The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies

by Erik Brynjolfsson

Some of the entrepreneurs, investors, researchers, and executives that we trust most have told us the technology-based future is arriving much faster than they thought.

Isnt thus relative? Does it really exist, or always in the future?

So this is a book about the second machine age unfolding right now—an inflection point in the history of our economies and societies because of digitization. It’s an inflection point in the right direction—bounty instead of scarcity, freedom instead of constraint—but one that will bring with it some difficult challenges and choices.

This book is divided into three sections. The first, composed of chapters 1 through 6, describes the fundamental characteristics of the second machine age. These chapters give many examples of recent technological progress that seem like the stuff of science fiction, explain why they’re happening now (after all, we’ve had computers for decades), and reveal why we should be confident that the scale and pace of innovation in computers, robots, and other digital gear is only going to accelerate in the future.

The second part, consisting of chapters 7 through 11, explores bounty and spread, the two economic consequences of this progress. Bounty is the increase in volume, variety, and quality and the decrease in cost of the many offerings brought on by modern technological progress. It’s the best economic news in the world today. Spread, however, is not so great; it’s ever-bigger differences among people in economic success—in wealth, income, mobility, and other important measures. Spread has been increasing in recent...

The final section—chapters 12 through 15—discusses what interventions will be appropriate and effective for this age. Our economic goals should be to maximize the bounty while mitigating the negative effects of the spread. We’ll offer our ideas about how to best accomplish these aims, both in the short term and in the more distant future, when progress really has brought us into a world so technologically advanced that it seems to be the stuff of science fic...

When ‘winner-take-all’ markets become more important, income inequality will rise because pay at the very top pulls away from pay in the middle.

But the evidence suggests that the spread of incomes continues at high levels of income with a fractal-like quality, with each subset of superstars watching an even smaller group of super-duper-stars pulling away.†

More often than not, when improvements in digital technologies make it more and more attractive to digitize something, superstars in various markets see a boost in their incomes while second-bests have a harder time competing.

But what if a technology arises that lets each seller cheaply replicate his or her services and deliver them globally at little or no cost? Suddenly the top-quality provider can capture the whole market. The next-best provider might be almost as good, but it will not matter. Each time a market becomes more digital, these winner-take-all economics become a little more compelling.

a traditional market, someone who is 90 percent as skilled or works 90 percent as hard creates 90 percent as much value and thus can earn 90 percent as much money. That’s absolute performance.

prettier icons—might completely dominate a market. There would likely be little, if any, demand for the tenth-best mapping application, even it got the job done almost as well. This is relative performance. People will not spend time or effort on the tenth-best product when they have access to the best. And this is not a case where quantity can make up for quality: ten mediocre mapping tools are no substitute for one good one.

Why Winner-Take-All Is Winning Why are winner-take-all markets more common now? Shifts in the technology for production and distribution, particularly these three changes: a) the digitization of more and more information, goods, and services, b) the vast improvements in telecommunications and, to a lesser extent, transportation, and c) the increased importance of networks and standards.

A similar dynamic comes into play when technologies like Google or even Amazon’s recommendation engine reduce search costs. Suddenly second-rate producers can no longer count on consumer ignorance or geographic barriers to protect their margins.

In a traditional camera store, cameras typically are not ranked number one versus number ten. But online retailers make it easy to list products in rank order by customer ratings, or to filter results to include only products with every conceivable desirable feature. Products with lower rankings or only nine out of ten desirable features receive disproportionately lower sales from even small differences in quality, convenience, or pricing performance.16

Thirdly, the increased importance of networks (like the Internet or credit card networks) and interoperable products (like computer components) can also create winner-take-all markets. Just as low marginal costs create economies of scale on the production side, networks can create ‘demand side economies of scale’ that economists sometimes call network effects. We see them at work when users prefer products or services that other people are flocking to. If your friends keep in touch via Facebook, that makes Facebook more attractive to you, too. If you then join Facebook, the site becomes more valuable to your friends as well.

In addition to the technical changes that have increased digitization, telecommunication, networks, and other factors that create superstar products and companies, there are more aspects at work in boosting superstar compensation for individuals. In some cases, cultural barriers to very large pay packages have fallen. CEOs, financial executives, actors, and professional athletes may be more willing to demand seven- or even eight-figure compensation deals. As more people get those deals, a positive feedback loop emerges: it becomes easier for others to make similar requests.

