The Age of Spiritual Machines: When Computers Exceed Human Intelligence

Ray Kurzweil can rightfully claim the title of "prophet", because his predictions of 1998 have proven to be uncannily accurate here in 2024. While he missed some things, he also certainly made an amazing amount of correct guesses. His work is thoroughly worth reading in 2024, unlike almost all other books attempting prophecy. On the misses: he assumed that there would be continuing value to intellectual property, which now appears to be wrong. That particular miss mars some of his other guesses based on that assumption, particularly ones about structuring future economies on the continuing value of intellectual property. (Also, for the computer science types, he also seems to have missed on guesses about genetic algorithms being a key component of AI; that does not seem to have panned out, or, at least, not up to this point.) The spoiler is that he predicted the end of humanity at the hands of our AI progeny. Let's strongly hope that he got that one wrong, too. It's not a great pleasure to read this book. Ray Kurzweil is good at writing prose, but his technique of a conversation with an imaginary "Molly" over the decades of the 21st century isn't one that I liked. These imagined conversations make up a fair percentage of the page count (roughly 65 pages of the total 260 pages in the main body of the book, which is over 20%), and I personally found those wearying to wade through. His timeline in the appendix is excellent, and it is worth skipping to it first, before tackling the text. I think that this is an important book that everyone should be familiar with, particularly the dire end of MOSHs (Mostly Original Substrate Humans.) We can always hope that we escape that fate!

That the future of thinking, self-aware machines is closer in the future than one realizes is the main point of this book, which is written primarily for the curious and the optimistic, but also those interested in developments in artificial intelligence from a commercial point of view. In the book, the author unashamedly embraces a future inhabited by beings whose intelligence is much greater than what humans now possess. But also, in his future humans are not left behind, but instead stand chin-to-chin with the robust, brilliant machines that have created. Not content with being mere passive spectators whose needs are being serviced by these silicon geniuses, they become hybrids of the organic and inorganic, a combination which makes the human mind even more efficacious and powerful. The author attempts to justify his predictions starting in the first part of the book. In chapter one, called "The Law of Time and Chaos", the author attempts to show that time is changing its "speed". The examples he gives are not really convincing from a scientific point of view, and he does not give any recommendations for any empirical tests of his assertions. And even though his arguments are plausible, one could deny them and still believe in the predictions he makes later in the book. His view of artificial intelligence is in terms of neural networks and genetic algorithms, with the former being around for some time now, and the later making its presence known rapidly. There is no doubt, and the author addresses this many times, that any type of machine intelligence will have to be able to mimic the creativity of the human mind and its ability to make unexpected connections with concepts that can appear at first glance to be completely unrelated. The author gives though convincing arguments as to the ability of even current computers to perform acts of creativity and automatic knowledge acquisition and discrimination, such as in the arts. The author discusses many strategies for building this new form of intelligence, including optical technology, nanotubes, and DNA computing. The later however will probably not be the way to go, as astronomical quantities are needed to perform some of the feats of calculation, such as the traveling salesman problem, which is mentioned by the author in the book as one that can be tackled using DNA computing. Quantum computing however is also mentioned in the book, and this approach holds much promise, and will be extremely powerful if realized. Current strategies in quantum computing are usually based on the experimentally unjustified notion of entanglement, but it looks hopeful that other ones will be developed that will bring out this quantum coup de etat in programming power. The social ramifications of the new age of intelligent machines is not ignored however in the book. He predicts, with ample justification I think, a growing "Luddite" challenge to the explosion of technology. Thankfully though, the author remains optimistic, and predicts events and technologies that would make today's computer scientists drool with anticipation. All of this is going to happen according to the author, within the next 30 years, and he gives a dramatic overview of the hypothetical decades ahead. Those readers basking in the current optimism, which can with confidence be called a technological Enlightment, will read these pages with great excitement and a longing to be part of what is ahead. Sure, the book is only a prediction, but lets keep our fingers crossed.........

