived combined. Kurzweil sets the date for the Singularity—the profound, disruptive transformation in human capability—as 2045. In that year, if the predicted trends hold, $1000 will buy a computer with a billion times more intelligence than every human on the planet combined. [But what is so special about that? Why not a trillion times more intelligent? Kurzweil does not explain. Furthermore, the creation of Strong A.I. hinges more on software development than it does on improvements in computing speeds. Where does Kurzweil prove that the rate of A.I. software progress is increasing fast enough allow for A.I. by the 2020’s and for the Singularity, which itself is simply the moment when A.I. becomes so advanced that it enters into a self-improving cycle that humans cannot affect, in 2045?] By that time, computers will be much smaller and will be integrated everywhere, including inside some human brains. Engines of Creation is described as the seminal book about nanotechnology. Kurzweil suggests that computing at the subatomic level (pico- and femtotechnology) could make computers even faster than 5 x 1050 operations per second, though even he isn’t sure if it is possible. With self-replicating nanotechnology, all matter could be made into computing devices, but the expansion of this “massive computer” would be limited by the speed of light. [Here, he might be going off the deep end]

7

Analysis of changes in the alpha (fine-structure) constant reveals that the speed of light (which is inversely proportional to the constant) has also changed. Increasing or surpassing the speed of light would not only allow true exploration of space, but it would allow for faster computational speeds. Todd Brun has theorized that time travel for the purposes of calculations is possible, without the danger of a “grandfather paradox.” Chapter 4: Achieving the Software of Human Intelligence Our ability to understand how the brain works is growing exponentially. The first artificial intelligence will result from the creation of a computerized version of the human brain. Brain scanning techniques are improving rapidly. Nanomachines will one day allow for perfect scans of the human brain from inside the bloodstream. Contrary to popular belief, the human brain’s complexity of function is not beyond our understanding. Kurzweil admits that raw computer processing power and speed will not produce sentience. Effective organization of the computer by emulating the human brain is the key to artificial intelligence. The first artificially intelligent computer will immediately render the human mind obsolete. Computers have the advantage of being able to learn instantaneously by downloading data. For humans, the task of learning is far less efficient. The structural features of the human brain exhibit a high degree of repetition. Only a relatively small part of the human genome codes for the human brain. The brain achieves its complexity through interaction with the environment. Models of how parts of the human brain function have been developed. Lloyd Watts’ analysis of the auditory cortex has resulted in the creation of a computer program that can localize sounds just as human hearing does. Kurzweil does not believe in scanning and electronically copying every neuron from a sample human brain. Rather, he wants to understand the coding and allow a self-organizing computer to create its own electronic brain using the blueprints. The human brain is composed of hundreds of specialized regions. Neurotransmitter release is considered “analog” communication. The human brain is massively parallel, with trillions of computations occurring simultaneously. This imparts its unrivaled abilities of pattern recognition. Analog computers have the advantage of being much more efficient than digital computers, but digital computers can be quickly reprogrammed while analog computers cannot be reprogrammed at all. Key differences between modern computers and human brains: -The brain works at very slow speeds. -The brain is massively parallel. -The brain combines analog (NT release) and digital (neuron firing?) functioning. -The brain rewires itself. -Most details concerning brain structure at the cellular level are random and don’t affect consciousness.

8

-The brain uses “emergent properties” (poorly explained). -The brain is imperfect. -The mind is capable of containing concepts that are contradictory. -The brain uses a system of “evolution” to rewire itself. -The brain is self-organizing. -The brain is holographic (poorly explained). -The neurons within the brain have multiple paths of synaptic connection so that the loss of any one brain cell will not affect the function of the whole. This is similar to the structure of the Internet. -There are hundreds of different regions in the human brain that perform discrete functions. -The design of each individual brain region is surprisingly simple and repetitious. The structure and chemical functioning of an individual neuron contained within is more complex. Thus far, our attempts at understanding the human brain have been crude and hampered by inadequate technology. The resolution of noninvasive brain scanning devices is doubling every twelve months— increasing exponentially as with other technologies. fMRI measures blood oxygen levels, has a resolution of 1-3 mm and has a low temporal resolution of a few seconds [wow, how poor]. MEG is a different technique that measures magnetic fields from outside the skull. It has a very low spatial resolution of 1 cm, but a high temporal resolution of 1 msec. Fritz Sommer is investigating the feasibility of combining the two techniques to create a superior brain scanning device. rCBF measures regional cerebral bloodflow, which is positively correlated with synaptic activity. TMS (transcranial magnetic stimulation) involves affecting different areas of the brain with strong electromagnets. This can be inhibitory or excitatory. Treatments have been known to often cause spiritual experiences and the temporary enhancement of certain mental abilities (such as artistic skills). Destructive brain scanning yields dramatically better results. Such scans on animals produce resolutions of less than 200 nm, which is nearing the point for reverse engineering. In this process, the brain is sliced into very thin segments for each scan. A number of new brain imaging techniques are in development. Internal brainscans using nanomachines will provide the most accurate pictures of brain structure and activity. Nanobots would enter the capillaries of the brain and transmit information wirelessly to computers. Kurzweil believes that brainscans will be one of the first principal uses nanomachines are tasked with, and that they will be able to perform this function in the 2020’s. Kurzweil’s suggestions on how to make nanomachines able to cross the blood-brain-barrier (dense capillary walls found in the brain): -Make the nanobots small enough to pass through the walls of the BBB. This would mean making them with a diameter less than 100 carbon atoms, severely limiting their functionality. -Make the nanobots large, but with thin, protruding arms that could be put through the fenestrations and used to scan the nerve cells. Workable since all cells are located next to capillaries. -Keep nanobots in the capillaries and equip them with miniaturized brain scanning devices (something that complex made so small?).

9

-Have the nanobots burrow through the BBB and then immediately repair the damage before scanning. -Design the nanobots with the ability to mimic organic molecules (such as glucose) that can freely pass through the BBB. -Direct injection of nanobots into brain tissue (wouldn’t that damage nerve cells?). Rob Freitas has envisioned techniques for monitoring sensory signals in the brain for reverse engineering the input systems. All techniques involve putting nanobots in contact with nerves of the different sensory and motor organs to measure activity. There is a difference between scanning the brain for the purposes of understanding basic function for the creation of artificial intelligence and scanning the brain to create a computer duplicate of a person’s mind. The former requires less exact knowledge of the individual’s brain anatomy and physiology. The first nanomachines will likely be used for medical purposes. Kurzweil “conservatively” estimates that, given such an emphasis on the medical uses, brain-scanning nanobots will be invented during the 2020’s, allowing for full scans of brain activity. Kurzweil believes that the past claims that human thought processes are incomprehensible are wrong and based on the poor performance of inadequate brain scanning technology. Joseph LeDoux: All human memories, skills, and personality traits (in essence, the consciousness) exist as synapses and combinations of connections between neurons. Anthony J. Bell: Subcellular features that affect neuron structure, performance, and maturation are also responsible for nuances of brain function (does not seem to inherently conflict with LeDoux). Synapses and dendrites are in a constant state of change (plasticity). Dendrites grow and die on a daily basis: only a small few remain permanently. Mathematical models of their growth cycles have been attempted. (It should logically follow that a detailed brainscan using nanomachines should preserve a perfect record of an individual’s personality and memories. With that being the case, cryonic freezing becomes a fully viable option for life extension, since any brain damage that may occur as a result of the process could be perfectly repaired using the brainscan as a guide.) Individual memories are stored scattered throughout the brain. Kurzweil says that this is similar to how a hologram stores information. Even adult brains can reorganize to a degree to meet new demands. Prions might also contribute to memory storage. Memories can be permanently erased through an active process in the prefrontal cortex. Brain cells of all types have their genesis in neural stem cells. The latter continue to form more of the former throughout a person’s life, though this process is seriously hampered by the physical impediment that preexisting brain tissue poses to the migration of the new. Direct injections of stem cells into damaged areas of the brain could be a treatment for some neurodegenerative disorders. Kurzweil explains research into models of sensory areas of the brain. The details are unimportant. [Kurzweil has really done a lot of research into brain physiology.] Only about 80,000 brain cells produce emotions. This relative simplicity raises the possibility that computers could be enabled with emotions much more easily that is commonly believed.

10

Animal experiments have shown that it is quite possible to directly interface organic brains with machines—in some cases to restore lost senses and in other case to give the animal direct control over a mechanical object linked to the computer. [Recall some of Ramez Naam’s citations] Neurons generate glial cells to coat foreign matter such as implants and protect the rest of the brain. Kurzweil believes that technology will offer a solution to this problem in the form of special coatings for implants that do not elicit the normal defensive cellular responses. Some brain implants have been approved by the FDA for medical uses in humans. Our ability to understand the human brain is directly related to the power of the computers that are used to analyze and simulate them, according to Kurzweil. [Same relationship exists between computing power and genetic mapping—Gene chips] Therefore, he believes that progress in understanding the brain and creating A.I. is proceeding exponentially, and he dismisses the “linear” view of his more conservative colleagues who estimate a hundred years or more as necessary for the task. “Uploading a human brain means scanning all of its salient details and then reinstating those details into a suitably powerful computational substrate. This process would capture a person’s entire personality, memory, skills, and history.” Kurzweil believes that the Turing Test will be passed by 2029. According to Kurzweil’s various calculations, brain uploading will be possible by the 2030’s. A nanobot brain scan would require that the nanobots monitor brain activity for an unspecified period of time. Chapter 5: GNR GNR stands for Genetics, Nanotechnology, and Robotics. Revolutionary advances in these three technologies will occur during the first half of the 21st century and will lead to the Singularity. An understanding of the human genome will lead to genetic engineering and radical life extension. But biology has its limitations and “will never be able to match what we will be able to engineer once we fully understand biology’s principles of operation.” Nanotechnology will allow humans to re-engineer themselves in more powerful and precise ways than genetic engineering by intelligently altering themselves on the molecular level. Kurzweil uses “Robotics” to refer to strong artificial intelligence. Each advance will bring perils: A greater understanding of genetics will allow highly dangerous new biological weapons to be created, nanotechnology will have the potential to—through accidental, uncontrolled replication or through deliberate misuse—completely destroy the Earth’s biosphere, and intelligent robots with mental and physical abilities far in excess of any human could pose a serious threat to the human race should they become pathological. Kurzweil admits this but also proposes countermeasures to each threat. The entire human genome is about 800 megabytes. Kurzweil predicted supercomputers with the ability to simulate protein folding by 2005. [This prediction did not come true—protein folding programs are only useful for small proteins. However, they can be highly accurate for such tasks.] [A number of factors exist that make protein structure prediction a very difficult task, including:

