“Automation should not be an enemy of employment. It never was before. The only difference between now and the past is that now we’re pretending that people who do the real work are actually not,” said Jaron Lanier, explaining why he is concerned that the current high-tech economy is not on a sustainable path.
In a talk at Techonomy 2014 in Half Moon Bay last week, the author, virtual reality guru, and tech consultant advocated for building a democratic and sustainable technologized economy. The problem with business schemes in Silicon Valley—Facebook and Google among them, he said—is not that they don’t create wealth. As Lanier’s audience knows, they do. The problem, he said, is the loss of a bell curve in economic outcomes.
Lanier called the bell curve “the setting for a stable society, an honest and fair democracy, and many other good things.” But when companies use digital technologies to “optimize the world,” he said, the economic outcome is a Zipf distribution (also known as a “winner take all” or a “long tail”). Instead of the bell curve’s normal distribution that has a small group of low performers, a big average, and small group of high performers, the Zipf looks like a hockey stick. A few are at the narrow tiptop, far away from the blade at bottom where most are clustered.
“Digital networking, computations, sensors and actuators to connect those to the real world ought to give us ability to make a world that makes more sense, is more optimized, more efficient, less annoying, and less dangerous,” Lanier said. But they also present a temptation to companies, “which is to turn yourself into a communist central planner, where whoever is closest to the most central and most influential computer gets to run everything.”
Organizations leverage such strategies to “reap incredible near-term benefits,” he said. But such concentration of power at the top eliminates the competitive market and results in a failure of democracy and economy.
As an example, Lanier described a strategy that he helped Amazon develop. In order to optimize its supply chain, Amazon collected data that enabled it to predict the absolute bottom line for anyone it would negotiate with. “It was kind of like counting cards: they could enter a negotiation knowing more than the people they were negotiating with. They had information superiority,” Lanier said.
As a result, of course, the company rapidly grew wealthy. But not in a sustainable way. “They’re starting to see limits to their growth and real problems because they’ve impoverished their customer base too much,” Lanier said. “This is the problem with a Zipf distribution: it’s not economically stable.” Lanier also pointed to Enron, Long Term Capital Management, and bundled mortgage-backed derivatives as other failures that were consequences of digital optimization of markets.
The same thing will happen to Google and Facebook eventually, he predicted. “All of these Zipf-creating schemes are not sustainable.” And since the turn of the century, Lanier said, bell curves have been disappearing in favor of Zipf distributions. Powerful computing technology has benefited the top 1 percent, he said—an unsustainable economic trajectory. “There aren’t going to be any customers to buy your stuff, eventually.”
His economic solution, he said, is simple: “For us in Silicon Valley and for those who start these schemes and for the world at large and in particular for democratic process, we’d all be better off if we could find our way back to the original idea of digital networking.” That idea, which he attributed to Ted Nelson, who proposed it as a 1960s Harvard grad student, is that “a universal micropayment structure” can create a stable outcome no matter how advanced and automated technology becomes.
“Ted thought in the ‘60s that if we have a richly connected graph and we automate the world, instead of everyone starving, we’ll get a nice bell curve distribution because people will be using each other’s information to run this thing somehow,” Lanier said.
Take language translastion, for instance. Rapid, automated translation services from Bing and Google have decimated the volume of work for professional translators. “A very typical reaction from the tech world is, ‘Well, that’s creative destruction … they have to be flexible and find new jobs,” Lanier said.
But in fact, he argued, automated translation tools rely on human translators. Early attempts to develop truly automated tools failed repeatedly. Instead, IBM researchers proposed a tool that used statistics to correlate sample sets from human translators. “You take real examples from real people. You correlate them, and you mash them up. It’s what we call Big Data and Cloud Computing now,” Lanier said.
Today “automated” translation services scrape millions of fresh translations from human translators around the world to support the translation process. “These people are still needed. It’s just that they’re needed in a new way. We haven’t made translation obsolete with some kind of brain implant or something; we’ve made it better. But we snuck in this disempowerment of people who do the work. The new way of doing work is adding data to the cloud—adding valuable data to a big data statistical system.”
Instead of compensating contributors to the system, Lanier said, “We’re putting all of those human translators behind a curtain and pretending they’re not there so we can say, “It’s AI.” But it’s not an AI; it’s an algorithm plus millions of people who are not getting paid.”
The solution, and possibly the only path to a sustainable future, he said, is to “accept true transparency, open the curtain, acknowledge that people are actually doing the work, but just in a new way, in a better way.”
The idea might sound “radical,” Lanier said, but added, “If we could acknowledge the improvements that we’re actually creating, we could create a sustainable and democratic very very high-tech future.”