Instead of being the thousandth-best children’s book author in the world, it may be more profitable to be the number-one author in Science-Based Advice for Ecological Entrepreneurs, or Football Clock Management.

For instance, research by Erik and his coauthors found that book sales at Amazon were characterized by a power law distribution.23 Power law distributions have a ‘fat tail,’ which means the likelihood of extreme events is much greater than one would expect to see in a normal distribution.24 They are also ‘scale invariant,’ which means that the top-selling book accounts for about the same share of the top ten books’ sales as the top ten books do for the top one hundred, or the top one hundred do for the top one thousand. Power laws describe many phenomena, from frequency of earthquakes to the frequency of words in most languages. They also describe the sales distribution of books, DVD, apps, and other information products.

a place like Silicon Valley or a research university like MIT, the rapid pace of innovation is particularly easy to see. Startups flourish, minting new millionaires and billionaires, while research labs churn out astonishing new technologies like the ones we saw in earlier chapters. At the same time, however, a growing number of people face financial hardships: students struggle with enormous debt, recent graduates have difficulty finding new jobs, and millions have turned to debt to temporarily maintain their living standards. FIGURE 11.1 Labor Productivity and Private Employment In this chapter, we’ll address three important questions about the future of the bounty and the spread.

‘Bounty’ doesn’t mean simply more cheap consumer goods and empty calories. As we noted in chapter 7, it also means simultaneously more choice, greater variety, and higher quality in many areas of our lives. It means heart surgeries performed without cracking the sternum and opening the chest cavity. It

technological advance is an awe-inspiring engine of betterment and bounty. It is also an engine driving spread, creating larger and larger differences over time in areas that we care about—wealth, income, standards of living, and opportunities for advancement. Some of these trends (particularly rising inequality) are also visible in other countries. We wish that progress in digital technologies were a rising tide that lifted all boats equally in all areas, but it’s not.

economist Larry Summers put it, “suppose the United States had 30 more people like Steve Jobs—. . . . [W]e do need to recognize that a component of this inequality is the other side of successful entrepreneurship; that is surely something we want to encourage.”3

We particularly want to encourage it because, as we saw in chapter 6, technological progress typically helps even the poorest people around the world.

If the strong bounty argument is correct, then we have nothing significant to worry about as we head deeper into the second machine age. But is it? We wish that were the case, but it’s not. As we saw in chapters 9 and 10, the data are quite clear that many people in the United States and elsewhere are losing ground over time, not just relative to others but in absolute terms. In America, the income of the median worker is lower in real dollars than it was in 1999 and the story largely repeats itself when we look at households instead of individual workers, or total wealth instead of annual income. Many people are falling behind as technology races ahead.

Many Americans believe that they still live in the land of opportunity—the country that offers the greatest chance of economic advancement. But this is no longer the case. As The Economist sums it up, “Back in its Horatio Alger days, America was more fluid than Europe. Now it is not. Using one-generation measures of social mobility—how much a father’s relative income influences that of his adult son—America does half as well as Nordic countries, and about the same as Britain and Italy, Europe’s least-mobile places.”11 So the spread is not only large, but also self-perpetuating. Too often, people at the bottom and middle stay where they are over their careers, and families stay locked in across generations. This is not healthy for an economy or society. It would be even unhealthier if the spread were to diminish the bounty—if inequality and its consequences somehow impeded technological progress, keeping us from enjoying all the potential benefits of the new machine age. Although a common argument is that high levels of inequality can motivate people to work harder, boosting overall economic growth, the inequality can also dampen growth. In 2012 economist Daron Acemoglu and political scientist James Robinson published Why Nations Fail, a sweeping account of hundreds of years of history aimed at uncovering, as the book’s subtitle puts it, “the origins of power, prosperity, and poverty.” According to Acemoglu and Robinson, the true origins are not geography, natural resources...

For some goods and services, such as automobile tires or household lighting, demand has been relatively inelastic and thus insensitive to price declines.