11

  





The number of possible structures that proteins may possess is extremely large, as highlighted by the Levinthal paradox. The physical basis of protein structural stability is not fully understood. The primary sequence may not fully specify the tertiary structure. For example, proteins known as chaperones have the ability to induce proteins to fold in specific ways. A particular sequence may be able to assume multiple conformations depending on its environment, and the biologically active conformation may not be the most thermodynamically favorable. Direct simulation of protein folding via methods such as molecular dynamics is not generally tractable for both practical and theoretical reasons. However, the distributed computing projects Folding@home are tackling such simulation difficulties. ]

As aging progresses, the cellular processes responsible for recycling misformed proteins start to degrade, causing the buildup of protein aggregates called protofibrils inside the cells. This is one proposed cause for Alzheimer’s disease. An understanding of protein folding will allow us to know the 3D structure of a protein based only on its DNA or RNA, which will lead to a breakthrough in pharmaceutical synthesis techniques. Kurzweil believes that the means currently exist to prolong the lives of the baby boomers until the Singularity, at which point techniques for extending life indefinitely will exist. So in essence, he believes that any person of middle age or younger could live forever. Kurzweil co-authored a book on the subject of personal fitness called Fantastic Voyage: Live Long Enough to Live Forever. Kurzweil claims to have cured his own Type II diabetes years ago through his own methods. He also managed to lower his cholesterol through another personal regimen. Kurzweil apparently became very obsessive with his health at age 40, and he is now (in 2006) 58. But the measure of his biological age shows him to have all of the vital signs of a 40 year old. [Photos of Kurzweil show that he clearly looks his chronological age] Kurzweil consumes 250 pills a day (!) and receives a half dozen intravenous “treatments” a week —none of which have been proscribed by a mainstream medical doctor (though Kurzweil did develop these treatments with the assistance of a physician who also co-authored the aforementioned book). He states that this has allowed him to “reprogram” his biochemistry so that his body operates in a manner much different from how it naturally would. Kurzweil also aggressively monitors many different indexes of his personal health (vital statistics, daily caloric intake, blood nutrient levels). Interestingly, Kurzweil claims that his personal treatments are “conservative” and exclude anything that is risky or unproven. A person who follows Kurzweil’s instructions will live long enough to see radical advances in medicine and nanotechnology. Those advances will allow rejuvenation of old bodies. [It should also be noted that Ray Kurzweil is a member of the Alcor Institute—the world’s leading cryonics firm. This might be proof of doubts he has about his own predictions about the progress of medical science, though Kurzweil would almost certainly disagree and claim that his actions are meant to preclude the possibility of premature death from injury] A full understanding of the aging process will allow scientists to create ways to stop and eventually reverse it.

12

Dr. Aubrey De Grey believes that techniques to extend the lifespan of animals will be demonstrated during the next decade, and that such developments will have a profound effect on public opinion about the inevitability to aging. Within a few years, the same techniques will be approved for human use. [Ramez Naam agrees exactly] The design of medical drugs will evolve from trial and error to a directed process based on our knowledge of biochemistry and molecular structures. RNA interference (RNAi) is a technique whereby double-stranded DNA complementary to mRNA translated from a bad gene is inserted into the cells. The DNA binds to the mRNA and prevents its transcription. Therapeutic cloning is a process where a person’s DNA and stem cells are used to regrow tissues and entire organs for their use. The regrowth may occur inside the body, or may be done externally with the finished product being surgically installed to replace the older one. Gene chips are small devices that allow a person to monitor their entire spectrum of gene expression. The technology has become dramatically cheaper and more accurate over time. Genetic profiling will allow: -Revolutionize the process of drug screening and discovery (poorly explained) -Improve cancer classifications (prevent misdiagnosis through a more accurate understanding of the changes in the body) -Identify genes, cells, and pathways involved in different biological processes such as aging -Determine the effectiveness of innovative therapies -Test the toxicities of compounds without the need for animal testing [Do more research into this] The first attempts at human somatic gene therapy were in 1990. Limitations of viral vector gene therapy: -Genes are sometimes too large to fit into certain cells like brain cells (neurons are small?) -Viruses cannot carry much DNA -Location of new DNA integration into host genome is currently uncontrollable Alternative new vectors: -Liposomes -Electric pulses -“Nanoballs” Kurzweil for whatever reason brings up the obvious need for genetic markers to be joined to the desired genes before infecting the person’s cells Kurzweil describes a number of conventional drugs under development. Kurzweil does not believe that human longevity was evolutionarily selected, and discounts the “grandma hypothesis.” Dr. De Grey proposes a preemptive treatment for curing cancer: use gene therapy to eliminate the genes in every cell that enable cancer to reproduce (a bit nebulous and fantastic). Mitochondrial DNA mutates faster than nuclear. Kurzweil proposes transferring the former into the latter as a “backup.” De Grey also wants to use somatic gene therapy to add new genes to the body that will enable it to break down intracellular chemical aggregates that interfere with normal biological functioning. As the body ages, it loses its ability to replace dead cells at a sufficient rate. Therapeutic cloning could cure this. Evidence suggests that there are only a few hundred genes involved in the aging process.

13

Kurzweil considers human cloning (creating a genetically identical copy of an existing person) to be unethical because current methods of doing such are imprecise (an electric “spark” is used to fuse the cell nucleus with an ovum) and cause a high likelihood of genetic mutation. However, once this issue has been resolved with advancements in technology, he does not see anything inherently wrong with the practice. However, Kurzweil doesn’t see any real value in human cloning. Kurzweil contemptibly describes human cloning as “a very weak form of immortality.” [Interesting that he is so adamant about transferring consciousness to electronic format, since he admits that he is not sure if the individual could be preserved through the process] Constructive uses of cloning: -Genetically engineering animals to produce chemicals (already being done with insulin) -Preserving endangered species and resurrecting extinct ones -Therapeutic cloning (current methods require the use of stem cells from human zygotes, raising ethical concerns among some) -Somatic engineering (includes the ethical conversion of one cell type into another and the rejuvenation of old organs by replacing them, one cell at a time, with youthful cells derived from the person’s adult stem cells—Kurzweil believes that this piecemeal “replacement” process could reverse the aging process) -Solving world hunger through the “hydroponic” growth of synthetic meat in food factories. This could, once the technology is perfected, be done very cheaply, and it would end all animal suffering. Some fear that genetic engineering would lead to homogeny since everyone would strive towards the same image of perfection. However, Kurzweil points out that this will not be true at least with respect to mental abilities: A person may have a genetic predisposition to a certain skill, but it takes practice and hence environmental stimulation to develop that inclination into an ability. People would still be defined by their experiences before genetic modification. [But what about once brain downloading of skills and experiences becomes possible? Then again, that’s not genetic engineering—that’s cybernetics.] Kurzweil believes that the genetic engineering revolution (i.e.- Gattaca world) will take decades to ramp up and will be overtaken by infinitely more important revolutions in cybernetics and computers. After all, what is the use of a perfect, genetically engineered body when an even better, externally identical, robotic body is available as well? At the current rate of development, computers and machines will be in the nano range—under 100 nanometers big—by the 2020’s. Biology has its limits. Engineered bodies and organisms will be superior. The conceptual birth of nanotechnology occurred in 1959 when Dr. Richard Feynman gave his speech: “There’s Plenty of Room at the Bottom.” The speech laid down his idea that there was nothing inherently impossible with manipulating matter at the atomic level to form any desired substance from another, and that such technology was inevitable and profoundly important. Eric Drexler founded modern nanotechnology in 1986 with the publication of Engines of Creation—originally his PhD thesis. No exact design for a nanomachine assembler has been drafted. The subsystems of an assembler: -Computer: Controls the nanomachine. Highly compact and efficient. -Instruction architecture: A more complex computer with detailed instructions for remotely controlling individual computers found in each assembler.

