4

The Expertise of Clouds

A Brief History of Experts

On February 3, 1986, President Reagan issued Executive Order 12546, establishing a Presidential Commission to report on why the Space Shuttle Challenger had blown up five days earlier, one minute and thirteen seconds after it took off.¹

The report begins with a calm, factual recounting of the disaster in intervals of a thousandth of a second: the first puff of gray smoke at 0.678 seconds; the “continuous, well-defined plume” of flame at 59.262 seconds; the “circumferential white vapor pattern” at 73.124 seconds; the Challenger “totally enveloped in the explosive burn” milliseconds later. The report concludes with recommendations in nine areas to remedy engineering faults, procedural flaws, and political pressures on NASA.

It is easy to see why this report is held in such high regard. The commission was headed by former Secretary of State William Rogers and included generals, physicists (including Richard Feynman), astronauts (including Sally Ride and Neil Armstrong), test pilots (Chuck Yeager), and rocket scientists. It took a broad look at the causes of the failure and produced an evidence-based document that led to needed improvements in NASA’s processes. The report saved lives. And it did so by embodying the very best of traditional expertise: A relative handful of highly trained and credentialed experts came together, followed a careful process, agreed on conclusions, wrote them down, and published them.

When Chaucer in the fourteenth century wrote in Troilus and Criseyde that those who are “expert in love” advise that it helps a man to have an opportunity to reveal his woe, to be expert was simply to be experienced.² The idea that being an expert could be a full-time, paid job rose with our culture’s increasing belief in science as a guide to social policy.

Some historians trace the rise of professional experts to a meeting held six months after the end of the Civil War,³ when one hundred reformers in various fields met in the Massachusetts State House and created the American Association for the Promotion of Social Science to advise their local communities and states about fixing everything from education to urban poverty, all based on the latest scientific research. ⁴ By the early 1900s, experts wielding “scientific management” techniques pioneered by Frederick Wilson Taylor—immortalized as the man with a clipboard and a stopwatch, timing the movements of workers—were sweeping through field after field.⁵ Even the home was now subject to the work of experts; as Ellen Swallow Richards, the founder of home economics and the first woman to get an engineering degree from MIT, wrote: “The work of homemaking in this scientific age must be worked out on engineering principles and with the cooperation of trained men and trained women.”⁶

Experts as full-time professional knowers needed professional institutions to support them. The first of these, the Brookings Institution, was founded in 1916, to provide policy advice to the government. By the 1950s, the Defense Department was relying on the RAND Corporation to help figure out questions of global life and death, including how nuclear war might be waged “successfully” and thus what types of bombs to build. RAND (the name comes from “Research and Development”) gave us our modern image of the expert, and he looked like Herman Kahn. The egg-shaped Kahn made a career by (as the title of his best-selling book put it) Thinking the Unthinkable: how to win a nuclear war. This hyper-rational approach could recommend a nuclear arms strategy because it would kill “only” 10 million people. Kahn entered public awareness as an egghead cheerfully detached from the carnage he contemplated, and was ridiculed in Dr. Strangelove, Or How I Learned to Stop Worrying and Love the Bomb, even as he wielded considerable influence on the Kennedy administration. Experts became cemented in our heads as highly rational, unswayed by personal and political considerations, and sometimes out of touch with lived reality.

In 1970, there were about two dozen think tanks. Today, there are over 3,500 worldwide, with about half of them in the United States.⁷ Our government has now relied on think tanks and their experts for a hundred years.

As the influence of think tanks on government policy increased, the cult of expertise spread to every corner of our culture. For example, after World War II, parents began relying on “parenting” experts, even though the experts disagreed among themselves. When should you potty-train your child? What are the effects of day care on very young children? What does a mild slap on the tush actually teach a child? A 2006 book that reviewed the advice of five top experts found little agreement even on such basic issues.⁸ Still, we continue to assume that it’s not only possible to be an expert at raising children but that experts can transfer that skill to us, often with just a few cogent mottoes.

Bringing smart people together is an ancient and effective technique for developing ideas. The Net also lets smart people connect and communicate. But the Net brings people together in new and occasionally weird configurations—a weirdness that is now being reflected in how expertise works....

From Crowds to Networks

If we were to diagram the way we think about the efficiency of the various ways we organize ourselves, until the past few years it would look roughly like a pyramid. At the bottom are mere crowds milling with multiple aims, like the crowds that pass by on a spring day in a city. As we go up the pyramid, the size of the social groups gets smaller, and their degree of organization increases. It’s for sure not a perfect pyramid—we can easily think of groups that are large and well-structured, or small and poorly structured—but the narrowing of the pyramid is based on the general truth of the observation that there is a type of social entropy, a waste of energy, when you try to get efficient results out of a large group. The organization can put in controls so that the whole sprawling collection of people works together, but it takes a lot of energy. Just ask your local army or multinational corporation.

