The Singularity is Near: When Humans Transcend Biologyby Published 26 Sep 2006
|The Singularity is Near: When Humans Transcend Biology.pdf|
For over three decades, Ray Kurzweil has been one of the most respected and provocative advocates of the role of technology in our future. In his classic The Age of Spiritual Machines, he argued that computers would soon rival the full range of human intelligence at its best. Now he examines the next step in this inexorable evolutionary process: the union of human and machine, in which the knowledge and skills embedded in our brains will be combined with the vastly greater capacity, speed, and knowledge-sharing ability of our creations.
"The Singularity is Near: When Humans Transcend Biology" Reviews
Can this book ever get to the point? Is there a point? In the future, when machines begin to express human discernment and burn books, I'm sure this endless and gigantic tome of wordy lists and nerd-spooge will be set alight, or edited towards readability. Either is fine with me. I would love to read the executive summary of this, but this book is too long.
This starts with the thesis: Technological change is exponential!
This has been true for many measures such as micro-processor size, cost of mass-produced goods, etc.
It is not, however, a general rule of thumb to apply blindly to all things "technological"!
This seems to be Kurzweil's big mistake.
He extrapolates features of technology to an unrealistic infinity.
For example, Moor's law is running up against the quantum limit, so micro-processor size is exponential up to a fast-approaching limit.
To take another example, the cost of an iPod may drop exponentially as you scale up production, but you can only sell so many iPods.
Once everyone has an iPod, you read a production limit, and the price becomes stable.
The major claim in the book is that brains will merge with computers.
Kurzweil argues that since transistors are faster than neurons, they will make better brains.
The fallacy here is that you would ever WANT to build a brain out of transistors in the first place!
Neurons can network widely at a low price, but wide networks of transistors are slow and costly in power and heat.
No engineer would try to build a brain out of transistors.
The lesson here is that biological evolution, while it's scope has been limited, will ALWAYS win in a contest with human engineers.
The better your artificial brain performs, the more it will look like a human brain.
This is not coincidence.
One important point made in the book is that our only problem need be to produce an intelligence greater than our own. Once this is accomplished, all other tasks can be up to that greater intelligence. While this is absurd applied to most practical problems like baking a cake, it makes some sense in the realm of AI. Here I disagree with Kurzweil, who asserts any such improved intelligence would be "non-biological". Heck, many parents achieve this goal by giving their child a good education!
Another thing he gets right is the demystification of Searle's Chinese Room thought experiment. Searle's objection to an artificial brain is wrong. Mind is platform-independent, but the port is yet to be written.
Ray makes the same tired arguments over and over, redundantly redundant.
You can almost hear him TRYING to keep alive his delusional dream of living forever!
Ray argues that a future AI will be produced with the ability to iteratively improve its own intelligence.
Among his many skills, he's an accomplished software engineer, so he should know better.
This will never, ever happen in this millennium.
Even if we assume computing resources increased by a dozen powers of ten.
Even if we "reverse-engineer the human brain" (whatever that means).
I know this review is getting long, but he does make some speculations about the year 2010, most of which never came to pass. He predicts we will have virtual assistants that can look-up movie actors, etc. that will respond to our vocal queues in virtual vision contact lenses. Instead, we have Wikipedia on our iPhones. Pretty far off the mark, if you ask me.
By far the most annoying thing Ray brings up no fewer than 10 times is that the speed of light may be "circumvented". I swear, there's nothing sacred to this man! He embarrassingly bungles an explanation of quantum entanglement, calling it "quantum disentanglement" and mistaking spin axis for wave function phase. Yikes.
This book is a big house of mirrors meant to disguise the lunacy of the thesis.
Don't get me wrong, the mirrors are interesting to look at in their own right:
Nano technology, genetic engineering, genetic algorithms, neural networks.
It's fun stuff!
But the singularity is not near, it is the delusion of an old man who would like very much to live forever.
(If you loved "Future Shock", and "The Celestine Prophecy" changed your life, this is the book for you)
But, wait! All those 5-star reviews gotta count for something, right? Well, let's take a look.
"We will have the requisite hardware to emulate human intelligence with supercomputers by the end of this decade."
Really, Ray. How's that coming along? You've still got a year, two if we're charitable. But, even despite the spectacular vagueness of the claim, things are hardly looking good.
"For information technologies, there is a second level of exponential growth: that is, exponential growth in the rate of exponential growth".