14

-Instruction transmission: Instructions sent to assemblers through electronic signals or mechanical vibrations -Constructor robot: Receives signals and has a single “arm” for moving atoms -Arm tip: Attaches to atoms via atomic attractive forces -Internal environment (inside the nanomachine body) free from interfering impurities -Energy: Some type of fuel system is built into the nanomachine, or the can be run remotely with “acoustic power at ultrasonic frequencies” A software description of an object will always be needed to replicate it. Therefore, the value of all things would be based on information. Drexler estimates the total manufacturing cost of any material produced by nanomachines to be 4 to 20 cents per kilogram. With that in mind, the production of some things would actually be cheaper with traditional means instead of nanomachines. An assembler is a machine that contains nanomachines and synthesizes new products. Each assembler would contain one type of nanomachine for the production of each type of subcomponent for the final product. Therefore, an assembler can contain many different types of nanomachines for making complex products. Instructions would be broadcast from the assembler’s computer to the nanites, and each nanite would be programmed only to heed the appropriate portions of the instructions. This is analogous to regulating gene expression in living organisms. Proof of the feasibility of nanomachines can be found in the architecture of life itself. Each cell contains a complete set of instructions, but only certain parts are turned on according to the cell type. The ribosomes act as molecular assemblers. Kurzweil proposes replacing cell nuclei and ribosomes with nanomachines. This would reduce transcription error rates to essentially zero, allow for intelligent, instantaneous, remote control of cellular processes (superior to traditional genetic engineering techniques), and defeat all viral infections and cancerous growths through control of the transcription of genetic information. But biology is limited by the requirement that all structures have a protein backbone. Engineered, synthetic cells could be thousands of times stronger than biological cells. Transmission speeds of neural impulses could also be accelerated by a factor of millions. Making nanomachines dependent on certain types of nonabundant energy sources should keep them from replicating in an uncontrolled manner, but more advanced nanomachines will doubtless be more capable and be able to obtain energy and materials from a wider range of substances. Kurzweil worries about this as a threat to the biosphere. Miniature gears, rotors, and levers have already been invented in nanoscale. Carbon nanotubules and DNA can both be assembled into a variety of structural arrangements and can store information. Across the world, scientists have already created primitive nanomachines that can perform simple functions. No serious flaw in Drexler’s nanoassembly concept has been found, according to Kurzweil. Nobelist Richard Smalley criticized Drexler’s idea of the nanoassembler. At the nanoscale, having any more than 1 or 2 manipulator arms would render the machine useless since atomic forces and direct physical interference would interfere with the movement of the arms (the “fat fingers” problem), and “molecular attraction forces” might make it difficult for the arms to release atoms from their grips (the “sticky fingers” problem). Kurzweil claims that Smalley has misunderstood Drexler’s proposal, and that cellular functions we normally observe are proof that

15

nanomachines can operate in principle. Kurzweil also claims that experiments conducted since Smalley’s thesis have debunked the “sticky-“ and “fat” finger problems. Nanomachines can only accomplish tasks at a reasonable speed when there are trillions of them working together. The only economical way to make trillions of nanomachines is through selfreplication. Most experts on the matter agree that nanomachines must be built from diamondoid or carbon nanotubules to withstand thermal extremes and atomic vibration effects. Quantum effects (i.e.—you cannot simultaneously know a particle’s location when it is in movement) are relevant only for electrons. Protons and neutrons—which are much more massive —could be manipulated by nanomachines. Nanomachines and their products could thus be assembled. It may be possible to use ATP and glucose/oxygen to fuel nanomachines, which will be advantageous since both substances are abundant in the human body. Kurzweil cites other academic attacks on Smalley’s ideas. Smalley’s main point seems to be that molecular assembling done in biological systems (Kurzweil’s proof of nanomachine validity) always occurs in an aqueous environment and with protein. These two substances are needed, he argues. Even if this is true, it does nothing against the potential uses of nanotech to augment and repair the human body. Kurzweil attacks the fears of uncontrolled nanotech reproduction by citing the early fears about vulnerabilities created by the Internet and worldwide communications systems. While both could theoretically be taken down for crippling effect, defenses evolved along with the tools of malicious people. [However, Kurzweil fails to mention that an uncontrolled nanotech breeding cycle would be far more devastating than even the worst Internet virus.] Medical applications of nanomachines have thus far been dedicated to developing antibody-like “nanotrackers” that bind to desired substances and make them easily detectable. Eventually, nanotrackers will allow a small sample of blood to be tested for hundreds of different substances simultaneously. Nanoscale scaffolds could be used to grow organic tissues. [Kurzweil gets a hit—Urinary bladders] Nanoparticles can be designed to deliver drugs to certain parts of certain cells, and to allow penetration of cell membranes and the blood-brain barrier. A machine implant is being developed that could be installed under a person’s skin to release precise amounts of medicines that the person needs. It could be on the market by 2008. Through internal monitoring of this manner, implants with nanoscale monitors and computer controls could restore normal chemical levels in people with health and organ problems. [Possibility of using brain implants to counteract chemical imbalances?] Kurzweil is against the use of fossil fuels and dislikes the geopolitical ramifications. Energy usage will grow much more slowly than computing power. The nanotech revolution Kurzweil predicts to happen will essentially make manufacturing and energy production into information technologies that will be subject to the Law of Accelerating returns. Kurzweil is aware of Peak Oil. Improvements in technology will make alternative energy cost-competitive with petroleum. Reversible computers will cut energy requirements for computers to a small fraction of what they currently are. Smaller logic gates and memory bits will further decrease energy consumption.

16

Such computers will be built with features on the nano- scale and will thus be examples of nanotechnology or will require nanotechnology to be constructed. Nanotech assemblers will be extremely efficient compared to traditional means of manufacture since everything will be built from the atomic level up. Very little energy will be wasted. Resources will not have to be expended moving raw materials to factories when nanotech assemblers could just synthesize the materials on site. Kurzweil cites Drexler’s claim that nanomachines could be energy producers (badly explained). Engineered synthetic materials will replace steel and titanium and be superior in all respects. Light bulbs will be replaced by cool, light-emitting diodes. Smalley’s ideas for using nanotech to improve energy transmission: -Nanotech can dramatically improve the economy of producing solar cells -Use of nanotech to produce hydrogen from water and sunlight -Ultrastrong, engineered materials could safely store hydrogen gas -Cost of making fuel cells will decrease -Improvements in battery storage efficiencies by factors of ten to a hundred -Strong and light nanomaterials will improve the safety and efficiency of automobiles and airplanes -Robots can create energy harvesting devices in space -Nanomaterial coatings could somehow reduce the cost of deep drilling (lubricant?) -Nanocatalysts could improve yields from coal -Nanofilters could capture soot from coal burning or separate different gases Transmission of energy captured by orbital solar arrays back to the Earth through wireless microwave beams. Major advances are also being made in microscopic fuel cells, which can serve as batteries for appliances such as laptop computers and boast dramatically superior lives. Advances in fuel cell design have yielded cells that can use stable fuels like ethanol for energy. Fuel cells have been made that run off of glucose and oxygen in human blood. Fuel cells will be used to power nanomachines. Fuel cells can utilize bacteria to convert substances into energy. Kurzweil believes that solar power is the future. If we were to take it seriously, it could meet all of our energy needs, and all the needs of future technology. Nanomachines could be used to cleanse the atmosphere of excess carbon dioxide. Conversion of solar energy to electricity is only 34% efficient. Evidence suggests that efficiencies greater than 60% are possible. Kurzweil doubtfully cites claims by Nanosolar, Nanosys and Konarka that the cost per watt of solar power could be decreased to $.50 per watt by 2006 (check it out). Nanotechnology will allow extremely thin solar panels to be produced for almost nothing. It could be put on any surface, including clothing. Kurzweil is a fan of the space elevator. Desktop fusion is a possibility, but Kurzweil is not banking on it. Nanotechnology will allow for much more efficient production of goods and hence much lower pollution. Existing pollution can be cleaned up. All energy needs can be met with solar power. Powerful filtration of water.

17

Improvements in fuel efficiency resulting from the use of lighter nanoscale composites in vehicles. Nanomachines could separate isotopes during the processing of nuclear fuel and could be used to dispose of nuclear waste. Elimination of dangerous chemicals used in electronics manufacturing by replacing them with nanomachines. Robert A. Freitas Jr. is a leading proponent of the medical uses of nanotechnology. He has designed nanomachines that could replace discrete cell types and organelles (RBCs, WBCs, DNA-repair bots) and perform their functions thousands of times more efficiently. Remote control of nanobots in bloodstream. Use of remotely controlled nanomachines to control internal biological processes will leave the body open to attack by computer viruses. Kurzweil dismisses this by citing the current existence of secure Internet firewalls as proof that failsafe systems can be engineered. Kurzweil views this is a risk outweighed by the benefits. For example, no one suggests destroying the Internet because of the potential damage computer viruses could inflict. Nanobots could reverse the aging process and repair any type of damage. “Robotics” refers to artificial intelligence that is superior to unenhanced human intelligence (Strong A.I.). Intelligent machines will have a number of advantages over humans that will make them superior. They can share information instantaneously, combine powers to work together more effectively than a team of humans can, store and perfectly recall enormous amounts of information (someday, an individual machine will “know” all human knowledge), and master broadly different sets of skills that no individual human could have the talent or time to master in a lifetime. For these reasons, the first synthetic brain that is capable of human pattern recognition and parallel processing will soar past human intelligence. Once Strong A.I. exists, all scientific discoveries will be made by it. We can expect the rate of technological advancement to explode. Kurzweil is not sure whether A.I. or nanotechnology will come first. Kurzweil believes that, most likely, advances in each field will simultaneously facilitate the other. Kurzweil says advanced nanotechnology will arrive in 2025 and Strong A.I. (pass the Turing test) in 2029. Runaway A.I.: The first A.I.’s will improve upon themselves and beget better A.I. in an accelerating cycle of enhancement. Kurzweil believes that the first A.I.’s will be primitive and only equivalent in abilities to average humans: faster processing speeds and better memory recall will be of limited use until they know how to apply their knowledge. Runaway A.I. will not occur until the mid 2040’s. Kurzweil dismisses the claims that A.I. is impossible and believes that average people are disillusioned because of the failure to create A.I. in the 1980’s (the “AI Winter”). “The technology hype for a paradigm shift—railroads, AI, Internet, telecommunications, possibly now nanotechnology—typically starts with a period of unrealistic expectations based on a lack of understanding of all the enabling factors required. Although utilization of the new paradigm does increase exponentially, early growth is slow until the knee of the exponential growth curve is realized. While the widespread expectations for revolutionary change are accurate, they are incorrectly timed. When the prospects do not quickly pan out, a period of