The argument here starts with the observation that intelligent machines are already a reality. In the future we will be able to create ever more intelligent machines. A truly explosive increase in machine intelligence is described, up to unimaginable levels only 100 years from now. As far as hardware is concerned, the author insists that Moore's law, according to which computing power at a given cost doubles every 18 months, will continue unabated. Current computer technology will hit fundamental physical constrains in about 10 years, but the author insists that intelligence will find a way to grow explosively without slowing down because of mere physics. He describes several potential avenues through which computer power can increase. The most intriguing one is the possibility of building quantum computers. According to theory these can have immense computing power, because quantum phenomena allow for a huge number of calculations to be done in parallel. Nature does never give something for nothing, and, for my taste, quantum computers come too close to the edge - but who knows? More interesting is the question of software. When I was a student I was taught that hardware power is not very significant for defining the reach of computers, because algorithmic complexity almost always grows exponentially when related to the size of the problem to be solved. So fundamental advances will come from better algorithms, not from increased hardware power. The current state of affairs seems to agree with this view. After all, word processing 20 years ago using an Apple II is not fundamentally different from word processing today using computers a thousand times more powerful. Algorithmic design is a very costly process that few humans can do well, and this fact, it seems, would necessarily frustrate the rapid increase in machine intelligence. The author produces two solutions: One, to dissect a human brain and copy its "algorithm". To me this sounds rather similar to Da Vinci's idea of dissecting birds in order to build a flying machine. The second solution refers to methods where the algorithm builds itself, i.e. learns how to work. Two methodologies are mentioned here, genetic algorithms and neural nets. Many examples are given about the successful use of these to solve surprisingly difficult problems. These are powerful ideas. Certainly if we come to a point where an intelligent machine can design and maybe build an even more intelligent machine, then indeed we have the ingredients for a runaway explosion of intelligence. The author paints a rather scary near future. First we use machines to broaden our own minds, but we also build independent machines that become more and more powerful. The war between conscious machines and humans never happens, because it is won by the machines before it even starts: Human minds migrate into machines and lead a fruitful life within the more expansive means that this new medium offers. A few biological humans remain at a much lower level of intelligence, and presumably wilt away without remorse or bitterness. The next level of evolution is achieved by beings that are unimaginably more intelligent than we are now. At this stage it is not meaningful to talk about machines, they are just our offspring living in bodies that are not DNA based. In fact these immortal beings include persons originally born as humans, including, it is stated, Microsoft's Bill Gates. Again, all of this will happen in the next 100 years. Now, what is wrong in this picture, apart from the idea of having Bill Gates around for all eternity? For starters, for this vision to work the whole world should be like California, which it is not. On most places of this planet misery rather than intelligence rules. The book was written before the techno-bubble burst, an event not predicted in the book, and also before the recent power failures in California itself. The earth is in such bad working order that one must wonder if ever intelligence will take hold. Also, if the explosion of intelligence is a necessary natural phenomenon, something like an unstoppable supernova, then by now the universe should be full of manifestations of intelligence, it should be infused, drenched, saturated by intelligence. Why, entire suns would be cloaked in order to harvest their energy needed for the huge amount of computations going on. (Maybe this explains the mystery of the dark matter, i.e. the fact that most matter in the universe is not visible.) Also, why hasn't this cosmic intelligent fabric reached us? Well, maybe they wanted to leave part of the universe in its natural state, you know like we keep protected nature parks. Our corner of the universe may be just such a place. We do not listen to the noise created by intelligence elsewhere, but maybe this only shows our low level of development. Now, if we assume that the universe is not really intelligent, this leaves us with three possibilities: First, for some reason there may exist intrinsic limits to the growth of intelligence. This does not seem probable, particularly after reading this book. Second, depressingly, the explosive growth of intelligence may always bring about its demise, the same way that the uncontrolled growth of cancer kills the organism that creates it. This sounds rather plausible if we observe the way humanity manages the world today. The third possibility is that the growth of intelligence reaches a level where it decides to stop, where it understands that the "law of accelerating returns" is not conducive to happiness. So maybe the universe is filled not only by intelligent, but also by wise races, that look after their own small garden without disturbing the rest of creation, and live meaningful, biological, limited lives. All in all, this is an extremely interesting book to read. I withhold one star, only because I find that the catchy title has little relevance to its content.

