In Ray Kurzweil’s amazing 1999 book, The Age of Spiritual Machines: When Computers Exceed Human Intelligence, Kurzweil predicts that computing power will continue along the exponential track of Moore’s Law, such that by the year 2030, a $1,000 personal computer will be 1,000 times more powerful than the human brain. At that point, computers will be capable of learning and creating new knowledge entirely on their own with no human assistance. By scanning the compendium of knowledge on the Internet, some computers will “know” literally every single piece of public information generated by human beings (every scientific discovery, every book and movie, every law and theorem).
A talk by Hans Moravec explores this concept further and illustrates this trend with an interesting chart on the evolution of computer power vs. cost.
As Kurzweil says in his book, your $1000 personal computer will be able to simulate the brain power of a small village by 2030, the entire population of the United States by 2050, and a trillion human brains by 2060. By 2100, one penny’s worth of computing power will have a billion times greater computing capacity than all humans on Earth.
Kurzweil further predicts that this rise of Artificial Intelligence will result in a “robot rights” movement in which there is public debate over what civil rights and legal protections machines should have. Given that many humans by that time will be part-machine themselves, with bionic limbs and cybernetic implants, there will be great debate as to what constitutes a “human being,” a subject explored briefly in the Will Smith movie I, Robot.
And of course you cannot discuss this robot future without considering the scenario proposed by the Terminator movies. Will computers eventually become sentient and take over the world?
Kurzweil points out that the more powerful technology — or rather, the civilization with more technological sophistication — always wins. This appears to be what happened when our Homo sapiens subspecies met the Neanderthals, when the technologically advanced Europeans met the indigenous people of America, and when the United States unleashed the atomic bomb.
By 2100, the computing power of machines will surpass that of humans so greatly that computers may perceive our human intelligence the same way as we humans perceive the intelligence of a hamster. We could find ourselves as household pets in a machine world, or worse, as a biological power source for a network of machines a la The Matrix.
One saving grace to this doomsday scenario is that computing power does not equal intelligence or innovation. My computer is already much faster than my brain and can remember facts I’ve long since forgotten, but Vista UAC still asks me three times to confirm when I want to delete a file on the C: drive. Oh that’s right, Vista UAC is stupid because of the human programmers who wrote it. Perhaps we humans are doomed.
Article published on December 9, 2008
If you like this article, please share it:
|
December 9th, 2008 at 5:09 pm
Ray Kurzweil is a fruitcake who is just cruising for a bruising. He did not learn from the 1960’s when AI researchers claim that AI would advance to a state where computer can replace lawyers, perform surgery and plan human-free space-missions and negotiate permanent between in the middle east. They were so wrong and lost so much public-confidence (aka research-funding potential) that a many research funding agencies would not fund the field sufficiently- like in England. Someone need to teach Ray about the limits of Moore’s law, the complexity of the human brain, NP-complete and other real-world circumstances that prevent day-dreams from meeting reality.
December 10th, 2008 at 11:37 am
Actually Ray Kurzweil is a genius, but he’s like most academics in that he doesn’t understand that the real world behaves very differently than the controlled environment of a classroom or laboratory.
History has shown that technological soothsayers overestimate the big changes (e.g., flying cars in Blade Runner, commercial space trips to the Moon in 2001 Space Odyssey) while totally missing the more subtle changes (like LCD screens today vs. the tube screens shown in 2001). This is because humans, while highly adaptable, are still rather set in our ways. Even if we could have flying cars, we’re not going to anytime soon because there are too many potential problems (accidents, for one).
So while I believe that computers will be as powerful as the human brain in a couple of decades — when compared in terms of raw computing power — it will likely be another century before computers can think and reason and plot like humans. So I think I and my kids are safe, but my great-great-great-grandkids better watch out!
December 10th, 2008 at 3:51 pm
Rise of the machines will happen also sooner than a century away. Why? Not because the machines will rise up. But because someone will create a militarized robot weapon, and it will get out of hand and wipe us out. Probably more like 2040.