18

disillusionment sets in. Nevertheless exponential growth continues unabated, and years later a more mature and more realistic transformation does occur.” “Narrow A.I.” is A.I. that performs specific functions better and faster than humans, and frees humans from the burden of that work. Narrow A.I. is everywhere. Kurzweil believes that the A.I. revolution will take much longer than the Internet and telecommunications revolutions since sentience is so complex. A.I. scientists and engineers have attempted to model consciousness without the full knowledge of the human mind that advanced brain scanning techniques will provide: -Expert Systems are computers with long series of Boolean rules which together yield complex abilities of judgment. The construction of Expert Systems revealed how truly complex human judgment is and how it often relies on subtlety. Attempts to create massive Expert Systems are still in progress. -Bayesian Nets gather data in an ongoing fashion to predict the probabilities of future events based on past events. Some SPAM filters use this process. -Markov Models -Neural Nets are simulated “brains” with electronic “neurons.” The neurons are put into multiple layers, with each layer randomly connected to the other. Responses to inputs are initially random. With external assistance (teaching), connections between neurons that yield correct answers are strengthened while errant connections are weakened or cut. Neural nets have to potential to mimic human pattern recognition. -Genetic Algorithms allow different “solutions” to a problem to compete with one another. The best solutions are allowed to “reproduce.” “Mutation” is also present. When the improvement of the solutions stops increasing from generation to generation, the optimal solution has been found. -Recursive Search is an algorithm for games with a clear set of rules and possibilities. Chess Recursive Search programs, for instance, examine each potential move and countermove in an enormous “tree” of possible options, and mathematically decide which move is best. Chess programs have improved and actually complete FEWER Recursive Search computations because their pattern recognition abilities have improved greatly. Unpromising “branches” of the recursive search tree and more quickly recognized and “pruned” without wasting any effort on further analysis. Kurzweil believes that the best way to create A.I. will be to create a computer capable of using all of these forms of computation. He also believes that the human brain’s regions operate differently, and combine in effect to yield intelligence. No one area of the brain is “sentient,” and all different methods must be combined for sentience and cognition. Narrow A.I. examples -Cruise missiles have pattern recognition software that allows them to alter course to follow the features of the terrain -Unmanned spacecraft have autonomous software that allows them to respond to mission threats without long-distance human intervention -Kurzweil’s company is working on clothing with embedded ECG sensors for monitoring cardiac health -Pattern recognition software is used in diagnosing diseases from tests -AI is used by businesses for data mining about consumers -Automated phone operators -Many other examples -Basically, limited AI has some small ability to think for itself and to adjust to changes

19

Kurzweil believes that truly convincing human speech will first require human-level intelligence. Without narrow AI, the industrialized world would come to a halt. By 2014, Kurzweil predicts house-cleaning robots will be common. Machines will surpass humans in one area at a time, but skeptics will only see the areas where humans are superior, so the dawn of strong A.I. will creep up on us. Kurzweil believes that a full understanding of the human brain’s functions by the late 2020’s is a conservative forecast. Strong A.I. will result from reverse-engineering the human brain and using varied models of computing for different tasks within the artificial brain. [Kurzweil may be underestimating the timeframe to keep the advancements within his own projected lifetime] Educating A.I. is the most difficult part. With access to human knowledge and human interactions over the Internet, computers will be able to learn sufficiently on their own. The parameters of the Turing test are vague, so there will be a period of time in which machines will pass certain versions of the test, but experts will disagree over whether it truly emulates human speech. [Fits in with “creeping” increase in capabilities previously stated] “The advent of Strong A.I…will mean that a creation of biology has finally mastered its own intelligence and discovered means to overcome its own limitations.” Chapter 6: The Impact As the Singularity approaches, people will have to reconsider the meaning of human life and will have to redesign human institutions. The human body 2.0: -Nanobots will be both embedded and mobile within the body, destroying pathogens, correcting DNA errors and controlling DNA expression, and eliminating toxins. In effect, the person becomes immune to biological diseases and enjoys eternal youth and medical immortality. -In the brain, nanobots will augment mental abilities and allow direct, mental interface to the Internet and full immersion in virtual realities. With humans so dependent on their technology, increasing emphasis will be placed on cyber warfare. The destructive potential of nanotechnology will also not be overlooked, and nano weapons will doubtless be created. With machines able to perform all routine functions of daily life, people will work to expand knowledge or to pursue the arts. Intelligence on Earth will continue to expand exponentially after the Singularity, and will eventually reach the limits of the ability of matter and energy to support computation. “Intelligence” will radiate outwards from the Earth at the speed of light, and possibly faster [Weird]. Synthetic organs and medical nanotechnology will prove their superiority over genetic engineering, and people will steadily upgrade themselves piecemeal with new, robotic organs. Human evolution has not favored long lifespans because older people are unable to engage in physical labor and are thus a waste of resources. [Aging and the concomitant reduction in abilities thanks to biological damage and imperfect repair mechanisms] With our technology freeing us from the burden of physical work, people can remain useful until their mental abilities start to degrade.

20

Kurzweil’s idea for altering the digestive system: -The human digestive system is a holdover form a much earlier period in our history in which all caloric intake had to be conserved due to sudden alterations in food supplies. -Today, that metabolic strategy is incongruous with our lifestyles and has caused an epidemic of obesity, diabetes, and other diseases related to overeating. -The “fat insulin receptor” (FIR) gene has been identified in mice. [Naam also mentions IGF-1 insulin gene] FIR knockouts ate without restriction but remained lean and healthy and lived 18% longer than normal mice. Humans may have a FIR gene. -Nanomachines in the bloodstream wirelessly controlled by computers monitoring nutrient and activity levels could perfectly regulate the intake of substances and the excretion rate. -Primitive artificial medicine delivery systems are already in development. -Nutrients could be “introduced directly into the bloodstream” by nanomachines for highly efficient food intake without actual eating. -Kurzweil thinks that these “food nanobots” could be stored in a reservoir located in a shirt or belt, and that they could enter and exit the body through the skin or other body cavities (!) as needed. He believes nanobot feeding will be a possibility by the late 2020’s. -Just as sex has become removed from its natural function of procreation and is now done mostly for pleasure, eating could be superseded by nanomachines as a mechanism of nourishment and could also be done just for pleasure. -Nanobots with the ability to filter wastes and deliver nutrients would render the kidneys and digestive organs unnecessary. -Eventually, the food nanobots will not need to be carried with individual people and will be “embedded” throughout the environment (?). -Food nanobots will also be able to efficiently store nutrients, including oxygen and glucose, so that a superhuman could survive without nourishment for extended periods of time. -The digestive system will be replaced gradually, just as typewriters remained for a time after the introduction of the first word processing computers. Programmable blood: -Freitas has designed synthetic blood cells (red, white, and platelets) that perform their functions far more efficiently than their organic counterparts. -It is estimated that his respirocytes (RBC’s) could store oxygen hundreds or thousands of times more efficiently than RBC’s, enabling a person to survive for hours without breathing. -Banning augmented people from competing in the Olympics would eventually create a very bizarre situation in which average “superhumans” of all ages routinely outperformed highly trained yet unmodified humans in tests of physical abilities. [It might eventually lead to a loss in interest in “unaugmented” professional sports and the Olympics since the achievements would be partially meaningless] -Artificial platelets could form clots hundreds of times faster than natural ones. -Programmed WBC’s could download instructions and immediately destroy pathogens or cancers. Natural pathogens could not develop any immunity to artificial WBC’s. The Heart and Lungs: -The heart is often the cause of death and usually fails before other organs. As such, Kurzweil considers it a weak point (his father also died from a heart attack). -Kurzweil ignores the idea of artificial hearts and looks to Freita’s idea to dispense with the heart altogether.

21

-Synthetic blood nanobots will have their own means of locomotion, so there is no need for a heart. -Nanobots will eventually be able to manage all gas exchange, so the lungs could be dispensed with (what about speech?). -The requirement of needing “breathable air” is “burdensome.” Hormones could also be synthesized by smaller, artificial organs and delivered by nanobots, eliminating the need for the natural endocrine system. Nanomachines could embed themselves in bone tissue and gradually replace the bone mass with some artificial material or with a “mesh” of nanomachines. The skeleton could be self-repairing. Kurzweil believes that dispensing with the skin altogether would be a mistake since it would lead to loss of sensation. Improving upon the durability (nanoengineered supple materials) and sensitivity of the skin would be desirous, however. Kurzweil doesn’t mention improvements to sensory organs. Artificial brain implants can remotely manipulate nerves, and can even electronically “reconnect” severed nerves. By scanning a disabled person’s brain activity when they try to move their disabled limbs, the computer can recognize the brain pattern generated to elicit a specific movement. A computer scanner can then be permanently inserted in the person’s brain and can transmit stimulatory signals to electrodes in the person’s muscles or somatic nerves that produce the desired physical action. It is possible that people who have been disabled for long periods of time could have some motion restored through this technology. These are “nanoelectromechanical” systems (NEMS). The integration of computers into the human body is merely a continuation of the trends of computer miniaturization and user-friendliness. The medical potential of computer implants will spur research, and eventually, the line between restoration of normal abilities and augmentation will disappear. Stephen Hawking has recognized the coming emergence of superior A.I. and believes it necessary that a means to link human and machine thinking be developed to ensure that the two groups do not exist in opposition. Human Body 3.0 -Possible in the 2040’s -Abandonment of a fixed, corporeal form -More focus on virtual reality, where people can change identities -Use of nanomachines on the exteriors of our bodies to alter external appearance at will -Possible use of “foglets” to alter reality and as a replacement for flesh People are often put off by the notion of replacing body parts and organs with machines because they fail to look beyond current conceptions of machines. Machines of the future will be much more advanced and amenable to a fusion with biology. Kurzweil’s predictions for the 2010’s: -Computers become increasingly small and integrated into everyday life. They are found woven into clothing, embedded in furniture, and in our eyeglasses. -Many of these computers and small, handheld computer accessories like iPods will be connected to the Internet, and they will all act as miniature web servers that pool their resources for computation. -Wireless, high bandwidth Internet access will be available almost everywhere. -Displays built into eyeglasses will project computer images onto retinas. -Cell phones built into clothing to project sound directly into ears (close to reality—BlueTooths)