Cutting the price of artificial light in half did not double the amount of light consumers and businesses demanded, so the total revenues for the lighting industry have fallen as lighting became more efficient.

Whole sectors of the economy, not just product categories, can face relatively inelastic demand. Over the years agriculture and manufacturing have each experienced falling employment as they became more efficient. The lower prices and improved quality of their outputs did not lead to enough increased demand to offset improvements in productivity.

On the other hand, when demand is very elastic, greater productivity leads to enough of an increase in demand that more labor ends up employed. The possibility of this happening for some types of energy has been called the Jevons paradox: more energy efficiency can sometimes lead to greater total energy consumption. But to economists there is no paradox, just an inevitable implication of elastic demand. This is especially common in new industries like information technology.21 If elasticity is exactly equal to one (i.e., a 1 percent decline in price leads to exactly a 1 percent increase in quantity), then total revenues (price times quantity) will be unchanged. In other words, an increase in productivity will be exactly matched by an identical increase in demand to keep everyone just as busy as they were before.

For instance, falling food prices might reduce demand for agricultural labor, but they free up just enough money to be spent elsewhere in the economy so that overall employment is maintained.22

KEYNES DISAGREED. He thought that in the long run, demand would not be perfectly inelastic. That is, ever lower (quality-adjusted) prices would not necessarily mean we would consume ever more goods and services. Instead, we would become satiated and choose to consume less. He predicted that this would lead to a dramatic reduction in working hours to as few as fifteen per week as less and less labor was needed to produce all the goods and services that people demanded.23

second, more serious argument for technological unemployment: the inability of our skills, organizations, and institutions to keep pace with technical change. When

The third argument for technological unemployment may be the most troubling of all. It goes beyond “temporary” maladjustments. As described in detail in chapters 8 and 9, recent advances in technology have created both winners and losers via skill-biased technical change, capital-biased technical change, and the proliferation of superstars in winner-take-all markets. This has reduced the demand for some types of work and skills. In a free market, prices adjust to restore equilibrium between supply and demand, and indeed, real wages have fallen for millions of people in the United States. In

In other words, just as technology can create inequality, it can also create unemployment. And in theory, this can affect a large number of people, even a majority of the population, and even if the overall economic pie is growing.

But in principle, machines can have very different strengths and weaknesses than humans. When engineers work to amplify these differences, building on the areas where machines are strong and humans are weak, then the machines are more likely to complement humans rather than substitute for them. Effective production is more likely to require both human and machine inputs, and the value of the human inputs will grow, not shrink, as the power of machines increases. A second lesson of economics and business strategy is that it’s great to be a complement to something that’s increasingly plentiful. Moreover, this approach is more likely to create opportunities to produce goods and services that could never have been created by unaugmented humans, or machines that simply mimicked people, for that matter. These new goods and services provide a path for productivity growth based on increased output rather than reduced inputs.

Thus in a very real sense, as long as there are unmet needs and wants in the world, unemployment is a loud warning that we simply aren’t thinking hard enough about what needs doing. We aren’t being creative enough about solving the problems we have using the freed-up time and energy of the people whose old jobs were automated away. We can do more to invent technologies and business models that augment and amplify the unique capabilities of humans to create new sources of value, instead of automating the ones that already exist. As we will discuss further in the next chapters, this is the real challenge facing our policy makers, our entrepreneurs, and each of us individually.

An Alternative Explanation: Globalization Technology isn’t the only thing transforming the economy. The other big force of our era is globalization. Could this be the reason that median wages have stagnated in the United States and other advanced economies? A number of thoughtful economists have made exactly that argument. The story is one of factor price equalization. This means that in any single market, competition will tend to bid the prices of the factors of production—such as labor or capital—to a single, common price.* Ov...

And indeed manufacturing employment in the United States has fallen over the past twenty years; economists David Autor, David Dorn, and Gordon Hanson estimate that competition from China can explain about a quarter of the decline in U.S. manufacturing employment.28 However, when one looks more closely at the data, the globalization story becomes much less compelling. Since 1996, manufacturing employment in China itself has actually fallen as well, coincidentally by an estimated 25 percent.29 That’s over thirty million fewer Chinese workers in that sector, even while output soared by 70 percent. It’s not that American workers are being replaced by Chinese workers. It’s that both American and Chinese workers are being made more efficient by automation. As a result, both countries are producing more output with fewer workers. In the long run, the biggest effect of automation is likely to be on workers not in America and other developed nations, but rather in developing nations that currently rely on low-cost labor for their competitive advantage. If you take most of the costs of labor out of the equation by installing robots and other types of automation, then the competitive advantage of low wages largely disappears.