A breathtakingly audacious claim. Without a scintilla of evidence provided to justify it. Graphs where the future has been conveniently 'filled in' according to the author's highly selective worldview do not count as evidence, and are nothing more than an embarrassment.
But then, most of the graphs in this book do not bear up under close scrutiny - their function is more cartoon-like. Even Kurzweil's more apparently reasonable claim - that of exponential growth at a constant rate - rests on a pretty selective framing of the question and interpretation of existing data.
"Two machines - or one million machines - can join together to become one and then become separate again. Multiple machines can do both at the same time: become one and separate simultaneously. Humans call this falling in love, but our biological ability to do this is fleeting and unreliable."
Say what now?
From a technical standpoint, as far as biotechnology is concerned (which is the area I am most competent to judge), there's hardly a statement that Kurzweil makes that is not either laughably naive or grossly inaccurate. Assuming that, indeed, drug delivery via nanobots and the engineering of replacement tissue/organs will at some point become reality, Kurzweil's estimate of the relevant timeframe is ludicrously optimistic. A relevant example is the 20 years it took to derive clinical benefit from monoclonal antibodies -- the rate-limiting steps had little to do with computational complexity. So the notion that, in the future, completely real biological, physiological, and ethical constraints will simply melt under the blaze of increased computing power is fundamentally misguided.
From a statistical point of view, things are no great shakes either. His account of biological modeling is such a ridiculous oversimplification it defies credulity. I'd elaborate, but frankly, the whole sorry mess is just starting to irritate me.
Given the density of meaningless, unsubstantiated, and demonstrably false statements in the first few chapters, it's hard to see the point in continuing. If one actually reads carefully what he's saying, and assumes that he is assigning standard, agreed-upon, meaning to the words he uses, then several possible reactions seem warranted:
* that sinking feeling that one inhabits a universe that is completely orthogonal to those who gave this a 5-star rating
* heightened skepticism and aversion to Kool-Aid
* bemusement at the gap between Kurzweil's perception of reality and one's own - in particular, the evident moral vacuum in which he "operates", as well as apparent ignorance or indifference to the lot of the vast majority of the planet's inhabitants
* wonder at the sheer monomaniacal gall of the man
Grandiose predictions of the future, the more outlandish the better, appear to have an undiminished appeal for Homo sapiens. For the life of me, I have never been able to figure out why.
I would consider this an 'impact book', one that truly changed the way I perceive the world. Kurzweil aims to convince his reader that we are on the cusp of an exponential growth in genetics, nanotechnology, and robotics (GNR) that will fundamentally change humanity, creating humans that are fully integrated with machines, live as long as they like, and frequently immerse themselves in virtual worlds. Its premise sounds a bit far-fetched but his meticulous research, incredibly broad grasp of current research, and history of success in predicting technological growth are surprisingly convincing. The book can be repetitive as Kurzweil feels the need to offer each and every criticism a full rebuttal, frequently reusing the same points, and to repeatedly explore the same concepts in various scenarios. If you can get over the repetitiveness, accept the fact that some of the science of this book will probably go over your head, and allow yourself to be open to the possibility that Kurzweil's futurist predictions might actually be feasible, this book will introduce you to an incredible new world, decades before you actually meet it.
Kurzweil has made a living of being a futurist and an inventor. Many of his inventions are the result of his predictions coming true, so there is good reason to listen to what he has to say on the topic. The main idea is that the evolution of technology is not linear (as most people think) but exponential. This exponential development of key technologies leads to dramatic changes in human history over relatively short periods of time. Good examples include the internet and cell phones. The book focuses specifically on 3 key technologies that will produce the human "singularity", an event where humans transcend their former selves and become something more than human. These key technologies (known as GNR) are genetics, nanotechnology, and robotics (or artificial intelligence). When these 3 things progress and converge in the next few decades, we will see humanity benefit by eradicating disease, prolonging life expectancy indefinitely, and promoting human intelligence to astronomical levels through direct neural connections to computer hardware. Effectively, we'll become so smart we'll be able to outlive and outhink "normal" humans to an unimaginable degree. He makes a compelling argument that the singularity is not a matter of "if" but of "when", and that we should be proactive in pursuing these technologies, not just for the benefit of humanity, but to keep amoral people from exploiting these things to an unfair advantage. It's a fascinating read and worth digging into if you have any appreciation for science in general.