22

-Projection of sounds over long distances to specific people using two ultrasonic beams that intersect at the desired point. (Minority Report advertisements) -The eyeglasses and audio hardware will allow for full-immersion virtual reality. -“Virtual assistants” will be computer programs built into wearable computers like the eyeglasses that will help the user during their daily routines. -Subtitles will give real-time translations of people speaking in foreign languages. Kurzweil’s predictions for the 2030’s: -Brain nanobots will allow for full-immersion V.R. (similar in principle to the world of The Matrix) by controlling our afferent sensory neurons. Real reality can be cancelled out. -People could browse the Internet using V.R., interact with other people, possibly with a false identity if desired, or visit new places. -Kurzweil found the experience of being Ramona to be very liberating. But is it just the novelty of the thing that makes it interesting? -Massively distributed brain nanobots will also be able to elicit emotional responses. -“Experience beamers” will be recorded or real-time brain transmissions of a person’s daily life that other people could remotely watch like a very advanced equivalent of a webcam. (Being John Malkovich) [Along the lines of the “Computing 2016” article] -Nanobots will allow people to greatly expand their cognitive, sensory and memory abilities, to directly communicate with computers, and to “telepathically” communicate with one another via wireless uplinks. People will be able to alter themselves in virtual reality starting in the 2030’s, and to alter themselves in real reality (foglets) in the 2040’s (hmm…). [What about downloading others’ knowledge and experiences?] Once the shift is made towards thinking with nanomachines, people will be able to alter their brain connections at will. [Potential for personality change?] Nanobots have advantages over brain implants: -Can be introduced without surgery -Can be withdrawn easily -Can change configuration and function -Can be massively distributed V.R.’s implications for prostitution, social affairs, and the workplace? Implications for marriage when also considering extreme longevity? [Kurzweil spends too much time thinking about virtual sex.] Biotechnology Advances: -Genomics (sequencing entire genomes of species) -Proteomics (study of proteins) -Gene therapy (suppressing expression of certain genes) -Rational drug design (more advanced, specific drugs) -Therapeutic cloning (entire limbs, organs, and individual cells—all from host DNA and with DNA errors corrected and telomeres restored) Advances in biotechnology and medical nanotechnology will eliminate medical causes of death. People with nonbiological computer brains could “back up” their minds, preserving a perfect copy of themselves in a safe place for activation should they die. This would be a form of immortality, though it would not allow for a continuation of consciousness. Kurzweil seems to disagree and believes that the “patterns” of a person’s brain [vague] are the keys to consciousness. [More likely you would just make a copy of yourself]

23

Kurzweil again states that he believes the humans of the distant future will have morphable bodies. The transfer of our consciousness to computers and the future Internet will allow people to “exist” without a corporeal form, being both everywhere and nowhere at the same time. Physical and virtual reality bodies will become just temporary manifestations that people take to complete certain tasks. Though this sort of existence, immortality could be achieved. Kurzweil believes this will happen by the middle of this century. The argument can be made that existence depends on having a constant physical form. But every cell in the human body, including the cells of the brain, are constantly replacing old damaged and aged structures and organelles so that every few weeks, every atom in the body is “new.” It is the pattern and arrangement of the atoms in a person’s mind that hold their consciousness. If this pattern could somehow be smoothly transferred into an electrical format, the person could continue to exist, but in a radically altered and infinitely more resilient form. Interestingly, Kurzweil does not give a clear pronouncement about whether or not he thinks this transfer could be made in a way that truly preserves the continuity of consciousness. If people believe that death is inevitable, then there will be little choice but to rationalize it as necessary and even ennobling. [Random: Nanobot brainscans will preserve a perfect record of brain structure and activity. This makes cryonic preservation a viable choice for life-extension, since brain damage from the freezing process could be perfectly repaired and the old brain reassembled without inaccuracy.] Information in an electronic brain will still degrade with time. Making backups, upgrading formats and other forms of maintenance are critical to immortality. Kurzweil is a member of the Army Science Advisory Group (ASAG) and shares some coming trends in weapons technology being pioneered by the U.S. military: -Greater emphasis on technology, smaller numbers, and speed to win battles. -“Future Combat System” (FCS) -The backbone of the Army’s ground forces will switch from the Division to the Brigade Combat Team (BCT). Each BCT will consist of about 2,500 men, will be self-sufficient and capable of undertaking all missions. All soldiers in the BCT will have a Common Operating Picture (COP) of the battlefield cobbled together from different intelligence and reconnaissance sources. Soldiers may receive appropriate COP information through a heads-up display, or in the more distant future, through direct neural transmissions. The mobility of the BCT and the individual soldier will be improved. -Robotic “mules” will carry equipment, lessening the load for soldiers. -Advanced uniforms made of Kevlar composites are being developed. They are flexible like normal fabric under normal conditions, but instantly harden when stressed by a bullet impact or knife stab. -Advances in engineering and countermeasures to anti-tank weapons will allow armored vehicles to be lightened without sacrificing protection for the crew. -Mini-UAV’s will be used for short-range reconnaissance. -The wireless communications/processing system that compiles and relays the COP will be distributed among all soldiers and all small units, eliminating central hubs and hence making the system more survivable. Significant efforts will need to be made to ensure that the encryption is adequate. -Certain types of UAV’s and land vehicles will be autonomously controlled by computers by 2025.

24

-Kurzweil believes that development times for new weapons will reduce as part of the overall acceleration in technological advancement. The development of more intelligent and detailed computer simulators will in large part be responsible for faster and more accurate weapon testing. -Substituting autonomous and remotely controlled robots for humans on the battlefield will also obviously save lives. -DARPA is developing remote sensing devices nicknamed “Smart Dust.” About the size of a pinhead, they could be dropped by the millions over enemy territory to provide reconnaissance. If the ability to read DNA were added, they would be able to identify individual people. -Nanotechnology has myriad military applications, many of which do not involve the “gray goo” effect. -Virtual reality devices might be needed to enable soldiers to remotely control advanced robotic devices on the battlefield. -Advances in warfighting technology will mean that controlling your own information while disrupting the enemy’s will become the keys to victory. Education remains unequal in quality across the world and even just between communities. The Internet and high quality computer teaching software programs will enable people to effectively learn online. MIT will have video recordings of all of its courses online by 2007. Online education can be disseminated much more cheaply than traditional education, and will free communities from the limited teaching skills of local teachers. Interactive programs can analyze the strengths and weakness of students and tailor a learning program. [GRE and some versions of the SAT] [But let’s remember that certain types of testing cannot be graded by computers. Essays, for instance, require humans for critical analysis. The bottom line is that learning cannot be entirely home-based, and that some interaction with a human instructor is needed.] Virtual reality will one day allow even more effective remote learning. People in any corner of the world will be able to enjoy a high quality education once the technology becomes cheap enough. Eventually, once nonbiological intelligence is introduced into the human brain, people will simply be able to download information. [Kurzweil again fails to understand that children are unable to learn from computers alone. Older people—yes—but not kids.] Because technology raises standards of living and because the rate of technological advancement and cost-performance is increasing exponentially, Kurzweil believes that the world’s “underclass” will disappear over the next 2 or 3 decades. [Uh…] Kurzweil also predicts a strong Luddite backlash against GNR. With the advent of nanotech manufacturing, any product will be able to be produced for almost nothing. The true cost will be in the information that codes for the product by the late 2020’s. Open-source and proprietary information will coexist as it does today. Intellectual property rights will have to be safeguarded in the future to protect the economy. Most routine and manual labor will become automated. Decentralization of the economy and of society will proceed thanks to advances in several fields of technology. Kurzweil believes that V.R. will be indistinguishable from real reality by the late 2020’s. Why Kurzweil believes we are the only intelligent species in the Universe: -The Fermi Paradox (aliens might not necessarily be here, but they will be impossible to miss)

25

- Even an advanced civilization with exotic means of communication will still use the electromagnetic spectrum for some purposes, so we should be able to detect them -Kurzweil’s own assignment of values to the variables in the Drake equation yields only one intelligent species in our galaxy—the human race -The Fermi Paradox is only a paradox if people accept the optimistic parameters applied to the Drake Equation -Once a certain point of enlightenment is reached (the Industrial Revolution), technological growth is explosive and leads to highly advanced capabilities in just a few hundred years. There is no “gradual increase.” -Kurzweil cops out on pondering reasons for the Fermi paradox that allow for intelligent alien life. [For instance, advanced aliens may choose not to interfere with human affairs for a number of reasons. Radical technologies may allow them to mask all evidence of their existence, including electromagnetic emissions.] The limits of computation revisited: -A civilization capable of 1042 cps per kg will be capable of finding a way to leap to 1043 cps per kg thanks to the problem solving abilities of their computers. -Kurzweil believes that we might reach a point where we convert entire planets into computers. -Concentric Dyson spheres or Dyson shells could harness the power of the sun and serve as massive computers. But Kurzweil doesn’t think they will be used for habitation since nonbiological beings will predominate by the time they are invented. -Kurzweil believes that making celestial body-sized computers will be possible by the end of the 21st century. [Uhhh…] -If the entire mass of the Solar System were used for computing, and if all mass were most efficiently used for the purpose, we would be capable of 1080 cps. -Kurzweil believes that there is an inward limit to transistor size, and that once this limit is reached, the only option will be to increase the size of computers to further increase computer power. The popular conception of space travel is wrong: Most interstellar probes will be intelligent nanomachines. Smaller size means that they could travel nearer the speed of light. The nanomachines could collect information and self-replicate. Kurzweil’s theories on the superluminal -Photon “flight time” is twice the speed of light, but is not useful since no information can be conveyed. -Quantum disentanglement means that two particles (photons only?) created at the same time will share some of the same properties, so that they “communicate” their properties to one another at speeds far in excess of the speed of light. While there may be a third, internal variable that is the same in each particle to explain the identical properties, most quantum physicists believe the particles are in communication. -Quantum disentanglement is not useful for transmitting information, but it could have uses in making unbreakable encryption codes that could also not be eavesdropped upon. -Galaxies are also moving away from each other at speeds that in some cases exceed the speed of light. Since this is due to the expansion of space due to the expansion of the Universe, it does not violate Einstein’s Theory of Relativity. -Einstein and Nathan Rosen formulated “Einstein-Rosen” bridges as subatomic wormholes wholly consistent with the Theory of Relativity. Later on, it was discovered that these wormholes could theoretically be enlarged (through extreme energy expenditure) to allow larger objects to

