The Singularity is Near deals with many of the same concepts as Kurzweil’s previous books but attempts to be a more thorough exploration of the idea.
It begins with a section called “The 6 epochs” which is a very long summary of history from the big bang to the evolution of life on earth to the dawn of man and the evolution of human technology. One interesting fact is that for roughly half of the time since life evolved the only form of life was single celled organisms.
The next section is a theory of evolutionary processes. This explains the mechanism Kurzweil believes is behind exponentially accelerating technological progress. To put it briefly he believes that it occurs because at each stage we are using the tools available to us to generate the next set of tools. As the tools get better the progress accelerates building on itself and growing faster and faster. To take extreme examples its easy to see why Google and Microsoft Excel speed technological progress and how a lack of basic tools like written language would slow it to a near halt.
There are separate sections on achieving the computational capacity of the human brain and achieving the software of human intelligence. Kurzweil estimates the computational power of the human brain and how long it will take computers to match this. One way in which these estimates is done is by looking at sections of the brain that are well understood and determining the equivalent amount of computer power necessary to achieve this on computer. If Kurzweil or even mainstream futurists are even remotely correct we will have no trouble achieving the computational power of the human brain. In fact in Kurzweil’s previous book, The Age of Spiritual Machines, Kurzweil predicts that by 2020 the computational power of the human brain will come at a cost of only $1000.
As I write this there are likely no computers in existence that have the hardware capacity of a human brain. The idea that that much computational power could be available so cheaply makes Kurzweil’s predictions about when we will have the software of A.I. seem almost conservative. I would imagine there will be an enormous economic incentive for businesses spending $30,000 a year on employees to create the software to replicate their jobs in $1000 machines. The incentive to create A.I. then cannot be overstated and given the hardware we have now there is not much economic incentive to do so, so we aren’t seeing a big push towards developing software yet. As I.J. Good said “the first ultraintelligent machine is the last invention that man need ever make”.
But creating A.I. is only the beginning. In one part of this book Kurzweil attempts to ascertain what the power of a computer would be if engineers could design it to the absolute limits set by physics. He calls this the theoretical computer the “Ultimate Cold Laptop” and the conclusions he reaches are staggering. According to Kurzweil’s estimate “the optimal computational capacity of a one-liter, one kilogram computer at around 10^42 cps, which is sufficient to perform the equivalent of ten thousand years of thinking of ten billion human brains in ten microseconds.” Our brains are roughly 3 pounds of matter powered by a small amount of energy. Science already tells us that the universe is staggeringly big and will last a staggeringly long time. The potential here for intelligent thought and experience is beyond comprehension. Simulations of every imaginable scenario would be possible.
Aritificial Intelligence skeptics’ more common criticism for the idea that A.I. is soon at hand is that even though we will reach the hardware capacity of the human brain we will not be able to create software. Kurzweil believes this issue can be addressed by reverse engineering the human brain. He describes the exponential progress in the ability of brain scanners to understand sections of the human brain.
One section in this book deals with the three approaching revolutionary technological paradigms of genetics, robotics and nanotechnology. Each of these new technologies could radically alter the world for the better or destroy it. Kurzweil believes we are in the beginning of the genetics revolution as usher in by the sequencing of the human genome. Many dystopian science fiction writers have written of the possibility of a future in which we can create designer babies; Gattaca is one example. Kurzweil’s vision is different, he writes of a potential for designer baby boomers in which adults alter their genes. Nanotechnology and Artificial intelligence will come later and have the potential to be even more powerful than genetics.
The only thing I disliked about the book is that that its results can come across a little too much like they are inevitable in places. Also Kurzweil risks sounding like a cult leader when he says he would like to label people who believe that the singularity will occur “singularitarians.” This language turned off a friend of mine who started this book but didn’t finish it. The book can seem a bit like a sacred tome when it discusses the “epochs” that the universe will go through. However, while this tone is sure to put some people off I do think it is forgivable. The implications of the singularity are so fantastic that it can’t help but step on the toes of traditional with religion at times. No matter how concepts such as eternal life and superhuman intelligence are presented they are going to be offensive to some.
I loved the way this book carefully deals with objections to the theory that “The Singularity is Near.” There are several legitimate criticisms that can be made for Kurzweil’s timeline and he deals with them honestly. Some of the criticisms include the criticism from Malthus, the criticism from incredulity and the criticism from software.
This book is a must read for anyone interested in the singularity. While most people probably consider it the definitive work on the singularity I’d still recommend starting out with Robot Mere Machine to Transcendent Mind or The Age of Spiritual Machines or even Kurzweil’s essay on the law of accelerating returns. The concept itself of the singularity is not overly complex and this book might seem daunting to someone who is not already sold on the idea.