This informal diagram of a social pyramid expresses a definite attitude. When we call people a “crowd” we mean that there is no additive social value to aggregating them. Worse, crowds have been associated with soul-deadening conformity and alienation in the age of mass communications. For example, in their 1950 best-seller The Lonely Crowd,⁹ David Riesman, Nathan Glazer, and Reuel Denney argued that America’s corporate culture was creating a generation of sheep who would conform in order to be socially approved. Even worse than a crowd is another social form at the bottom of the pyramid: a mob. A mob is a crowd stirred to the basest of actions. For example, an anti-draft crowd in New York City in 1863 turned into a mob when it threw stones, started fires, and looted. Before too long, it had lynched black men and torched the Colored Orphan Asylum on Fifth Avenue.¹⁰

So, it’s interesting that in the past few years we’ve grabbed onto the terms “crowd” and “mob” and applied them as positive characterizations of Internet sociality. Howard Rheingold’s Smart Mobs in 2003 applied the term to people connected through instantaneous digital communication,¹¹ and James Surowiecki’s The Wisdom of Crowds¹² in 2004 pointed to ways that unassociated groups of people can come up with more accurate answers than can individuals. Both books—each excellent—had titles that played upon our negative feelings about groups of people who are sharing space. “You see,” both books in effect said, “there’s a new positive potential in bringing big groups of strangers together.”

Smart mobs and wise crowds represent only two ways that knowledge can be developed on the Internet, often simply by being connected. Let’s look at five of the most basic properties of the Internet, starting with the simplest and moving to the more complex. Each of these is giving rise to its own types of networked expertise.

1. The Internet connects lots of people

The first and most obvious fact about the Internet is that it’s the biggest crowd anyone has ever seen.

Just as James Surowiecki says, there is a type of expertise that can come from people who are in the same place without being any further organized. The Wisdom of Crowds opens with what is now the canonical example. At a county fair in the eighteenth century it was noticed that if you wanted to know how much a particular ox weighs, the average of the total guesses of fair-goers was likely to be closer than the estimate of any particular expert. Surowiecki’s book is careful to lay out the precise conditions under which crowds do better than experts—it depends on there being a diversity of opinion, independence, decentralization, and a way to derive a collective decision—but almost as soon as he published it, “the wisdom of crowds” was used to refer to everything from choosing presidents, to the making of best-selling fashions, to voting for your favorite on American Idol. That we’ve so stretched this phrase shows just how excited we were about the new possibilities for social knowledge.

Knowledge always has been social. We clustered experts into think tanks and academic departments because we recognized that they’re smarter when together. In the eighteenth century, the great Western thinkers constituted what they called a “Republic of Letters,” in which they shared their ideas in correspondence, arguing back and forth at the speed of ponies and sailing ships. Even in ancient Greece, where the idea of knowledge was invented, the most famous thinker reached toward knowledge exclusively through dialogue with others.

But there used to be a natural size to such networks. Few people were admitted to the Republic of Letters, and it really helped to be a leisured white man. University departments are small enclaves. Books and then radio and TV are one-way media, and only a small group of people get to broadcast through them. Within those limits, we constructed a system of knowledge that concentrates expertise in a relative handful of people: If it costs so much to communicate to lots of people, we better give the microphone to those with the most expertise per square inch.

The Internet undoes those constraints. Its massiveness alone gives rise to new possibilities for expertise—that is, for groups of unrelated people to collectively figure something out, or to be a knowledge resource about a topic far too big for any individual expert.

The simplest forms are what Jeff Howe called “crowdsourcing” in a 2006 article in Wired.¹³ He intended it as a play on “outsourcing,” and his examples mainly were of “plugged-in enthusiasts” who will work for much less than traditional employees. But the term was so good that it quickly escaped its creator’s tether and now applies to just about any instance in which the mass of the Net lets us do things for little or no cost that otherwise would have been prohibitively expensive.

The examples of crowdsourcing are familiar at this point. When Members of Parliament were found to be routinely taking frivolous deductions, the British newspaper The Guardian set up a site where 20,000 people pored through 700,000 expense claims. When technologist Jim Gray and his sailboat went missing on the ocean, Amazon enabled people to scour thousands of satellite images for traces—fruitlessly, it turned out. The Internet enables us to gather and interpret information simply because the Internet is so damn big that you need only a tiny fraction of people to volunteer.

Sometimes the fact that the Internet covers so much physical ground is enough to create crowdsourced expertise. For example, in 2009, the Defense Advanced Research Projects Agency (DARPA)—the R&D branch of the US Department of Defense—decided to celebrate the fortieth anniversary of its creation of ARPANet, the precursor to the Internet. It moored ten 8-foot red weather balloons at accessible locations around the United States and offered a prize for the first accurate report of all of their locations.¹⁴ DARPA wanted to see if social networking sites such as Facebook, MySpace, and Twitter could serve as platforms for quickly gathering intelligence nationwide, a task with implications for national defense. About 4,000 teams entered to compete for the $40,000 prize. Within nine hours, a group at MIT had beaten them all simply by putting up a Web site that promised a substantial cut of the winnings to the first people to report the coordinates of the balloons and smaller cuts to the chains of people who had invited the finders onto the MIT team.¹⁵ What would have been extremely difficult for an individual to know turned out to be a snap for a network.

