AT&T Tech Support 360September 29, 2005 4:00 AM PDT
Newsmaker: Ray Kurzweil deciphers a brave new world
See all Newsmakers
Now Kurzweil, who can claim credit for developing the first text-to-speech synthesizer and the first CCD flat-bed scanner, is busy inventing a future in which humans merge with machines and the pace of technological development accelerates beyond recognition.
The concept is known by its proponents as the Singularity, and until recently it's been the province of science fiction authors such as Vernor Vinge and Ken MacLeod.
Now Kurzweil, in a new book called "The Singularity is Near: When Humans Transcend Biology" (Viking, $29.95), claims that the inexorable pace of technological development and its exponential growth will usher in the Singularity by 2045.
To appreciate the dizzying scope of Kurzweil's predictions, read the book. But the condensed version goes like this: Thanks to Moore's Law and other exponential growth rates, by 2030 a $1 computer will be as powerful as the human brain. Information technology's exponential curve will fuel advances in biology, robotics, nanotechnology and artificial intelligence--with world-shattering results including radical life extension and practically omniscient and omnipotent abilities for humans who elect self-augmentation.
Adding some heft to this prediction are Kurzweil's earlier books, including "The Age of Intelligent Machines" and "The Age of Spiritual Machines," which show that he has a better-than-average record of predicting the future. In 2002, he was inducted into the National Inventors Hall of Fame and he has received the $500,000 Lemelson-MIT Prize and the National Medal of Technology.
CNET News.com spoke with Kurzweil on Wednesday about his book tour, his views of the melding of man and machine and the political ramifications of having hyper-intelligence initially available to the very wealthy.
You've been busy promoting your new book. How's the tour going?Kurzweil: It's going great. Seven presentations last week including a keynote at the Accelerating Change Conference, spoke at Google, the Software Development Forum. I gave a keynote at George Gilder's Telecosm, which he named after the book. I'm going to speak at MIT tomorrow.
You said at a speech last week in San Francisco that you were working on a project with former Microsoft CFO Michael W. Brown that'll result in a hedge fund. Can you tell me about that?
Kurzweil: It's been a major project for about six years. It's applying my field, detecting subtle patterns, and using technology forecasting. Six years ago the project wasn't fully feasible because we didn't have rapid access to all ticker data for stocks. You really couldn't place trades very effectively online. The technology wasn't there--it can't take two weeks for the computer to make a decision that needs to be made in five seconds.
(My system) doesn't make perfect predictions. But what we can do is predict them substantially better than chance. That puts us in the position of being the house in a casino. It places lots of bets, some win and some lose, but it consistently makes money. We haven't had a down month yet. It makes 80 to 100 percent returns a year.
When will this be open to people qualified to invest in a hedge fund?
Kurzweil: We've been investing our own funds just to demonstrate it works. We're launching a hedge fund starting in January. But this will be principally open to our own investors.
You talk a lot about exponential growth, including the stock market. But the Dow Jones Industrial Average was around 1,300 points higher in 2000 than it is now. Adjust for the cost of trades and real inflation and you might have lost 6 to 7 percent a year, no?
Kurzweil: It's the power and adoption of information technologies that moves exponentially. The stock market includes many businesses with old business models...You do have long-term exponential growth of the stock market. The measures we find are highly predictable, surprisingly so. But there are business cycles.
How about the Nasdaq index, which features newer business models, and has fallen by over 50 percent from its top.
Kurzweil: Not every technology company succeeds. You have as much creation as destruction. Companies need to reinvent their business models quite regularly. It was only three years ago that people were saying you couldn't make any money in Internet advertising. The psychology of the stock market has its own dynamic. The boom-bust dynamic is a harbinger of (technological) revolutions...Now e-commerce is growing at a point where it's significant.
It's not a theory of the overall dynamics of the stock market, which have a lot of psychological aspects, some overly pessimistic and some overly optimistic.
Your concept of the future relies heavily on "strong AI," the idea that artificial intelligence will become self-aware and eventually surpass human intelligence. But it seems like AI researchers have abandoned that idea for focused real-world applications like face recognition. I was at a speech this week where computer science professor Rudy Rucker said that strong AI was dead.Kurzweil: There are hundreds of applications where AI is performing projects that would have required a human level of intelligence a few years ago. Those include diagnosing heart disease, routing e-mail messages, cell phone routing, landing planes.
We are now in an era of narrow AI, meaning it's not strong AI. It's not the full range of human intelligence. But it's performing functions that used to require human intelligence. Looking for credit card fraud is one example of that. These were research projects 15 years ago.
