The Singularity Is Near

From Wikipedia, the free encyclopedia

Jump to: navigation, search
Cover of the book
Cover of the book

The Singularity Is Near: When Humans Transcend Biology (Viking Penguin, ISBN 0-670-03384-7) is a 2005 update of Raymond Kurzweil's 1999 book, The Age of Spiritual Machines and his 1987 book The Age of Intelligent Machines. In it, as in the two previous versions, Kurzweil attempts to give us a glimpse into what awaits us in the near future. His reasoning rests on the combination of four postulates:

  1. That a technological-evolutionary point known as "the singularity" exists as an achievable goal for humanity (the exact nature of the point is an arbitrarily high level of technology).
  2. That through a law of accelerating returns, technology is progressing toward the singularity at an exponential rate.
  3. That the functionality of the human brain is quantifiable in terms of technology that we can build in the near future.
  4. That medical advancements could keep a significant number of his generation (Baby Boomers) alive long enough for the exponential growth of technology to intersect and surpass the processing of the human brain.

Contents

[edit] Postulates

All four of Kurzweil's primary postulates must be correct in order for his conclusion to be true.

Acceptance and striving for the idea of living forever

Kurzweil asserts that a technological-evolutionary jump known as "the singularity" will be recognized as an achievable goal by humanity. Kurzweil acknowledges that belief in the singularity promotes the paradigm shift necessary for its advancement. By promoting the "truth" of its coming through predictions that seem remarkable at the time but inevitable after the fact (a global computer network, a computer beating the chess champion, etc.), Kurzweil's popular series of books reinforce the belief that a singularity is unavoidable.

The law of accelerating returns

Kurzweil asserts in his Law of Accelerating Returns that technology is progressing toward the Singularity at an exponential rate, relying almost entirely on empirical data. He expands on Moore's Law with models showing that not only the return, but the rate of return is increasing exponentially.

An objective measurement of cerebral processing power

Kurzweil asserts that the functionality of the brain is quantifiable in terms of technology that we can build in the near future. Kurzweil's earlier books showed cerebral processing power as primarily the number of computations in a square inch multiplied by the area of the brain. In this update, however, he finally acknowledges the possibility of Penrose-Hameroff Microtubule quantum processing (Orch-OR) and states that if his calculations of the processing capability of the brain are off by a factor of a billion, the double-exponential growth of technology will still catch up to it twenty-four years after his original projections. The Orch-OR theory is generally discredited among neuroscientists. In a rebuttal paper at KurzweilAI.net, Hameroff asserts that the quantum processing power required for consciousness is at an order of magnitude greater than what can be expressed through conventional systems of processing measurement - an argument that seems again to ignore Kurzweil's premise that accelerating returns in development of present technologies (as well as the inevitable paradigm shift by advent of another periodically) could ameliorate even such a barrier, a relevant example being more advanced a form of quantum computing capable of full neural emulation, at a scale and with accuracy of computation equivalent to the biological human brain.

Sufficient medical advancements

Kurzweil asserts that medical advancements will keep his generation alive long enough for the exponential growth of technology to intersect and surpass the processing of the human brain. Kurzweil explains how nanobots will eventually be able to repair and replace any part of the body that wears out, but relies on other methods of medical technology to prolong our lives long enough to reach the singularity. The usefulness of this medical postulate then becomes a function of how long it will take to reach the singularity, something that has been thrown into question due to the possibility of quantum brain processing in many recent books, such as Roger Penrose's The Road to Reality: A Complete Guide to the Laws of the Universe.

[edit] Organization

  • Prologue: The Power of Ideas
  • Chapter 1: The Six Epochs
  • Chapter 2: A Theory of Technology Evolution
  • Chapter 3: Achieving the Computational Capacity of the Human Brain
  • Chapter 4: Achieving the Software of Human Intelligence
  • Chapter 5: GNR
  • Chapter 6: The Impact
  • Chapter 7: Ich bin ein Singularitarian
  • Chapter 8: The Deeply Intertwined Promise and Peril of GNR
  • Chapter 9: Response to critics
  • Epilogue

[edit] Chapter One: The Six Epochs

Kurzweil first defines the Singularity as a point in the future when technological advances begin to happen so rapidly that normal humans cannot keep pace, and are "cut out of the loop." Kurzweil emphasizes that this will have a profound, disruptive effect on human societies and on everyday life, and will mark the end of human history as we know it. In place of normal humans, Strong Artificial Intelligences and cybernetically augmented humans will become the dominant forms of sentient life on the Earth. The Singularity will be initiated once self-improving Artificial Intelligences (sentient, highly intelligent computer programs capable of quickly redesigning themselves and their offspring to improve performance) come into existence.