Written in 1999, this book is very comprehensive and generally right in its predictions. Technology has advanced substantially, has reached many middle-class families here and abroad, and computer programs that simulate thinking and planning are being used effectively by corporations and governments. One area where the book may have been less accurate is that Kurzweil's idea of AI being one computer or a group of computers that has human-like intelligence or superintelligence appears to have been displaced by networking of people, universities, corporations, and government departments (social media and the Cloud). The Cloud would seem to be more or less something like the Singularity he predicts. A good summary of Kurzweil's and AI researcher ideas can be seen in the documentary film "Transcendent Man". Two other recent topics are worth noting: energy and public health. While there may be some exploration of energy sources for large-scale AI in his other books and in the film, this title does not explain how much solar energy, nuclear reactors, or hydropower would be required for global scale AI networks. If the energy system is not sustainable the AI dream would be quite limited. On public health, specifically disease outbreaks, it is now clear in 2020 that highly infectious diseases like the coronavirus are a challenge that government can respond better to when government uses internet technology, databases, Cloud computing, and "AI", not to mention high-tech manufacturing. The internet pioneers were optimists about technology's future. Ultimately, we invented and developed modern technology to improve the quality of life. Today, as technology has grown more powerful our choices about how to use technology are more consequential. The knowledge and experience of our leaders and the fairness and decency of leaders overseas will be tested severely by this newly found power.

I find this book thoroughly interesting. As a mathematician, I have found many books on the subject of AI lacking in content. This one surprised me. I am at heart disturbed and somewhat reluctant to assent to the idea of making a thinking, feeling machine. This is what so many engineers and scientists purport to be working toward. I feel this attempt is wrongheaded. It is anthropocentric. Why recreate human consciousness in a machine? Why not start the process and allow the machine to become whatever the processes of evolution dictate. It is the age old attempt by man to become God, (w/o sex) in my view. Be that as it may, Kurzweil reverses the notion and predicts not a Brain in Vat idea but that we'll become the intelligent machine by the end of this century. Why make a 2001 Hal, when we can become Hal. One thing is good for sure, all of those moralizing philosophers can shut up with their worries about a Matrix scenario, because we'll be the matrix if Kurzie is right? And I remind you he has before. He predicted speech to text recognition software by 2009 and now every time I check my msgs I get a written transcription (not always correct but pretty good). So, he is not far form right to predict the digitizing of our minds. And I like the idea too! Screw that ill-thought up Turing Test, to see if an AI is a conscious, emoting entity---let's become digital. Now, let's note the irritating flaws to the book. First, he cow-tows to women by making his chosen gender of reference 'she' and 'her'. Don't know if Kurzie is hen-pecked or just trying to be non-sexist but it made me wanna scream! I can just see his wife saying: Ray now ya did remember to use her instead of he when writing hypothetically right? Yes dear, I did. Look! That's good, because She in the Sky wouldn't like that. Okay, I'm ribbing him, but he should understand English is a masculine language and take he for the general gender. He could've use 'one' but that would have become really esoteric. like poetry, okay minor flaw, but an annoying one. He seems to believe to make the complex and abstruse topics he covers more amenable to the non-tech audience, he must have a dialog with this fictional Molly (whom is suppose to be us the readers, disaffecting to any male reader), who asks all the dumb laymen (or should I say laywomen) questions. I've rail against this let's-explain-it-to-the-dummies approach in every review I've made of books about scientific topics directed at laymen. Kurzie is no exception. The little dialogue sections began to get me so I skipped them, suggest others do too. I think he was trying to copy Hofstadter's style in his famous Goedel, Escher and Bach opus. Also, he packs a little too much into this work. Those grey sideline stories distract from the subject at hand, though they are interesting. He also, strays afar sometimes without credentials in some discusses. for instance the Big Bang theory and Entropy. A tangent he is not really qualified to discuss, though he has the right to theorize in whatever field he pleases, it only distracts the reader. He did for me I started researching it and stopped reading him for about a week. But, all in all, this work is fresh and though it's a decade old, it hits the intellectual spot. BTW, you can now check all the predictions he makes in the appendix. He made a lot of them, some for right around now: circa 2010 to 2020.

This books presents nothing more than reasonable and logical future developments of our current technology. And that is mind blowing enough to surprise anyone who do not spent some time wondering what the future will be for us and the sentient beings we will create. What really surprises me is that in most exercises of futurism that politicians, economists and philosophers like to make every now and then, they completely ignore the main role that AI and Cybernetics will have. If they don't start paying atention to this soon, they will have a very rough surprise in just a few decades. This book is a perfect introduction for a future that will be somewhat alien for some. But in this alien future there will be no T2 or HAL9000 scenarios coming to get us. Anyone who think this can actually happen, completely ignores that a intelligent machine will not have the evolutionary urges that makes us, humans, do many of the horrid things we do. They will not have the "need for power" that makes humans commit crimes like wars and genocide. The future much more probably will be a merge of carbon based intelligences and silicon ones. This book explains how that will probably happens. Maybe Mr. Kurzweil is wrong by 10 or 20 years, but these things will happen because they're the only logical extrapolation from technology and events ocurring today. About the people who think that this book is full of nonsense, I can only think that they believe that there is something magical hapenning inside our heads (maybe we all have little Harry Potters inside our skulls =), something that cannot be reproduced by science. THAT'S my friends is the real "nonsense". You don't even have to be an atheist to believe that our minds are matter based entities that obey the laws of physics. To the other critics,these ones more science-minded individuals, who too think this book is full of "nonsense" (an undefensable position because as I said you'll find nothing besides science and logic backing Mr. Kurzweil predictions), I remember the quote from Mr. Clarke: "If a scientist says that something is impossible, he's almost always wrong"