26

pass through them. [Negative energy is required to keep wormholes open, and as yet, there is no way to produce appreciable quantities of this.] -Kip Thorne designed an artificial wormhole that has two units. A wormhole linking the two is opened and stabilized, and one wormhole is transported to a different location. “Thorne offered the example of moving the remote entrance via a small rocket ship to the star Vega, which is twenty-five light-years away. By traveling at very close to the speed of light, the journey, as measured by the clocks on the ship, would be relatively brief [3 months at 99.995 percent the speed of light]…Although the time for the voyage, as measured on Earth, would be around twenty-five years, the stretched wormhole would maintain the direct link between the locations as well as the points in time of the two locations. Thus, even as experienced on Earth, it would only take three months to establish the link between Earth and Vega…By traveling arbitrarily close to the speed of light [i.e.—99.9999999% the speed of light = almost no time passing inside the spacecraft], the time required to establish a link—for both communications and transportation—to other locations in the Universe, even those millions of billions of light years away, could be relatively brief.” -Natural wormholes may also exist in abundance and could be used by humans. -Astronomer John Webb discovered that, through examining the fine structure constant, the speed of light varies by a very small amount [cited earlier in book] in the Universe. Others believe that humans could artificially alter the speed of light to allow faster space travel.

The Fermi Paradox Revisited -John Smart has suggested the “transcension” scenario where civilizations move on to different, more advanced universes once they reach a certain level of advancement.

27

-Kurzweil contradicts himself by saying that alien observers will certainly be able to hide their presence thanks to advanced technology. [Shouldn’t this also be true for masking radio waves from reaching the Earth?] The Laws of Physics and the Universal physical constants (such as Planck’s constant) are perfectly balanced to enable the existence of organized matter and of life. Small differences in constant values would have led to a totally lifeless Universe. Some people feel the fact is irrelevant (i.e.—if the Universe didn’t have the appropriate constants, no one would be around to comment on it), while others believe it is proof of divine engineering. String Theory suggest that we may be part of a Multiverse made of a large, yet limited, number of Universes. Susskind and Smolin have suggested that Universes “reproduce” by creating new Universes through black holes. Universes with physical constants similar to ours were favored for poorly explained reasons. Apparently, stable carbon is necessary both for life and for black holes. James Gardner combined the Susskind/Smolin theory with Kurzweil’s own theory to theorize that intelligent life is capable of creating new Universes once it reaches a sufficient level of technology. Thus, Universes most conducive to supporting life through their physical properties were also more likely to beget new Universes. [If our Universe was created by intelligent beings, does that mean there is/was a God? Why would such beings create our Universe and then choose not to populate it?] Gardner and Kurzweil agree that the ultimate destiny of the Universe is to be converted from randomness into a thinking mind by human intelligence. Once the limit of transistor shrinkage has been reached, the only way to increase the power of a computer will be to expand its size. Once a certain size is reached, gravity will cause the computer to collapse into a black hole. Kurzweil thus believes that an “organized” black hole could be the most powerful computer design possible. Stephen Hawking and Kip Thorne concluded in 2004 that black holes were indeed capable of computation. If two, quantum entangled particles are created near the event horizon and one goes into the black hole while the other does not, computation affecting the black hole particle will be transmitted to the free particle. Kurzweil believes that intelligence (technology) can overcome physics so that sentience will shape the structure and destiny of the Universe, and he admits that he is in the minority with this view. Engineering a desired Universe is the ultimate goal of the Singularity, according to Kurzweil. Kurzweil believes that not all of the matter in the Universe will have to be converted into computers, and that many Universal events regarded as worth preserving could be left unmolested. Provided that the speed of light can be circumvented and that current trends in increasing computer sophistication hold, Kurzweil predicts that we could saturate the matter in our Universe with intelligence by the end of the 22nd century. A failure to break the light barrier would impose a severe limit upon the expansion of computation, so that it would take billions instead of hundreds of years to saturate the Universe. Trying to understand Posthuman values and life outlook is like having ancient bacteria attempt to conjecture what it would be like to be human. Chapter 7: Ich bin ein Singularitarian

28

A Singularitarian is a person who understands and believes in the Singularity. Belief in the Singularity is not a matter of faith, but rather results from an examination of scientific trends. Ray’s personal philosophy: -The technology exists today for baby boomers and following generations to live to the Singularity if they aggressively apply what is currently known about human health (the “bridge to a bridge to a bridge”). Consider Ray’s personal regimen of medical treatments. -The body is temporary: The particles composing every person completely turn over during the course of a month. Only the pattern has continuity. -Merging with our technology will allow us to vastly expand our minds. MNT fabrication will allow morphing bodies. -Technology has the power to solve age-old human problems which have no other solution. -There is a difference between information and knowledge. Intelligence selectively destroys irrelevant information and preserves knowledge, which usually reveals a pattern. -Death is a tragedy since it represents a permanent loss of knowledge. -People who claim “death is necessary” are merely trying to rationalize something that is inevitable—to try and put the best face on something horrible that they can do nothing to prevent. -The Posthuman world will give new richness to life and will make eternal life worth living. -The purpose of life is to create more order and hence more knowledge. -The purpose of the Universe is the same purpose as human life: to create greater order and to gain control over itself with greater intelligence. -Human-inspired, nonbiological intelligence will radiate outwards from the Earth by the end of this century. Kurzweil is against “passive Singularitarianism”—a tendency to grow detached from mundane, everyday concerns because of the belief that all problems will be eventually solved by technology. Kurzweil believes that technology must be aggressively pursued to save as many lives as possible now, and at the same time, the application of technology must be carefully controlled to prevent evil uses. Kurzweil believes that the antitechnology movement will fail and points to the ineffectiveness previous upheavals of the 20th century had on technological advancement. [Yet he fails to look to more distant times and see that regimes are indeed capable of halting and even reversing science.] Kurzweil only believes that the Luddites will at best stall the Singularity and cause a lot of unnecessary suffering in the process. As with Naam and McKibben, Kurzweil believes that there is no clear dividing line between therapeutic repairs and cosmetic upgrades. Technology is a slippery slope. Kurzweil sidesteps the issue of GNR technologies creating a new species by saying: “…the whole idea of a species is a biological concept, and what we are doing is transcending biology.” Kurzweil’s ideas about a new religion: -The first principle is to respect all intelligent life. Essentially, it is the Golden Rule. -The second principle is to have respect for knowledge. -Animating the matter of the Universe will create a supreme intelligence which humans could merge with at will. This would be the closest thing to God. Kurzweil regards biological intelligence’s inability to combine to mental powers as an unfortunate limitation. Computers can combine or separate at will, making them superior. Intelligent machines

29

-Reverse engineering of the human brain will allow artificial intelligence to be created. -Brain uploading of humans will allow computer duplicates of people to be created, but there is the question of whether there would be a transference of consciousness. -Kurzweil’s “most compelling” thought is that humans will augment their organic brains with progressively more and more advanced cybernetics until the nonbiological portion predominates. [Very similar to my idea! But what about gradually dispensing with the organic tissue altogether?] There exists no objective test for the presence of consciousness, making the future judgment over whether machines will have sentience a subjective one. Posthumans will possess the “finer qualities” of human nature to a much greater degree, making them, in one sense, more human than Homo sapiens. [Whatever…] Kurzweil argues that, regardless of whether future A.I.’s are really sentient, their behavior will be so convincing that it will be impossible to prove it either way. Kurzweil thinks consciousness is important because our morals and laws are built around the concept. Eventually, the law will have to recognize machines as sentient and thus entitled to rights currently reserved for humans. Kurzweil’s interpretation of cryonic revival proposals is that revived people will have their bodies and brains rebuilt from new materials to repair the damage inflicted by the freezing process, creating a “new” person. Kurzweil’s views on gradual replacement of the human brain with computers and its impact on consciousness continuity are unclear. Kurzweil believes that it is only the pattern of the brain that persists and defines consciousness. [Explains why he doesn’t have such problems with uploading or with reconstructing cryonics patients.] Chapter 8: The Deeply Intertwined Promise and Peril of GNR The Internet makes it much easier to find and spread information. Research that used to take hours at a library can now be accomplished in a few minutes over the Internet. Terrorists can use the resources of the Internet for their destructive purposes. Bill Joy, founder of Sun Microsystems and inventor of Java is a critic of GNR and focuses on its dangers. Bill Joy and Eric Drexler both advocate banning self-replicating nanomachines, but pursuing many other applications of nanomachines. Kurzweil and Joy believe that gene sequences of pathogens should not be available on the Internet. Banning technological advancement as Bill McKibben advocates would require a totalitarian world government with values inimical to democracy, and would merely drive research underground where irresponsible people would wield the power of the new technologies. “Technology empowers both our creative and destructive natures.” A deadly pathogen like smallpox or anthrax, genetically modified to resist conventional treatments and then released into the population, could kill tens of millions of people in the U.S. This is an example of a potential misuse of genetic technology. However, technology has also improved our ability to cure diseases. SARS, a previously unknown virus, was genetically sequenced within 31 days of the first outbreak, allowing a