Underlying all of Kurzweil's ideas regarding the progress of technology and the Singularity is the Law of Accelerating Returns. The Law states that technological progress occurs exponentially instead of linearly, meaning that each new advancement enables several higher advancements instead of just one higher advancement, and concordantly, every year, more useful inventions and discoveries are made than were made in the last.

The Law of Accelerating Returns has a very important consequence in that extrapolation of exponentially improving technology trends into the future suggests, by Kurzweil's analysis, that highly advanced technologies will arrive far sooner than linear-thinking people assume. The creation of the modern Internet and the completion of the Human Genome Project are prominent examples illustrative of this point. Both were multi-year projects that relied on computer technology to reach completion. In both cases, critics derided them as hopeless since, in the beginning, both relied on computers that would have taken decades to process all of the necessary data. However, these critics had failed to take into account the exponentially improving nature of computer processing speeds and price-performance, and thus failed to see that, within a few years, the two projects would have access to vastly superior computers that would drastically shorten their timelines for completion.

Kurzweil understands that a linear view of historical progress and of anticipated future change is instinctive to the average human mind, but insists that it is wrong.

Kurzweil lays forth his idea that the fate of the universe is to progress through six different epochs, characterized by the major paradigm shift that takes place after each one. The universe has already passed through four of these epochs and we are entering epoch five. Each step indirectly produces the next through chemical-, biological-, and technological Evolution (respectively). The Six Epochs are subject to the Law of Accelerating Returns, which states that each transition occurs more rapidly than the last. Kurzweil supports this final postulate with logarithmic graphs of the chronology of important events in the history of the Universe (i.e. - the Big Bang, the origin of life, the birth of the human race, the creation of the first computer).


Epoch 1. Physics and Chemistry

This epoch starts at the beginning of the universe (for example, the Big Bang). In this epoch, information is mostly held in subatomic structures such as particles and atoms. That is, the most complicated stable objects in the universe do not exceed the molecular scale in size or complexity.

Epoch 2. Biology and DNA

This epoch starts with the beginning of life on Earth (or elsewhere), suddenly giving rise to more complicated, yet stable, organisms that are capable of growth and self-sustainment. However, in this epoch organisms do not change within their lifetimes as evolution takes thousands of generations. Moreoever, in this stage, genetic information is stored in DNA molecules.

Epoch 3. Brains

The evolution of life gradually produced more and more complex organisms, necessitating the need for fast central control and thus giving rise to the evolution of brains. With brains, organisms can now change their behavior dynamically to suit changes in the environment and can also learn from past experiences. Evolutionary information is, in this stage, stored in neural patterns, with the meme emerging as the primary information unit.

Epoch 4. Technology

Evolution of brains culminates with the evolution of humans, which possess the ability to create technology. In this stage, technological designs are also subject to evolution and information is held in hardware and software designs.

Epoch 5. The Merger of Human Technology with Human Intelligence

This epoch, which Kurzweil argues we are in the process of entering, is where technology reaches a level of sophistication and fine-structuring comparable with that of biology, allowing the two to merge to create higher forms of life and intelligence.

Epoch 6. The Universe Wakes Up

After mastering the methods of technology and biology, Kurzweil predicts that human/machine civilization will expand its frontiers into the universe, gradually (or perhaps explosively) consuming the contents of the cosmos until the universe reaches a 'saturated' state where all inanimate matter has been converted to substrates for computation and intelligence, and a truly universal super-intelligence takes form.


Next, through logical deduction, Kurzweil briefly lays out a number of postulates and consequences in the first chapter as they relate to the Singularity. As stated, the Law of Accelerating Returns is real and implies exponential- and non-linear growth in technological capabilities and human knowledge. Information technologies are experiencing superexponential growth since human economic investment in them increases as they become more powerful.