This is already beginning to happen. Terry Guo of Foxconn has been aggressively installing hundreds of thousands of robots to replace an equivalent number of human workers. He says he plans to buy millions more robots in the coming years.

similar argument applies outside of manufacturing. For instance, interactive voice-response systems are automating jobs in call centers. United Airlines has been successful in making such a transition. This can disproportionally affect low-cost workers in places like India and the Philippines. Similarly, many medical doctors used to have their dictation sent overseas to be transcribed. But an increasing number are now happy with computer transcription. In more and more domains, intelligent and flexible machines, not humans in other countries, are the most cost-effective source for ‘labor.’

you look at the types of tasks that have been offshored in the past twenty years, you see that they tend to be relatively routine, well-structured tasks. Interestingly, these are precisely the tasks that are easiest to automate. If you can give precise instructions to someone else on exactly what needs to be done, you can often write a precise computer program to do the same task. In other words, offshoring is often only a way station on the road to automation.

In the long run, low wages will be no match for Moore’s Law. Trying to fend off advances in technology by cutting wages is only a temporary protection. It is no more sustainable than asking folk legend John Henry to l...

So as technology races ahead, will it leave a generation behind in all areas, or at least most of them? The answer is no. Even in those areas where digital machines have far outstripped humans, people still have vital roles to play. This sounds like a contradiction in terms; the game of chess shows why it’s not.

The teams of human plus machine dominated even the strongest computers. The chess machine Hydra, which is a chess-specific supercomputer like Deep Blue, was no match for a strong human player using a relatively weak laptop. Human strategic guidance combined with the tactical acuity of a computer was overwhelming. The surprise came at the conclusion of the event. The winner was revealed to be not a grandmaster with a state-of-the-art PC but a pair of amateur American chess players using three computers at the same time. Their skill at manipulating and “coaching” their computers to look very deeply into positions effectively counteracted the superior chess understanding of their grandmaster opponents and the greater computational power of other participants. Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.

We’ve never seen a truly creative machine, or an entrepreneurial one, or an innovative one. We’ve seen software that could create lines of English text that rhymed, but none that could write a true poem (“the spontaneous overflow of powerful feelings, recollected in tranquility,” as Wordsworth described it). Programs that can write clean prose are amazing achievements, but we’ve not yet seen one that can figure out what to write about next. We’ve also never seen software that could create good software; so far, attempts at this have been abject failures.

These activities have one thing in common: ideation, or coming up with new ideas or concepts. To be more precise, we should probably say good new ideas or concepts, since computers can easily be programmed to generate new combinations of preexisting elements like words.

Ideation in its many forms is an area today where humans have a comparative advantage over machines. Scientists come up with new hypotheses. Journalists sniff out a good story. Chefs add a new dish to the menu. Engineers on a factory floor figure out why a machine is no longer working properly. Steve Jobs and his colleagues at Apple figure out what kind of tablet computer we actually want. Many of these activities are supported and accelerated by computers, but none are driven by them.

Computers are not useless, but they’re still machines for generating answers, not posing interesting new questions. That ability still seems to be uniquely human, and still highly valuable. We predict that people who are good at idea creation will continue to have a comparative advantage over digital labor for some time to come, and will find themselves in demand.

Ideation, creativity, and innovation are often described as ‘thinking outside the box,’ and this characterization indicates another large and reasonably sustainable advantage of human over digital labor. Computers and robots remain lousy at doing anything outside the frame of their programming. Watson, for example, is an amazing Jeopardy! player, but would be defeated by a child at Wheel of Fortune, The Price is Right, or any other TV game show unless it was substantially reprogrammed by its human creators. Watson is not going to get there on its own.

Zara relies on its store managers around the world to order exactly, and only, the merchandise that will sell in that location over the next few days.