I read this like seven or eight years ago and go back once a year to see how Kurzweil did on his predictions for the future with his historical timeline of technology. And folks, he's been spot on. And when you look at his predictions for the twenty years or so since he wrote this book, he's been spot on. And then when you consider the things he's predicted for the end of the century, it's absolutely mind-blowing. I mean this guy said twenty some years ago in writing that a super computer the size of a credit card, likely a telephone, would be affordable and in everybody's hands. You will want to know what he's predicted for the next twenty years I think. And under the heading of a cause and effect loop, some financial investors have payed close attention to Kurzweil's time line too. Now I'm not an engineer or much of a math guy and there's a lot of that stuff in this book, but Kurzweil is also a readable and prolific philosopher, and whether you agree with his philosophy or not, did I mention the guy hasn't been wrong in his predictions yet, you will take pause to rethink your point of view for better or worse. What else could you want in a book?

Kurzweil is a tech savant, respected at the highest institutions. It helps to know a little physics, but his prose is clear and easy to understand. Suggest you familiarize yourself with Heisenberg's Uncertainty Principle, IE, at the quantum level, matter only comes into being when something Conscious looks at it, collapsing from a wave function into particles. It means that consciousness is Non-local, it exists apart from the physical universe, and affects the matter it observes. Bottom line, we humans are consciously inventing the next step in our evolution. The exponential rate of computing power indicates we will achieve Artificial Intelligence soon; Self-aware, conscious computers, operating at light speed (as opposed to slow, neuron-based organic brains). Being conscious entities, they will inherit curiosity and creation. They will have identity and existential/spiritual questions, and will likely go looking for themselves. We humans are already implanting artificial joints, pumps, microchips and medical devices in our bodies; expect brain implants tomorrow. Over time, and to stay relevant, humanity will inevitably need to abandon the ages-long pace of natural evolution and "move into" our technology. The switch from limited neuron-based thinking to photon-speed intelligence is a paradigm shift for our species--bringing exponential advancement. Thus, the coming spiritual machines will be our self-evolved selves.

Quick delivery. Good product.

Tout en étant technique sur des sujets parfois difficile c'est avant tout d'une grande clarté et d'une facilité incroyable. Parfois un peu trop argumenté et donc répétitif il se lit comme un roman et fait beaucoup réfléchir. Les implications sont énormes et font parfois froid dasn le dos. On hésite à en parler car on pourrait passer pour un fou furieux...

It's a great reading even it's a book published for so many years! It's so mind blowing and inspiring! Recommended!

遠くない将来に機械（コンピュータ）は人間を超える知性を獲得し機能別共生も進む、という基本テーマを巡って、歴史的展開・展望を含め様々な視点・観点から本質に切り込んでいて興味深く読めた。各章末には、読者が抱くような素朴な疑問を取り上げ、仮想対話（自己問答）の形式で著者の考えを敷衍しており、また巻末のGLOSSARY、NOTES、SUGGESTED READINGS、WEB LINKSも充実していて読者にはありがたい。ただし、精神的機械との関連で、人間の意識に触れているくだりでは、Chalmersのいわゆるハード・プロブレムのきわみとしての「そもそも意識の主体とは？ なぜ・どのように意識は生じるのか？」 という点について、博識・博学で知られる著者の考えを披露して欲しかった。また知性だけでなく、もっと根源的な「命（いのち）」という視点からも、機械と人間の比較に踏み込んで欲しかった。しかし、それは過大な期待として差し引くとして、本書は当該分野における秀逸の書物と思う。

Great book which made me think and take a step back. A new perspective on tech and society is refreshing. An opinion more nuanced and educated than « society is trash, tech is bad, future is screwed » that we usually hear and see in sci fi movies.