Sean King

My photo
Knoxville, Tennessee, United States

Sunday, June 29, 2008

The Looming Technological Singularity

I recently read Ray Kurzweil's The Singularity is Near and found the concepts its discusses to be both compelling and fascinating.

One of Ray's main points is the exponential growth rate of technology. He notes that technology not only grows exponentially (i.e., in multiples rather than via additions), but the exponent (i.e., the multiplier) grows exponentially as well. This has been best documented in Moore's Law, where the number of transistors on a computer chip, and therefore its computational ability, doubles over a given period of time. Twenty years ago it took approximately two years for the computational capacity of a given chip to double, and now it takes only about 1 year due to growth of the exponent.

Kurzweil notes that standard exponential growth (where the exponent remains constant) can appear to be very slow at first, but at a certain point, referred to as the "knee of the curve", the growth rate quickly becomes nearly vertical. The deceptive power of exponential compounding is demonstrated, for instance, in a common (joking) bet among golfers: The bet is that they will start out playing golf for two dollars a hole and then double the bet every hole. Thus, the first hole is worth two dollars, the second 4 dollars, the third 8 dollars, the fourth 16 dollars , the fifth 32 dollars, the sixth 64 dollars, the seventh 128 dollars, the eighth 256 dollars, and the ninth 512 dollars. Now these numbers are getting large enough to cause the golfers a little heartburn, but they are still manageable, at least for some. However, after the 9th hole, they have reached the "knee of the curve", and the numbers quickly get out of hand. So much so that the last hole is worth an incredible $262,144, and this is with only a constant (i.e., non-increasing) exponent of two. Imagine how large the number would get if instead of doubling the bet every hole, you doubled it only after the first hole, then tripled it after the second, quadrupled it after the third, etc. This latter example is more like how technology works (though the exponent grows more slowly), and one can easily see how quickly the numbers get inconceivably large.

Ray's point is that we are currently at the knee of the computing curve, and that the rate of growth of in the computational power of computers will quickly become apparent to everyone in the most profound of ways. While we should rightfully be amazed by what has happened over just the last ten years--things such as GPS navigation, Google, Google Earth, social networking sites, etc., these are nothing compared to what the near future holds. For instance, simple math demonstrates that, if the exponential growth in processing power continues, and companies like Intel insist it will, a thousand dollar computer in the year 2020 will have the processing power of the AVERAGE HUMAN BRAIN and, because of continued annual (or faster) doubling thereafter, a single thousand dollar computer will have the processing power of ALL HUMAN BRAINS ON EARTH by only 2050! By the mid 2040's, the total ANNUAL INCREASE in artificial intelligence (adding all computers together) will be a BILLION times greater than the intelligence represented by all human intelligence on earth today, and very soon the millions of computers that comprise the Internet (including yours) will be able to combine their resources to achieve extraordinary things at little to no additional cost to humanity. The practicality of this last thing has already been demonstrated through, for instance, the,2283-order,1-page,1-c,screensavers/description.html project.

The significance of this increase in technological intelligence cannot be overestimated, and the resulting consequences are not likely to be all good, though, like Kurzweil, I'm an eternal optimist. One outcome of this technological expansion, which Kurzweil and others refer to as The Singularity, is that humans will ultimately be forced to merge with their technology, or else be overrun by it. Now "forced" is probably too harsh of a word, because the benefits will be such that most of us will voluntarily merge when given the chance. But, either way, sooner than anyone realizes humans will "transcend biology." If Kurzweil is right, the age of the cyborg looms nearer than we suspect.

In the meantime, Kurzweil argues that something else that is far closer than most realize is near limitless human life expectancies. Soon, we will reach a tipping point where the benefits from scientific research over the course of any given year will result in MORE than a year's increase in average life expectancies. At that point, human natural lifespans will increase immeasurably, and Kurzweil theorized that those humans who are willing to merge with their technology--as many of us already do today via artificial joints, artificial hearts, cochlear implants, etc.--may well gain an element of virtual immortality (pun intended).

While virtual immortality sounds admittedly far-fetched, the tipping point where life expectancies increase by more than a year for each year that passes does sound feasible. That this tipping point may be closer than most realize is evidenced even in recent news stores such as this:

That's right. US life expectancy increased by an amazing 4 months just between 2005 and 2006(!), reflecting the exponential increase in life expectancies that has continued now for over a century. If Kurzweil's correct, this trend will continue to accelerate until, in only 10 to 15 years, we may be able to prolong life indefinitely.

Many on today's left may ponder the ecological consequences of substantially increased life expectancies, but rest assured that if our intelligence (some biological, but mainly technological) advances to the point where our natural lives are effectively limitless as contemplated above, that same technology will be more capable of dealing with the ecological fallout that such long lives entail, or so Kurzweil insists. He spends a significant portion of his book explaining how this will be so.

The more interesting implications of indefinite life expectancies are not the practicalities of managing the environment, but rather the philosophical/religious ones. Many insist that death gives meaning to human lives, and that fear of death is important to our spiritual/psychological/philosophical well-being. But, I have to say, once one comes to truly believe as I have that indefinite life expectancy is not only possible but is only a decade or two away, contemplating living forever, or even for several centuries, is nothing short of terrifying. As a result of this fear, isn't an axiom like "eat, drink, and be merry for tomorrow you may die" equally as compelling, maybe more so, when rephrased as "eat, drink, and be merry, for you will be living one hell of a long time and you might as well enjoy it"? In other words, which is more terrifying and which results in more soul-searching: The thought of dying (and all that entails), or the thought of living forever?

For me, it's the latter.

1 comment:

Anonymous said...

Heh. I'm thinking of Battlestar Galactica now. (By the way, have you watched the recent re-make, because I think you'd like it.) Computers will take over and kill humans.

Seriously, though, I'm waiting for the chip that they put into my brain that will allow me to send messages to anyone else on earth just by thinking. Though, really, that'd be information overload. I guess what I'd most like is a chip that would let me record my thoughts at will, so that all of the great ideas that I get when I'm not where I can act on them (like thoughts about my novels) can be captured immediately.

Apparently, I can ask for that for Christmas within a few years, yes? ;)

-- Your Sister