The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (9 page)

Waze makes GPS what it should be for drivers: a system for getting where you want to go as quickly and easily as possible, regardless of how much you know about local roads and conditions. It instantly turns you into the most knowledgeable driver in town.

The Economics of Bits

Waze is possible in no small part because of Moore’s Law and exponential technological progress, the subjects of the previous chapter. The service relies on vast numbers of powerful but cheap devices (the smartphones of its users), each of them equipped with an array of processors, sensors, and transmitters. Such technology simply didn’t exist a decade ago, and so neither did Waze. It only became feasible in the past few years because of accumulated digital power increases and cost declines. As we saw in chapter 3, exponential improvement in computer gear is one of the three fundamental forces enabling the second machine age.

Waze also depends critically on the second of these three forces: digitization. In their landmark 1998 book
Information Rules
, economists Carl Shapiro and Hal Varian define this phenomenon as “encod[ing information] as a stream of bits.”
⁴
Digitization, in other words, is the work of turning all kinds of information and media—text, sounds, photos, video, data from instruments and sensors, and so on—into the ones and zeroes that are the native language of computers and their kin. Waze, for example, uses several streams of information: digitized street maps, location coordinates for cars broadcast by the app, and alerts about traffic jams, among others. It’s Waze’s ability to bring these streams together and make them useful for its users that causes the service to be so popular.

We thought we understood digitization pretty well based on the work of Shapiro, Varian, and others, and based on our almost constant exposure to online content, but in the past few years the phenomenon has evolved in some unexpected directions. It has also exploded in volume, velocity, and variety. This surge in digitization has had two profound consequences: new ways of acquiring knowledge (in other words, of doing science) and higher rates of innovation. This chapter will explore the fascinating recent history of digitization.

Like so many other modern online services, Waze exploits two of the well-understood and unique economic properties of digital information: such information is
non-rival
, and it has
close to zero marginal cost of reproduction
. In everyday language, we might say that digital information is not “used up” when it gets used, and it is extremely cheap to make another copy of a digitized resource. Let’s look at each of these properties in a bit more detail.

Rival goods, which we encounter every day, can only be consumed by one person or thing at a time. If the two of us fly from Boston to California, the plane that takes off after us cannot use our fuel. Andy can’t also have the seat that Erik is sitting in (airline rules prohibit such sharing, even if we were up for it) and can’t use his colleague’s headphones if Erik has already put them on to listen to music on his smartphone. The digitized music itself, however, is non-rival. Erik’s listening to it doesn’t keep anyone else from doing so, at the same time or later.

If Andy buys and reads an old hardcover copy of the collected works of science-fiction writer Jules Verne, he doesn’t “use it up”; he can pass it on to Erik once he’s done. But if the two of us want to dip into
Twenty Thousand Leagues Under the Sea
at the same time, we either have to find another copy or Andy has to make a copy of the book he owns. He might be legally entitled to do this because it’s not under copyright, but he’d still have to spend a lot of time at the photocopier or pay someone else to do so. In either case, making that copy would not be cheap.
⁵
In addition, a photocopy of a photocopy of a photocopy starts to get hard to read.

But if Andy has acquired a digital copy of the book, with a couple keystrokes or mouse clicks he can create a duplicate, save it to a physical disk, and give the copy to Erik. Unlike photocopies, bits cloned from bits are usually exactly identical to the original. Copying bits is also extremely cheap, fast, and easy to do. While the very first copy of a book or movie might cost a lot to create, making additional copies cost almost nothing. This is what is meant by “zero marginal cost of reproduction.”

These days, of course, instead of handing Erik a disk, Andy is more likely to attach the file to an e-mail message or share it through a cloud service like Dropbox. One way or another, though, he’s going to use the Internet. He’ll take this approach because it’s faster, more convenient, and, in an important sense, essentially free. Like most people, we pay a flat fee for Internet access at home and on our mobile devices (MIT pays for our access at work). If we exceed a certain data limit, our Internet Service Provider might start charging us extra, but until that point we don’t pay by the bit; we pay the same no matter how many bits we upload or download. As such, there’s no additional cost for sending or receiving one more chunk of data over the Net. Unlike goods made of atoms, goods made of bits can be replicated perfectly and sent across the room or across the planet almost instantaneously and almost costlessly. Making things free, perfect, and instant might seem like unreasonable expectations for most products, but as more information is digitized, more products will fall into these categories.

Business Models When the First Copy is Still Expensive

Shapiro and Varian elegantly summarize these attributes by stating that in an age of computers and networks, “Information is costly to produce but cheap to reproduce.”
⁶
Instantaneous online translation services, one of the science-fiction-into-reality technologies discussed in chapter 2, take advantage of this fact. They make use of paired sets of documents that were translated, often at considerable expense, by a human from one language into another. For example, the European Union and its predecessor bodies have since 1957 issued all official documents in all the main languages of its member countries, and the United Nations has been similarly prolific in writing texts in all six of its official languages.

This huge body of information was not cheap to generate, but once it’s digitized it’s very cheap to replicate, chop up, and share widely and repeatedly. This is exactly what a service like Google Translate does. When it gets an English sentence and a request for its German equivalent, it essentially scans all the documents it knows about in both English and German, looking for a close match (or a few fragments that add up to a close match), then returns the corresponding German text. Today’s most advanced automatic translation services, then, are not the result of any recent insight about how to teach computers all the rules of human languages and how to apply them. Instead, they’re applications that do statistical pattern matching over huge pools of digital content that was costly to produce, but cheap to reproduce.

What Happens When the Content Comes Freely?

