hen my mom was on the brink of marriage, her mother dragged her out to spend her meager teenage life's savings on a set of silverware. Not flatware and ceramic dishes, but actual silver—something of enduring value that Mom could sell if hard times came along. But I can tell you firsthand, real silverware dents and scratches and oxidizes easily, and who's got the time to polish a hundred pieces? It also conducts heat really well, so that a spoon dipped into hot soup can easily singe the fingers holding it. It has an unpleasant taste, too, at least to my palate. Apparently I'm not the only one who thinks so, because today the whole dining set could be replaced on eBay for under $50, including shipping. The joke is on my grandmother, who simply couldn't envision a future where silver was deemed, if not valueless, then at least inferior to stainless steel. When she was a young girl, aluminum was precious, too, but the relentless march of technology has cheapened it.
Electronics face an even worse situation. In 1965 Gordon Moore, one of the founders of Intel, observed that the number of components that could be placed on a given area of computer chip was doubling every 12 months. He later revised the figure to 24 months, then to 18, where it has held surprisingly steady for the past 40 years. In fact, here at the turn of the millennium "Moore's Law" has become something of a cliché; we expect new computers to grow denser and faster every year, on an endless exponential climb. As a result, used electronics are painfully obsolete and almost completely worthless.
Notice how the left-hand side of the graph looks flat, while the right hand side looks nearly vertical? Interestingly, this shape is insensitive to scale; we can zoom in or pull back, changing the units from kilohertz to gigahertz, from days to years to decades, and the graph will have that same look, that same sudden lurch toward infinity. In the real world, it's alarming when things are suddenly changing that quickly around us; think of hyperinflation, stock market bubbles, rampant real-estate speculation ... These things aren't a lot of fun. Trends like this led science-fiction writer Vernor Vinge to speculate, in his 1987 novel Marooned in Realtime, that the growth of technology, information and resource consumption was flinging us toward a vertical wall—in his word, a singularity—where everything about our civilization would pass through a spasm of overwhelming change, emerging on the other side as something unrecognizably strange. Do we go extinct? Become gods? Change form so dramatically that we no longer recognize ourselves? Will machine intelligences take over the world and (if we're lucky) keep us around as pets?
A misunderstood cliché
Since then, the singularity has become a science-fiction cliché, and a misunderstood one at that. To understand why, it helps to look even further back in time, to the 18th-century British economist Thomas Malthus, who studied the relationships among population, food supply and industrial production. Human populations also have a doubling period—the global average is currently around 50 years—and therefore population grows along the same exponential curve you see above. Malthus argued, basically, that this growth could continue only until the number of people exceeded the carrying capacity of the environment, at which point war, famine and disease would step in to bring the numbers back down with a heavy crash—a different kind of singularity. And then the exponential growth could begin again. This exact pattern can be seen in animal populations like locusts, and in a murkier way it has left its mark on human history as well.
Now, Malthus has been widely criticized as an oversimplifying pessimist; the world is a lot more complicated than this. Malthus never anticipated birth control and family planning, for example, and didn't quite grasp that if our technology improves exponentially, so does the carrying capacity of the land. Although it's unpleasant, I can imagine a world where all the plants have been replaced with superefficient, solar-powered processing stations that remove wastes from the soil and water and atmosphere, producing food and oxygen as byproducts. Cover the Earth in such machines and it could support a trillion people; extend them out into space and the only limit would be the energy output of the sun itself, as noted by Freeman Dyson in a famous 1959 essay in Science. But if you can't leap to the distant stars, that number really is a hard limit, impossible to exceed. Ultimately, sunlight is a limited resource.
Humans will always be humans
Similarly, there are physical limits to Moore's Law; the incredible shrinking transistor can't continue past the size limit of the atom itself. In the best possible case, this leaves room for about 11 more doublings over perhaps the next 17 years—but that's all. And in fact, every doubling in chip density follows a quadrupling in tooling costs; those end-of-the-line superchips will require a $4 quadrillion factory. Since the entire global economy today contains only about $30 trillion, those growth rates will have to level off pretty soon now. The result is the classic S-curve of resource-limited growth. We saw it in the railroads, too; from 1830 to 1900 the American railroad system appeared to be growing exponentially, but then it suddenly slowed, and even declined a bit over the latter half of the 20th century.
Of course, during that same period the railroads were supplanted by superhighways and airlines and rocket ships moving at ever-higher speeds, prompting some observers to predict an endless exponential growth in the velocity of human travel. But that, too, seems to have leveled off; the human speed record was set by the Apollo astronauts on their way to the moon and hasn't been equaled since the final moon mission (Apollo 17) in 1972. In the same way, we can expect mind-boggling advances in stuff like quantum computing over the span of the 21st century. I can already hear the economists speculating that the growth might never end. But if history is any guide, it will end.
I'm not criticizing Vinge here, but I do think the fans and writers following after him have missed the point. Literally: because the moment of maximum change and uncertainty comes at the center of the S, when the upward curve of exponential growth suddenly—and without warning—turns over. In mathematical terms this is called an inflection point, and in socioeconomic terms it spells ruin for investments and forecasts and business models based on continued exponential growth. This is why the bubble economy of the late '90s Internet stocks has been so damaging; we're only now learning to cope with lower rates of network growth.
Don't get me wrong; the singularity (or inflection point) is a force to be reckoned with. We'll hit several of them over the next 100 years—in population, computing power, ecological resources and the explosive growth of Third World economies. It promises to be a rocky ride. Australian writer/philosopher Damien Broderick, author of The Spike, has offered an interesting opinion, that if you connect the centers of these many S-curves you'll find an even larger exponential curve behind them, representing the sum total growth in human capability. And when that one goes vertical, Vinge's predictions—ranging from the transcendent to the nightmarish—will very suddenly come true, for better or worse. Maybe he's right; it's certainly a hard argument to refute.
But again, if history is any guide, that moment of shock, if it comes, will be very brief—a few years at most—after which things will level off, settle down and become a lot less interesting. And while the world on the other side of that event may look very different from the world of today, my personal prediction is that we'll still see ourselves in it, with habits and motives largely unchanged. The desire for comfort, for wealth, for entertainment and novelty and pleasure ... these things will never go away. We'll just be richer, more powerful and wiser for our troubles. And what, exactly, is wrong with that?
Sources used for writing are:
"The Periodic Table in Earth and Sky," 3rd edition, Jenner Scientific LLC, 2005
Wikipedia: ("Moore's Law," "Freeman Dyson"): http://www.wikipedia.org
McCarthy, Wil: Hacking Matter, Basic Books, 2004
Tehranchi, Mike: "Bulls, Bears, and Mathematicians," University of Texas Department of Mathematics
McCarthy, Wil: "Runaway Train," Wired, January 2002
The National Rail Museum: www.nationalrailmuseum.org
The New York City Subway Resources: www.nycsubway.org
The Encyclopedia Britannica, 2004 Edition ("Malthus," "Transportation, History of")
Wil McCarthy is a rocket guidance engineer, robot designer, nanotechnologist, science-fiction author and occasional aquanaut. He has contributed to three interplanetary spacecraft, five communication and weather satellites, a line of landmine-clearing robots and some other "really cool stuff" he can't tell us about. His short writings have graced the pages of Analog, Asimov's, Wired, Nature and other major publications, and his book-length works include the New York Times notable Bloom, Amazon "Best of Y2K" The Collapsium and most recently Lost in Transmission. His acclaimed nonfiction book, Hacking Matter, is now available in paperback.