This isn't 2029. We'll make a billionfold increase in hardware capacity between now
See more CNET content tagged:
hedge fund,
prediction,
Artificial Intelligence,
project
Biologically, human is a ?machine? of things, no different from all the ?things?. As a spiritual machine, human assembled words such as Singularity, Intelligence, Strong AI and Love, Money etc, while no one really know the ?real? definition of these words. Well, these words did get your attention ? for the money in your wallet.
Inventor is nothing but an assembler. Prediction is nothing but another ?one-man opinion? for you to read while you don?t have anything else to do or think. Of course, reading grows knowledge by brainwashing your belief ? what a tricky way to control you!
Therefore, Tech will merely oscillate us with greater diverging resonance. Singularity can not be accomplished by Tech. Conclusion, Kurzweil?s prediction will fail. What?s in your belief?
but really, you have some very good points about the fallacies of language, and your argument against the concept of the singularity is correct in my opinion.
These Fundamentalists have forgotten the basis of the humanist agenda is to achieve greater well being for humanity through means such as advanced technology. Attacking it in a shallow pseudo-Marxist anti-corporate manner, and trying to always show things in the most negative manner possible in order to show off how much of a "realist" he is, just shows how far he and his pals have strayed from rational optimism.
It's too bad Ray didn't request to be interviewed by someone who could at least attempt to be objective.
What would you have asked him? Where do you see his ideas breaking down? What is your view of the future?
.
Now by “hardware capacity” he could mean almost anything, but most readers would assume computers --especially since he’s discussing machine intelligence. Moore’s law is an order of magnitude increase every 5 years so in 25 years “hardware capacity” will have increase by 10^5 not 10^9 – off by a factor of ten thousand -- assuming of course that Moore’s law (which is only empirical after all) holds that long.
Another problem is that Kurzweil shifts definitions in mid-sentence. In discussing machine intelligence he cites examples of tasks that formerly required a human but are now done by machine. Not the same thing at all – we don’t call tractors artificial horses. I’ve no doubt the future will have very capable machines – but intelligence is another matter. Kurzweil assumes such capable machines will start setting there own agendas, something that is difficult to define (why do humans do what they do?) much less implement.
Finally the choice of the singularity metaphor is interesting. Presumably this is a take on black holes. An interesting aspect of black holes is that someone falling into one never arrives – the singularity is always just ahead; yet to an outside observer the person falling in disappears rather quickly (except for the possibility of ghosts – which is another topic). So in this metaphor – the singularity will never occur (except for alien observers).
Similar to your point, if our technology could not 'simply' use H (hydrogen) as a source of power - like tree and human body, what are we talking about the much 'complex and compound'singularity?
So, if your calculations show an order of magnitude increase in 5 years, to get his number for 25 years you don't just multiply your number by 5. You say that the next magnitude increase will take,(pulling numbers out of my thin air) , 4.5 years, then 4, 3.5, 3, 2.5, 2, 1.5, 1 and pretty quickly you reach his singularity when things are increasing at an infinite rate.
BTW, if you take my numbers, in 24.5 years you have an increase of 10^7 which is still off by a factor of 100, but to reach that would only take an additional 2.5 years (or 27 all up)
My problem is that Ray in his past books has started with 'todays doubling rate' at the overly optimistic end of the range of estimates.
I think he takes todays rate of doubling as a figure of 1 year whereas 1.5-2 years would have made his calculations more in line with the common peoples understanding of Moores Law doubling rates.
First, Law(s) of Acceleration could go positive (+) or negative (-) direction.
Next, Strong AI machine(s) do exist. We (and God) call them ?Human?.
Next, if humans need AI to achieve our perfection, then how could human design a perfect AI, while an AI like human will still be ?Human?.
Next, who are going to write the code? Of course, The human alike AI = the human.
Speaking of the Code, in order to reach Singularity, therefore, we got to find out The One Line (Singularity) Code for the Universal Energy Law (UEL), then we can use it to understand how ?energy? ?process? in human brain ?structure?, then AI is done.
By the way, e=mc2 is not a Universal Energy Law, it only good for calculating material thing, not for measuring emotional = intelligent energy.
Can anyone write an equation for UEL?
As an airline pilot and testpilot I am used to do my work in between robots: autopilots connected with navigation computers, assisted by failure management systems. Currently I fly the A330, one of the most advanced planes today. Although it can land fully by itself, all systems are still "stupid". There is no intelligence at all.