30

screening procedure for carriers to be quickly formulated. Global communication also allowed a coordinated response against the outbreak. Trillions of nanobots working together will be necessary to make the technology useful for building anything. The only practical way to generate this many nanobots is to start off with just a few and to have them reproduce. Uncontrolled self-replication resulting from a defect in the reproductive termination instruction could lead to the “gray goo” scenario. The principal structural element of nanomachines will probably be carbon because of carbon’s optimal bonding properties and versatility. Thus, an uncontrolled replication of nanobots would hurt living organisms (including humans) the most since all are carbon-based and also contain useful energy molecules and trace elements that the nanobots might also use. In a real “gray goo” scenario, the complete consumption of the Earth’s biomass would take several weeks thanks to the facts that nanomachines can’t move far due to their small size, and only the nanomachines on the “front of destruction” would be involved in self-replication—the nanomachines within the gray goo would not have access to new biomass. [What about a terrorist gray goo attack on several different locations? Also, can’t the wind move nanomachines very fast?] The ultimate nanobot weapon would come in the form of a two-stage nanobot attack in which the “seed” nanobots first spread themselves across the planet in small, undetectable concentrations, and then, at some optimal time, begin self-replicating without restraint. If the first stage were executed properly, the entire biosphere could be consumed in a 90-minute second stage. Kurzweil believes that the threat requires the creation of a nanobot immune system. Kurzweil believes that the self-replicating feature is necessary for some nanobots. Aside from the fact that it is the most efficient way to produce nanobots in useful quantities, self-replication will be necessary for immune system nanobots, for spreading intelligence beyond the Earth, and will have military applications. Safeguards against self-replication could also be defeated by a determined individual. Freitas’ “gray plankton” scenario: nanobots use carbon from methane and CO2 dissolved in the oceans. Nuclear terrorism is not an existential threat. All-out nuclear war is highly improbable. Genetically engineered viruses that are easily transmitted and highly deadly could pose an existential threat to the human race, but the window on this threat will close in the 2020’s with the advent of nanotechnology. [Part of Kurzweil’s more general view of cybernetics overtaking biotechnology] Precautionary principle: If the consequences of an action are judged to be potentially too negative, the action must not be pursued. Tantamount to technological relinquishment. Proactionary principle: Balancing the risks of action and inaction. It has been theorized that particle accelerator experiments might destabilize matter and destroy our entire planet. After all, the destructive potential of mass seems to increase as size decreases (dynamiteUraniumHydrogen isotopessubatomic particles). However, Kurzweil believes that this is not a considerable risk since the odds of exactly configuring a subatomic explosive device are extremely low, akin to building a thermonuclear weapon “by accident.” There is the possibility that our entire reality is merely a computer simulation and that it may be shut off someday, killing us all. Kurzweil is unbothered by this threat since the best possible way to avert it is to make our “simulation” interesting to our masters by achieving the Singularity. The threat of asteroid impact will be nullified by technology within the next few decades.

31

GNR cannot be abdicated out of fear of their potentially dangerous uses because the technologies also have the ability to alleviate human suffering on a global scale (ex—famine in Africa). Broad relinquishment will require a worldwide totalitarian regime. Kurzweil believes that, even if the leaders of the world had the will for such a thing, it would be impossible to impose such a system thanks to the democratization enabled by the Internet and other communication technologies. [What about Internet censorship in China? Can’t technology also lead to newer and better ways to censor communications?] Nanobots could also be used to conduct surveillance and even control people’s actions and thoughts. A government, group, or skilled individual could use the technology for these ends. Kurzweil predicts the emergence of countermeasures. [Can nanomachines be used to clean up radioactive wastes by actively transferring neutrons to cause fission in radioactive elements?—Probably not.] GNR is being advanced for benign and widely accepted reasons. The progress cannot be arrested. A Posthuman future is inevitable. Kurzweil believes that the constructive uses of GNR will predominate, but that it will be necessary for the government to be proactive in preventing misuses and to prepare defenses before disasters occur. “Should we tell the millions of people afflicted with cancer and other devastating conditions that we are canceling the development of all bioengineered treatments because there is a risk that these same technologies may someday be used for malevolent purposes?” Along with improving the human condition, economic growth is a key motivator for advancing technology. Relinquishment of one area of technology is impossible because all modern technological fields are interdependent on one another. Kurzweil and Freitas agree that some relinquishment needs to be done. Most self-replicating nanotechnology needs to be banned. A control on self-replication could be imposed through a “broadcast architecture” design in which nanobots must go to a central source to get their own design blueprints, allowing humans to control what the nanobots build and preventing any mutation of the code. For instance, in human cells a nanocomputer could replace the nucleus and mechanical ribosomes could replace their biological predecessors. The ribosomes would have to get assembly instructions from the nucleus and could not function otherwise. In turn, the nanocomputers could be remotely controlled, allowing for complete control over all gene expression. Fundamentalist environmental groups pose a serious risk to proper progress. Kurzweil cites the example of Greenpeace’s opposition to certain types of genetically modified crops which contain elevated amounts of vitamins that could save millions of children in the Third World annually from crippling developmental defects stemming from poor nutrition and which are also pest resistant and hence have lower pesticide and fertilizer requirements. While the GM crops will eventually prevail, millions of people will suffer unnecessarily for a few years because of the environmentalist opposition. Kurzweil does not believe that all GM crops are automatically safe and believes that each needs to be rigorously analyzed and tested before introduction into the environment. “Fundamentalist humanism” will also arise as a movement opposing any genetic or cybernetic upgrading of the human body. Kurzweil believes it will fail in the end. People often advocate broad relinquishment consider only the dangers that new technology will bring and fail to realize that new countermeasures will be developed from the same technology.

32

Kurzweil believes that the biotechnology sector is too tightly regulated by the government, stifling productive innovation and allowing developers of harmful biotechnology to pull ahead. He also believes that the government’s balancing of risks is skewed and cites the example of how all human gene therapy research, which has the potential to save the lives of millions of people, was halted for years after one person died during a clinical trial. Kurzweil repeatedly stresses the need to change public attitudes and restrictive government policies about biotechnology and to establish defenses against bioengineered pathogen terrorism. Kurzweil and others suggest creating a worldwide “immune system” against rogue selfreplicating nanobots. Monitoring stations would continually scan for the presence of selfreplicating nanobots, and if detected, would release their own “police nanobots” to consume the former. Kurzweil believes that the police nanobots should themselves be self-replicating to ensure that they are of sufficient number to digest the gray goo, while others believe that simply storing a whole lot of nanobots at each release point would be sufficient [what about nanobot missiles and bombs?]. All parties admit that there is a possibility of a mechanical “autoimmune response” where some of the self-replicating police nanobots themselves begin attacking the biosphere, but Kurzweil sees this as a lesser danger than having no defense against rogue nanobots at all. Analogously, while the human body’s immune system always poses the risk of initiating a deadly autoimmune response, this occurrence is extremely rare, and disabling the human immune system out of fear of such a disorder would lead to certain death within a few days. Foresight Institute guidelines for nanotechnology: -Replicators must not be able to replicate in the natural environment. -Evolution within self-replicating populations must be discouraged. -MNT devices should limit proliferation and provide traceability of any replicating systems. -Replicating nanotechnology should be restricted to agencies that agree to the guidelines. -Replication should require materials not found in the natural environment. -Replicating nanobots should only be able to produce non-replicating construction bots, which in turn do the actual building of end products. -Replication should require encrypted and time limited encryption codes that can only be obtained from a controlled source (like broadcast architecture). [Considering the expenses that would be incurred by safety protocols at nanotech assembly facilities, wouldn’t traditional means of manufacturing and agriculture prove more cost efficient for some purposes?] While the guidelines could prevent accidental replicating nanobot release, deliberate misuse is another matter. Kurzweil admits that determined people with the necessary skills could bypass the safety measures in nanobots and enable self-replication. The only option is to continually invest in defensive technologies to stay one step ahead. Kurzweil sidesteps the issue of whether or not self-improving A.I. could be made permanently friendly to humans. Decentralization provided by distributed energy, the Internet, and virtual reality will make cities obsolete and will allow people to live happily wherever they want. Pollution and vulnerability to military or terrorist attack will also diminish thanks to decentralization. A program for GNR defense: -Accelerate funding for defensive measures against genetically engineered pathogens since this is the most immediate threat

33

-Begin development of countermeasures against nanotechnology and hostile A.I. and intensify the progress as the impact of these technologies near. -Streamline bureaucratic regulatory processes for the development of new drugs -A global program for random, confidential serum monitoring to check for the development of new pathogens (apparently this is not Kurzweil’s original idea and has been around for decades but never pursued) -Well-defined moratoriums halting specific avenues of research until investigations of the risks have been completed -Agreement over the proper safety protocols regarding the use of nanotechnology -Raising public awareness of the promise and perils of GNR to gain the necessary grassroots support for political action on the issue of defensive preparations -International cooperation on all the aforementioned issues -With advanced technology, a balance must be maintained between monitoring the populace to defend against terrorist actions and preserving civil liberties. -Pathological A.I. will be the most difficult threat since, by its nature of being more intelligent that humans, pathological A.I. would be able to circumvent the safeguards of a lesser intelligence (humans). If humans gradually convert themselves into A.I., our values may be preserved. Police nanobots would have to self-replicate to keep pace with self-replicating gray goo bots. Malevolent individuals could hack immune system nanobots, which is why electronic control countermeasures are badly needed. Still, there is no viable alternative to going without an immune system [what about using lasers or some kind of energy pulses to disable nanobots?] Chapter 9: Response to critics The Criticism from Incredulity -Most people simply find it hard to believe that so much progress can be made in the 21st century, or that machines could ever become as intelligent as humans -Kurzweil reminds us that his predictions are based on extrapolating the logarithmic increases in computer power that have been consistently steady for the past 100 years. -People are often threatened by ideas that diminish the uniqueness of human beings. Copernicus’ idea that the Earth was not the center of the Universe, Darwin’s idea that humans had evolved from primates, and the Transhumanists’ idea that machines can be made more intelligent than humans thus all met resistance when they were first announced. The Criticism from Malthus -Exponential trends slow and stop growing as resources are exhausted and conditions change. -Kurzweil believes that the exponential growth of technology will eventually end, by not until well after the Singularity. -Computer chips will shrink and so will their energy requirements (reversible computing), preventing overheating. -Eventually, a minimum possible size for a computer will be reached and the only way to make computers more powerful will be to enlarge them. The Criticism from Software -While hardware power (memory storage, processing speed, communications bandwidths) is increasing exponentially, software capabilities are not. -Kurzweil argues that the reliability and capabilities of software programs stays on the cutting edge of what is allowed by the abilities of modern computers.