Human brain-scanning techniques are reliant upon computers and advanced machines, which are all subject to the Law of Accelerating Returns. Therefore, our ability to understand the way the human brain works is improving tremendously every year. Based on the current trends, brain scans should give us a sound understanding of how the human brain works by the mid-2020's. Using that information, scientists should be able to create simulated human brains inside of computers, leading to the first Artificial Intelligence (a thinking computer capable of passing the Turing Test) by 2029. The computer hardware of that era should also be powerful enough to meet the cognitive needs of a simulated human mind.

Artificial Intelligence would inevitably prove superior to human intelligence since the former would combine the fast speeds, memory capacity and recall, and instant downloading/learning abilities of computers with the creativity and pattern-recognition abilities of the human mind, effectively constituting a being with the strengths of both and the weaknesses of neither. Artificial Intelligences would also have the ability to edit their own software at will to instantly improve themselves, which is an ability humans naturally lack and can only approximate through laborious genetic engineering techniques.

Biology possesses inherent and insurmountable limitations that originate at the molecular level. These limitations handicap both potential mental abilities and physical abilities. This reality will inevitably lead to the creation of synthetic replacement parts that are infinitely more durable and capable than natural human organs, limbs and cells. Some humans will choose to incorporate these engineered prostheses into their bodies, becoming cybernetically enhanced.

Nanotechnology will play a major role in human brain augmentation with the advent of brain nanomachines designed specifically for interacting with human neurons. Brain cybernetics of this sort would allow human users to vastly expand their cognitive abilities, to experience full-immersion virtual reality, and to directly interface with computers and other mentally augmented humans. Being naturally more powerful, the cybernetic portion of augmented human minds will come to predominate.

A very broad range of advanced nanomachines will also be created, allowing the humans and machines of the future to alter their world with incredible power. Aside from the aforementioned brain augmentation role, nanotechnology could be used for a variety of tasks including a total repair of the Earth's environment and the instantaneous manipulation of physical objects via "foglets" so as to blur the distinction between "real" and "virtual" reality.

Kurzweil foresees the computers of the future using three-dimensional computer chips composed of nanotubes. This architecture will allow them to operate at terahertz speeds. Computer advances like this will eventually make A.I.'s so powerful that they will completely take over the cutting edge of all scientific research and development, and will generate new advancements (including improvements to their own programming) so quickly that normal humans will not be able to follow what is going on. By this point, the Singularity will definitely have been reached.

In the far future, A.I.'s and cybernetic humans will live almost exclusively in full-immersion virtual reality worlds, which will themselves be contained in advanced computers. Computers will grow ever more advanced, but at some point, the bottom limit to transistor size as defined by the laws of Physics will be reached, and performance improvements will only be possible through the creation of new computers or the expansion of existing ones. More and more inanimate matter on the Earth will be restructured to form useful computer substrate, but it will eventually prove insufficient, and the future machine race will radiate out from the Earth in all directions, "saturating" the entire universe with "intelligence" by converting unstructured, inanimate "dumb" matter (i.e. - Moon rocks, dead gas giant planets, meteoroids) into structured "smart" matter that lives in the sense that it supports thinking, feeling A.I.'s. Entire celestial bodies of countless number across the universe will be totally converted into computer substrate. The length of time it will take to finish the task of "waking up" the universe depends heavily on whether the speed of light can be surpassed or circumvented--the maximum rate of travel of course limiting the speed at which the advanced Earth civilization can spread across space to new locations.

Kurzweil concludes the chapter by stating his belief that, while A.I.'s will inevitably prove vastly superior to humans in every way, he expects them to respect human life and to embody human values.

[edit] Chapter Two: A Theory of Technology Evolution

Each stage of biological evolution and technological innovation increases the level of biological or technological order and enables entry into the next stage. Order in this case is defined as how well the form (genetic or physical) suits the given function. For example, a modern car is better suited for transporting human occupants than a Model T because the newer vehicle possesses numerous design improvements the older one lacks.

An increase in order also generally correlates with an increase in complexity. Getting back to the previous example, a modern car has more moving parts than a Model T, and the newer car's parts are also more precisely and elaborately made. Thus, the modern car is more complex. The overall trend towards increasing complexity can also be observed in biology, where genomes of living species have gradually grown in size along with levels of physical diversity and specialization.

