In one of the buildings at NASA's Ames Research Center, within walking distance of the Googleplex, elite groups of very smart people are trying to prepare for a future so advanced we can't even predict what it'll look like. This Singularity University is a hub for forward-thinking experts to learn about robotics, artificial intelligence, and other key technologies of the next century.
What's conspicuously absent is a serious discussion of this future's economy. For hundreds of years people have brushed aside Luddite complaints and kept on creating new jobs out of the ashes of dead industries. However, the link between technological gains and employment growth is becoming frayed.
The singularity for which the university is named is one theoretical end result of all human technological progress. It's a point at which we finally create machines that vastly exceed human ability in every possible way, starting a runaway process of computer self-improvement that ends in sci-fi utopia. Man and machine will merge. Angels will sing, perhaps in binary code. Underpinning this is Ray Kurzweil's assertion, backed by numerous statistics, that technological progress is exponential, and thus gets faster and more dramatic over time.
The inconceivably computational future, Kurzweil says, is going to be our ultimate achievement, a networked nirvana. That is, if we're ready for it.
The data divide
Unfortunately, not everyone is ready for the future. Many people are barely ready for the present. Pundits talking about the possibility of a lost decade often overlook the previous decade, which was the first since the Great Depression with no net job creation. Despite the lack of hiring, real GDP grew by nearly $2 trillion, adjusted for inflation -- to put that in perspective, it's about the size of Italy's economy. What made up the difference? Productivity.
Source: Economic Policy Institute.
This optimistic productivity trajectory may be underestimated in many ways, since it's virtually impossible to calculate the productivity society gains from free digital resources like Google and Wikipedia. But wages simply haven't kept up. I showed in a previous article that average income grew far less than the income at the top of the scale over this period of time, but median income -- which may be closer to the reality for most workers -- actually fell in the last decade.
Rise of the machines
Productivity growth that doesn't correspond to a rise in employment or income has to come from somewhere. Are the workers still on the job that much more productive with fewer coworkers? The explanation, according to MIT researchers Erik Brynjolfsson and Andrew McAfee, coauthors of Race Against the Machine, is rising automation. The trend's gotten more pronounced since the recession, as you can see on the chart. Equipment and software spending is up 26% since the recession, though payrolls are essentially flat.
When you stop to consider the many things automated processes can now accomplish that were once thought exclusive to humans, this isn't so surprising:
- Image recognition technology from Xerox (NYSE: XRX ) accurately analyzes and ranks photos by appearance in ways resembling human judgment.
- IBM (NYSE: IBM ) simulated a mouse brain with 512 standard processors earlier this year, close to Kurzweil's prediction of commercially available mouse-brained computing power around the start of this decade. IBM is also working on simulating the human brain, and is about 4.5% of the way there.
- Google (Nasdaq: GOOG ) is investing in robotics research, and has already produced teleconferencing robots and automated cars.
- Northeastern University researchers have created a virtual nurse, which offers rudimentary social and medical interactions for lonely patients and has proven an effective stand-in for the real thing.
- Remote-controlled or autonomous robotic vehicles developed by iRobot (Nasdaq: IRBT ) and Lockheed Martin (NYSE: LMT ) are increasingly common tools in the U.S. military arsenal.
Historical moments of mechanical triumph over man now run the gamut from John Henry to Garry Kasparov to Ken Jennings on Jeopardy. You say John Henry won? The machine could keep working after the contest was over. Henry couldn't. He died.
Debunking the hype
In The Singularity is Near, Kurzweil claims that "the law of accelerating returns is fundamentally an economic theory" and predicted a tripling of stock prices from their 2005 levels. The Race Against the Machine authors claim that we can stem the tide of job losses simply by reforming education and emphasizing entrepreneurship from an early age. Human ingenuity and creativity will win the day, and we can learn to get along with the machines as they make us all rich.
Kurzweil's claim ignores the obvious fact that our economy -- and much of our market prices -- is built on consumption, and a world run by machines is one that won't support the same levels of consumption if those displaced have no easy way back into the workforce. The MIT authors, in their prescription for the future, choose to ignore a point brought up midway through their own book, which is that "Many industries are winner-take-all or winner-take-most competitions, in which a few individuals get the lion's share of the rewards."
No easy answers
Here's a hypothetical situation: someone's lost his job as a construction worker. He wants to become a solar panel installer, which is a kind of construction. This is an easy enough career change. Now, let's say there are 750,000 unemployed construction workers, and perhaps 5,000 new solar panel installation jobs available each year. These aren't hypothetical figures; they're based on numbers available from the Bureau of Labor Statistics.
Let's say instead that the construction worker wants to become an electrical engineer. There are less than half as many electrical engineers as there are unemployed construction workers, and the roster isn't expected to grow by much over the coming decade. High-tech jobs aren't made by the millions, and when they are created, most people just aren't able to do the work.
It's easy (and popular) to talk about making education work in the 21st century, but our schools still function much the same as they did two centuries ago. Real reform would gut the whole system and start with a fresh blueprint, but even that might be too little, too late. If change, as Kurzweil says, keeps speeding up, those unemployed construction workers might graduate to find their newly specific technological knowledge bases already outdated.
Creativity and entrepreneurship are not panaceas either. If 20% of creative professionals and business owners get 80% of the work, that still minimizes the earning potential for others across a broad swath of industries. Some people can be starving artists, but when you've got a family to feed, playing a gig at the local dive for free beer and publicity just won't cut it. Selling knickknacks on eBay to make a living seems like a great idea when few people are doing it, but not when the same thing is listed by thousands at the same time.
Thou shalt not make a machine in the likeness of a human mind
What are we left with? The singularity has great appeal to technology's wealthy elite, but its benefits to the rest aren't as understood. Where will the jobs come from in an economy driven by computerized production? Where will the economy come from when there aren't enough jobs? Accepting a world in which human labor is largely supplanted by technological solutions requires us to radically reimagine the foundations of our economy, which is a step most are not prepared to take.
Day by day, however, the machines are gaining ground upon us; day by day we are becoming more subservient to them; more men are daily bound down as slaves to tend them, more men are daily devoting the energies of their whole lives to the development of mechanical life. The upshot is simply a question of time, but that the time will come when the machines will hold the real supremacy over the world and its inhabitants is what no person of a truly philosophic mind can for a moment question.
A fallacy until proven otherwise
The quote above isn't from a recent book. It was written by Samuel Butler in 1863. Many have predicted a world run by machines and have been wrong before -- but machines have always been narrow in scope and limited in function before. The Luddite fallacy continues to exist because we've always had new ideas for work that computers only could assist with, not replace. There's no clause, however, that says it must exist permanently.
Maybe human ingenuity can still save the day. It always has before. Let me know your thoughts, or your ideas for sustainable change, with a comment.