34

-Kurzweil claims that software price performance has been changing exponentially along with hardware price performance, and he cites speech recognition software of 1985 to that of 2004 as proof [this is a very narrow analysis]. -Software development times have been increasing exponentially [according to Kurzweil’s anecdotal claims] thanks to better technology enabling faster use of information resources. -Software size and complexity is also increasing, possibly in an exponential fashion. -Algorithmic innovation has led to new software programs that solve the same problems and perform the same functions with dramatically greater efficiency. -Kurzweil admits that software innovation is much more inconsistent that the steadily exponential increases in hardware performance. The development of a single new algorithm can surpass an entire year of progressive improvement. -Rules-based expert systems will never result in A.I. Emulation of the human brain will. -People fail to appreciate A.I. because of its narrow nature, and they always look at its shortcomings as proof that it is not truly intelligent. Only when an A.I. matches all human abilities will people suddenly believe. The Criticism from Analog Processing -Apparently, some experts believe that digital computing, due to its nature, cannot emulate the analog processing of the human brain [badly explained]. Kurzweil simply disagrees. The Criticism from the Complexity of Neural Processing -The macro- and microscopic structures of the brain are inextricably linked to function. Since every structural detail cannot be electronically reproduced, the function (sentience) cannot be reproduced. -Kurzweil disagrees and says that mathematical models have made great progress in emulating human mental functions. Further development will yield A.I. -Some will say that, due to its complexity, the brain cannot be modeled in computer code. But Kurzweil again brings up the fact that the portion of the genome coding for the brain is only a few megabytes, and that the resulting complexity stems from growth and development patterns programmed into the original code. -The brain is self-organizing, chaotic, and fractal in design, meaning that a lot of complexity can result from a relatively simple set of rules. These rules can be understood and emulated with computer software, resulting in A.I. -The brain assumes a functional structure from meaningful interactions with the environment. Order comes from chaos. -A brain’s development and functioning can be mathematically modeled and instantiated in a computer. Kurzweil admits that we are a long way from being able to do this, but it is theoretically possible. The Criticism from Microtubules and Quantum Computing -Some people argue that structural microtubules in the brain’s neurons engage in quantum computing and are also responsible for sentience. -There is debate among experts over the existence and necessity of quantum computing in the human brain. -If it were discovered that quantum computing existed in the brain, it would be a valuable insight since it would prove that quantum computing was possible. Kurzweil believes that quantum computing could be done synthetically, so if it performed any function critical to sentience, machines could be empowered with it.

35

-Due to the Uncertainty Principle, exact copies of anything cannot be made, but Kurzweil questions the necessity of ever making an exact copy of anything. Even an inexact copy of a person’s mind into a computer would still produce a being with thought patterns indistinguishable from the original. -Many estimates of human intelligence account for mental abilities through extrapolation of known neural properties and do not require the element of quantum microtubule computing for completion. The Criticism from the Church-Turing Thesis -Gödel, Turing, and Church were all mathematicians who independently concluded that there were certain mathematical problems for which no logical means of finding solutions existed. The Church-Turing Thesis is an extension of this finding and stipulates that there are limits to what artificial computation can accomplish, and that these limits to logical abilities are already known. -In consequence, adherents to the Church-Turing Thesis believe that, since there are many problems that computers cannot currently solve that humans are capable of solving, computers are incapable of intelligence. -Kurzweil believes that, in fact, humans are no more capable of solving “unsolvable” problems than machines are. All of our methods pertaining to such matters involve random guessing that is sometimes right or clearly defined processes. Both of these things can be emulated by computer algorithms. The Criticism from Failure Rates -Some experts believe that, as our computer programs have become larger and more complex, failure rates have climbed higher as well, suggesting a limit to the exponential growth in the power of computing. -Kurzweil simply disagrees that failure rates have been increasing, and again says that A.I.’s won’t have expansive program codes commensurate with their intelligence. A relatively small set of well-constructed algorithms with randomizing and fractal elements should enable the computers to become sentient. The Criticism from “Lock-In” -“Lock-in” is the phenomenon where old technologies resist displacement because considerable resources are devoted to maintaining them. -Some people cite the transportation sector as something that suffers from lock-in. -MNT manufacturing will be pursued because of its potential to produce goods anywhere for extremely low cost. Solar technology and microscopic fuel cells will be pursued once MNT makes their manufacture extremely cheap and efficient. -These advances in manufacturing, energy production, and energy storage might lead to a revolution in transportation. Affordable flying machines could be made to transport both human and material cargo without producing any pollution. The Criticism from Ontology: Can a Computer be Conscious? -Some critics believe that consciousness requires a biological substrate. -The “Chinese Room” argument shows how a computer may appear intelligent when in fact it is merely programmed to appropriately match outputs with inputs. -Kurzweil first disagrees with the idea that a nonsentient computer could convincingly emulate human speech. Any program that coded for every possible sentence and statement would be millions of pages long, meaning that responses would be very slow [what about really, really fast computers of the future?]

36

-No individual neuron or connection in the human brain is responsible for any human ability. Intelligence, speech, and all the other cognitive processes that define human are emergent properties that result from the sum of all the brain patterns. By the same token, no one part of a computer brain will be sentient. Rather, the sentience will result from the interactions of all parts. -Consciousness is a subjective experience and is incompatible with objective measurement. Determining the biological cause of human consciousness would not necessarily allow us to rule out animals lacking this property as being unconscious. The Criticism from the Rich Poor Divide -There is the fear that rich people could gain control over cutting edge technologies and deprive the poor of it. -Kurzweil points out that rich people always have access to new technology first since it is expensive initially. But over time, the price performance increases, making the technology available to poorer people (Kurzweil cites the examples of cell phones and HIV drugs). So, the poor are always a few steps behind in having access to technology, but they eventually get it. [He ignores the issue of whether a small group of people could control key technologies for their own benefit] -Kurzweil’s opinion is clearly that, despite all perceptions to the contrary, world living conditions are steadily improving thanks to technology. The Criticism from the Likelihood of Government Regulation -Government restrictions will slow or even stop technological growth. -Kurzweil again states that, historically, government intervention has had little overall effect on technology, and that only a worldwide dictatorship could truly block development. -Restrictions on embryonic stem cell research, for example, led to greater research on other types of stem cells which are moving our knowledge of biotechnology in the same direction (making totipotent cells from a person’s own DNA). -The growing use of GM crops to feed increasing populations shows how government restrictions eventually give way to necessity. -Kurzweil believes that technology will hasten social change. Decentralized technology will allow people to overcome government barriers. The Criticism from Theism -The belief that consciousness results from a soul, and that the soul cannot be reproduced in a machine. -Kurzweil counters that humans are machines, but this is not apparent since we are made of biological tissue. -The fact that the human mind can be understood mathematically does not diminish its complexity or beauty. -People are wrong to view today’s limitations on computers as inherent. More advanced computers will be capable of intelligence. The Criticism from Holism -Biological brains operate holistically, with each part interacting with the other. Computers, on the other hand, are rigidly organized hierarchies. The massively parallel interaction is what creates intelligence. Hence, computers cannot be intelligent. -Machines can be made to emulate human brain patterns once the brain patterns are understood. -Biological brains are not totally holistic just as computer brains are not totally modular. Both exist on a continuum.

37

-Emergent properties derive from patterns of interaction and make things more than just the sum of their parts. -Computers can be programmed with randomizing algorithms to mimic biological systems. Epilogue -The Singularity does not imply infinite computational abilities. We are limited by the matter in the Universe. -Kurzweil suggests that we may be able to create or invade new Universes to expand our computer abilities.

Kurzweil’s CSPAN2 interview -Moore’s Second Law will be overcome by self-organizing and self-repairing microchips that are on the verge of technical feasibility. [Makes nanoscale manufacture cheaper.] -Translating telephones, retinal beaming eyeglasses (virtual reality), 3D teleconferencing -Kurzweil believes that acceptance of the Law of Accelerating Returns is increasing among scientists as the increased pace of progress becomes more evident to them independently. His personal interactions with academics have led him to believe this. -Gradual integration of humans and machines via cybernetics simultaneous with development of A.I. -“Emotional intelligence is not a sideshow to human intelligence, it is actually the cutting edge.” [Agreement with Francis Fukuyama?] -Bionic brain implants could download new skills that the person could use -improved cost performance=deflation exponential growth=50% deflation -rate of cost performance improvement accelerating -accelerating pace of change, while not intuitive, emboldens Luddites -Like Fukuyama, Kurzweil foresees possible violent resistance to technology -“nanotechnology-based fly devices” (flying cars) in 20 years -Luddites will have their arguments weakened by the fact that it is almost impossible to establish consensus over what constitutes “normal” or “natural” human abilities since even unmodified humans differ tremendously. [Consider the weakness of Fukuyama’s arguments about “human nature”] -Potential benefit of cybernetics outweighs negatives such as electronic eavesdropping or mind control -Deflation counterbalanced by the discovery of new applications for technologies as their costperformance improves -Medical discoveries used to randomly occur, and scientists seldom understood how the new drugs worked. Today, medical technology is information-based: ends can be systematically and scientifically pursued. -Need for political policy reforms to ensure the safe and accelerated development of biotechnology. (ex—extended FDA approval process prevents fast development of antipathogenic drugs to combat new pandemics) -Writing a new book on how the human brain works

38