If something goes wrong, common sense is what saves the day. Every day again. I have NEVER seen anything that even looks like common sense in my career. I must admit that on the other hand computers have saved the day many times. But that is because they lack common sense.
I'm not saying that common sense will never be achieved by computers, but it's my opinion that common-sense is related to quantumphysics. So only when computers are able to use quantum-events in neural nets, I'll be happy to be passenger on a plane without pilot....
Regards, Peter
Mr. Kurzweil talked about his system can predict and make a decision and consistently makes money ? 80 to 100 percent returns a year, according to Ray in the posted interview by Mr. McCullagh.
As we all know, stock investing involves Science (S) and Control (C) of emotion, where S should cover all the economic ?numbers? of the investing stock (company), while C should include the emotional ?numbers? such as how all the investors (include you and me) and analysts ?feel? ? control - about it.
S and C not only set the direction, distance, speed and acceleration for the path and pace, but also provide (reasons and results of) sudden impulses ? discrete gaps of big drop and jump - of the stock. ?Discrete(s)? often occur in the pre-market and after-hour market ? especially at the earning report day or FOMC meeting day ? are tough for average (all?) investors.
To consistently make money, Mr. Kurzweil?s system should be able to calculate both ?numbers? of S and C ? emotional factors.
Ray also mentioned that his system is not perfect yet. Well, emotional impulses (discrete nature) are particularly harder to predict. By the way, like the stock market, the occurrence of ?discrete? might just well be the limitation, boundary (the end?), and the ?pull back? of our exponential technology expansion. Whether it is a mathematical behavior or natural phenomenal, when a divergent function approaches ?near? the infinity (undefined) boundary, ?discrete? takes place. It happens and it definitely going to happen again ? how?
Besides stock investing, basically our daily life can be measured (encode and decode) by the ?energy? ?numbers? of S and C. Therefore, a Universal Energy Law can be written as Energy = nS + mC + lR. For instance, the term nS could represent Einstein?s mc2 in measuring material (with mass) of energy. Naturally, mC term is for emotional energy measurement ? how?
Again, an approval from News.com is needed to further discuss UEL.
For all the guys and girls out there, good luck for all your S and C ? including all your portfolios.
What will actually happen, as Mr. Kurzweil and many others point out, is that we will integrate ourselves with machine intelligence in a way that makes us as intelligent and competent as any artificially intelligent computer.
For example, through nanotech, we'll have millions of cellular-small mini-computers in our bloodstreams and brains, acting in concert with out neurons to make use of both our own massively parallel biological minds, as well as the lightning quick electronic speeds, and memory capacity, of computers.
The goal is to increase our intelligence as we increase the intelligence of machines: Darwinian symbiosis not cutthroat competition between man and metal.
It's easy to fear the future. But it is more empowering and more interesting to embrace it--and help to create it!
- Technology is growing Smarter
-
by
October 16, 2005 11:38 AM PDT
- Have you called an automated phone answering system recently? Instead of having to respond to the techno-woman voice using dial-tones, you can actually respond verbally- and it understands you!
-
Reply to this comment
-
See all 27 Comments >>This software isn't magical, impossible, a long way off (or even inevitable!). Businesses demand better, smarter technology because it makes them more money. How long until our cars, computers, etc. have voice recognition? Our phones have had it (although the software is somewhat frustrating) for quite a while.
What next after voice recognition? Facial recognition will be ushered in to usage. Software will be more intelligent. It happens, it makes a lot of money for businesses, and it gets better. How often do you use a spell checker?
Who is going to create the first Google-bot? You offer it a voice command about some topic you need information about, and it googles your topic, analyzes and organizes the information, and produces a report for you in seconds. This is an example of lo-fi human intelligence enhancement.
Will computer chips get smaller? Of course they will. Moore's Law (a debateble "law") is a driving *motivation* in the industry. People PAY MONEY for faster equipment. Multi-BILLION dollar corporations are pouring money into nanotechnological research and new chip making techniques.
People want to store gigabytes of songs in their cellphone/ipod/digitalcamera/camcorder, and major corporations will provide them with this capability or certainly die trying!!!
You can't deny that things are getting better- and getting better at a faster and faster rate.
Programmable nanotechnological molecular machines are on their way. Intelligent computers are on their way. Things are going to start getting massively, massively different and soon- and we are going to be caught completely, utterly unprepared because nobody expects it.
I urge you to visit the Singularity Institute at www.singinst.org