But what would happen to the digital world if information were no longer costly to produce? What would happen if it were free right from the start? We’ve been learning the answers to these questions in the years since
Information Rules
came out, and they’re highly encouraging.

The old business saying is that “time is money,” but what’s amazing about the modern Internet is how many people are willing to devote their time to producing online content without seeking any money in return. Wikipedia’s content, for example, is generated for free by volunteers all around the world. It’s by far the world’s largest and most consulted reference work, but no one gets paid to write or edit its articles. The same is true for countless websites, blogs, discussion boards, forums, and other sources of online information. Their creators expect no direct monetary reward and offer the information free of charge.

When Shapiro and Varian published
Information Rules
in 1998, the rise of such user-generated content, much of which is created without money changing hands, had yet to occur. Blogger, one of the first weblog services, debuted in August 1999, Wikipedia in January 2001, and Friendster, an early social networking site, in 2002. Friendster was soon eclipsed by Facebook, which was founded in 2004 and has since grown into the most popular Internet site in the world.
⁷
In fact, six of the ten most popular content sites throughout the world are primarily user-generated, as are six of the top ten in the United States.
⁸

All this user-generated content isn’t just making us feel good by letting us express ourselves and communicate with one another; it’s also contributing to some of the recent science-fiction-into-reality technologies we’ve seen. Siri, for example, improves itself over time by analyzing the ever-larger collection of sound files its users generate when interacting with the voice recognition system. And Watson’s database, which consisted of approximately two hundred million pages of documents taking up four terabytes of disk space, included an entire copy of Wikipedia.
⁹
For a while it also included the salty language–filled Urban Dictionary, but this archive of user-generated content was removed after, to the dismay of its creators, Watson started to include curse words in its responses.
¹⁰

Perhaps we shouldn’t be too surprised by the growth and popularity of user-generated content on the Internet. After all, we humans like to share and interact. What’s a bit more surprising is how much our machines also apparently like talking to each other.

Machine-to-machine (M2M) communication is a catch-all term for devices sharing data with one another over networks like the Internet. Waze makes use of M2M; when the app is active on a smartphone, it constantly sends information to Waze’s servers without any human involvement. Similarly, when you search the popular travel site Kayak for cheap airfares, Kayak’s servers immediately send requests to their counterparts at various airlines, which write back in real time without any human involvement. ATMs ask their banks how much money we have in our accounts before letting us withdraw cash; digital thermometers in refrigerated trucks constantly reassure supermarkets that the produce isn’t getting too hot in transit; sensors in semiconductor factories let headquarters know every time a defect occurs; and countless other M2M communications take place in real time, all the time. According to a July 2012 story in the
New York Times
, “The combined level of robotic chatter on the world’s wireless networks . . . is likely soon to exceed that generated by the sum of all human voice conversations taking place on wireless grids.”
¹¹

Running Out of Metric System: The Data Explosion

The digitization of just about everything—documents, news, music, photos, video, maps, personal updates, social networks, requests for information and responses to those requests, data from all kinds of sensors, and so on—is one of the most important phenomena of recent years. As we move deeper into the second machine age, digitization continues to spread and accelerate, yielding some jaw-dropping statistics. According to Cisco Systems, worldwide Internet traffic increased by a factor of twelve in just the five years between 2006 and 2011, reaching 23.9 exabytes per month.
¹²

An
exabyte
is a ridiculously big number, the equivalent of more than two hundred thousand of Watson’s entire database. However, even this is not enough to capture the magnitude of current and future digitization. Technology research firm IDC estimates that there were 2.7 zettabytes, or 2.7 sextillion bytes, of digital data in the world in 2012, almost half as much again as existed in 2011. And this data won’t just sit on disk drives; it’ll also move around. Cisco predicts that global Internet Protocol traffic will reach 1.3 zettabytes by 2016.
¹³
That’s over 250 billion DVDs of information.
¹⁴

As these figures make clear, digitization yields truly big data. In fact, if this kind of growth keeps up for much longer we’re going to run out of metric system. When its set of prefixes was expanded in 1991 at the nineteenth General Conference on Weights and Measures, the largest one was
yotta
, signifying one septillion, or 10
²⁴
.
¹⁵
We’re only one prefix away from that in the ‘zettabyte era.’

Binary Science

The recent explosion of digitization is clearly impressive, but is it important? Are all of these exa- and zettabytes of digital data actually useful?

They’re incredibly useful. One of the main reasons we cite digitization as a main force shaping the second machine age is that digitization increases understanding. It does this by making huge amounts of data readily accessible, and data are the lifeblood of science. By “science” here, we mean the work of formulating theories and hypotheses, then evaluating them. Or, less formally, guessing how something works, then checking to see if the guess is right.

A while back Erik guessed that data about Internet searches might signal future changes in housing sales and prices around the country. He reasoned that if a couple is going to move to another city and buy a house, they are not going to complete the process in just a few days. They’re going to start investigating the move and purchase months in advance. These days those initial investigations will take place over the Internet and consist of typing into a search engine phrases like “Phoenix real estate agent,” “Phoenix neighborhoods,” and “Phoenix two-bedroom house prices.”

To test this hypothesis, Erik asked Google if he could access data about its search terms. He was told that he didn’t have to ask; the company made these data freely available over the Web. Erik and his doctoral student Lynn Wu, neither of whom was versed in the economics of housing, built a simple statistical model to look at the data utilizing the user-generated content of search terms made available by Google. Their model linked changes in search-term volume to later housing sales and price changes, predicting that if search terms like the ones above were on the increase today, then housing sales and prices in Phoenix would rise three months from now. They found their simple model worked. In fact, it predicted sales 23.6 percent more accurately than predictions published by the experts at the National Association of Realtors.