Of course, greater order does not necessarily always entail greater complexity. Sometimes, the simpler solution proves superior both in biology and technology, but the general trends in both have been towards more complexity and more order.

The Law of Accelerating Returns states that biological and technological evolution utilize positive feedback, with each improvement building upon the last and enabling the next.

Technological progress in any field consists of a series of "paradigms"--particular methods used to solve certain problems. A given example of a paradigm would be the shrinking of computer transistors to make the computers more powerful. While there are innumerable technological paradigms, all share the same basic life cycles. The advances in the cost-performance of a particular technology, or the Life Cycle of the Paradigm, if graphed, will appear as a sigmoidal S-shape with three distinct phases.

  • Slow Growth Phase

At this point, the kinks in the technology are still being worked out, and a it is still struggling the establish a market base. Growth in price-performance and capabilities is exponential, but still at such an early stage that the growth appears deceptively flat and linear.

  • Rapid Growth Phase

Begins after the exponential growth passes the "knee of the curve" and explosive growth in the technology's capabilities and user base starts.

  • Leveling Off Phase

The technology matures as scientists find it increasingly more difficult to make improvements to the same technology in an effort to further address the original need. Growth in usefulness levels off.

Once a technology has reached maturity, it is replaced by a newer, totally different technology, meaning a paradigm shift occurs. This occurred during the 1960's when scientists found it impractical to shrink computer vacuum tubes any smaller and instead switched to transistors, which were newer and allowed the process of miniaturization to continue.

This process of periodic exponential growth parallels biological evolution in two ways. First, biological evolution also occurs in spurts (Theory of Punctuated Evolution). Second, some biological innovations make organisms exponentially better or speed up evolution from that point onwards in an exponential manner. For example, the advent of DNA allowed life forms to evolve much higher levels of complexity and order.

Society's acceptance of new technologies is speeding up exponentially

  • U.S. Phone company revenues and daily number of American phone calls (increasing exponentially)
  • Number of U.S. cell phone subscribers (increasing exponentially)
  • Time until major new inventions reach mass use by American consumers (decreasing exponentially)

Technologies experiencing exponential changes

  • Dynamic RAM size (smallest feature sizes decreasing exponentially)
  • Dynamic RAM price performance (improving exponentially)
  • Average Transistor price (decreasing exponentially)
  • Transistor Manufacturing costs (decreasing exponentially)
  • Microprocessor clock speeds (increasing exponentially)
  • Microprocessor costs (decreasing exponentially)
  • Transistors per microprocessor (increasing exponentially)
  • Processor performance (increasing exponentially)
  • DNA sequencing costs per base pair (decreasing exponentially)
  • Random Access Memory bits per dollar (increasing exponentially)
  • Magnetic data storage bits per dollar (increasing exponentially)
  • Wireless Internet and phone services price performance (increasing exponentially)
  • Number of Internet hosts (increasing exponentially)
  • Bytes of Internet traffic (increasing exponentially)
  • Internet backbone bandwidth (increasing in a very terraced, quasi-exponential manner)
  • Mechanical device sizes (decreasing exponentially)
  • Number of scientific citations for nanotechnology research (increasing exponentially)
  • Number of U.S. nanotech patents (increasing exponentially)

While Gordon Moore first observed in 1965 that the transistor densities of integrated circuits were doubling every two years, an extended analysis shows that computers have been experiencing exponential improvements to their cost-performance (maximum number of calculations per second per $1,000) since at least 1900, when the very first electromechanical computers were invented. This trend in increasing performance has held steady across five computer paradigm shifts (electromechanical, relay-based, vacuum tube, transistor, and integrated circuit) and is now encapsulated by Moore's Law. While integrated circuits will--like all paradigms--ultimately reach the limits of their possible capabilities, the exponentially growing performance trend will likely continue via a paradigm shift to a newer technology like three-dimensional molecular computing.

Extrapolating this rate of improvement, supercomputers should be capable of the same number of calculations per second as a human brain by 2010, and personal computers should be at this level around the year 2020.

[edit] See also

[edit] External links

Retrieved from "http://en.wikipedia.org/wiki/The_Singularity_Is_Near"
Personal tools