The popular technology site Engadget pooh-poohed the significance of this contest. “DARPA would have you believe that it’s the brilliance of modern-day social networks that led an MIT-based team to win its red balloon challenge this weekend” when it was just that MIT offered to split the money.¹⁶ That objection misses the point: Without the network, the offer of money would have gone nowhere.

Indeed, some of the most powerful ways to crowdsource expertise involve paying people. Amazon’s Mechanical Turk, launched in 2005, enables vast numbers of people to work on small, distributed tasks for a small amount of money per transaction. (It’s named after an eighteenth-century chess-playing “machine” that beat almost all comers, including Napoleon and Ben Franklin, by concealing a human chess expert within it.) Businesses have used Mechanical Turk to get thousands of online images labeled, find duplications in yellow-page listings, and rate the relevancy of a search engine’s results. As my colleague Jonathan Zittrain points out, there is a potential for abuse, such as someone using Mechanical Turk to engage the crowd in matching photos, never letting on that they’re being employed by an oppressive government to identify people at protest rallies.¹⁷ But whether Mechanical Turk is being used for good or evil, it is a type of networked expertise that previously would have been prohibitively expensive. Before the network made crowds so available, labeling millions of photos would have required hiring hundreds of professional cataloguers. The professionals at, for example, the digital image vendor Corbis do a more precise job, because they are trained and experienced. But the crowd does a good enough job, at a fraction of the cost. Further, the crowd can scale up to handle massive numbers of photos in a short period if necessary. And it is indeed arguable that crowd-based cataloging (or “tagging”) can produce systems of labels that more closely match how users think about matters.¹⁸

Most of the examples of crowds networking themselves into expertise do not involve money. Far more common are examples like that documented by Calvin Trillin in a light-hearted New Yorker article about food-enthusiasts’ attempts to track down Peter Chang, a Chinese chef who moves unpredictably from strip mall to strip mall.¹⁹ This informal network of “foodies” found each other through sites such as ChowHound.com, announcing sightings via email and blogs. The primary motivation for this networked collaboration was low on the Maslow scale: Chef Chang is a hell of a cook. But the overall process is so common now that it’s practically invisible to us: We rely upon work done by people who don’t know one another, spread out across a global network, who have answered a question, gathered data, refined results, contributed to a web of blogs, or even built an encyclopedia. And at the end of the day, the contributors to this networked expertise split a prize, whether it’s money, reputation, or “crispy eggplant cut like French fries and salt-fried with scallion greens, a hint of cumin, and hot pepper.”

2. The Internet has many different types of people in it

John Davis was a chemist in Bloomington, Illinois, who didn’t know a lot about oil. But it turned out that he knew enough about cement to solve a problem that had the oil experts stumped.

In 2007, the Oil Spill Recovery Institute, a nonprofit created by Congress in response to the 1989 Exxon Valdez oil disaster,²⁰ offered $20,000 to the first person who could figure out how to get the oil from that spill out of the bottom of the ocean where it had been sitting for eighteen years. Simply pumping it up did not work because when it got to the surface, the Alaskan air solidified the mix of oil and water, making it impossible to pump off the barges.²¹ But Davis knew that cement wouldn’t harden so long as you keep vibrating it. Perhaps if the oil were kept stirred up on the barge, it wouldn’t harden either.²² Problem solved. Davis spent part of his prize money to fly himself to the Alaskan site and has offered to work for the Institute for free on other projects.

Davis’s wisdom was not that of the crowd. But get a crowd large enough, and you’ll find experts like Davis—so long as the network contains many different sorts of people who know many different sorts of things, and so long as it has mechanisms by which the experts can be located. This is exactly the nature of the Net: huge and incredibly diverse. It enables a type of expertise just about impossible to actualize before the Internet existed.

Contests are one way to separate the individual experts from the crowd. It might be a $1,500 first-place prize for the undergraduate who writes the best paper about long-term pavement performance in a contest the Federal Highway Administration has run since 1988.²³ It might be the $500,000 the Department of Transportation has kicked in to the X-Prize for innovation in renewable fuels for the aviation industry. ²⁴ Or it might be the million dollars Prize4Life is offering through InnoCentive for finding a cure for ALS (Lou Gherig’s disease). In each of these cases, the network of experts has value only because that network contains many different types of people.

InnoCentive, a company that spun out of Eli Lilly in 2000, has become the leading contest-based expertise broker, typically offering $10,000 to $100,000 for solutions to problems posed by clients such as Procter & Gamble, NASA, Novartis, and the Rockefeller Foundation. Individuals from anywhere, with any degree of training, can submit solutions. For example, Ed Melcarek, who was earning a living installing HVAC systems, won $25,000 for figuring out a novel way of getting fluoride into toothpaste tubes for Colgate-Palmolive: Ground the tube and give a positive electric charge to the fluoride power.²⁵

As Karim Lakhani, a professor at Harvard Business School, has pointed out, such contests occurred well before the advent of the Net. In 1714, the British Parliament offered £20,000 to the first person who could come up with a way of determining longitude at sea.²⁶ John Harrison was finally granted the money in 1773, at the age of seventy-nine, after a literal lifetime of work on the project.²⁷ But before the Internet, it pretty much took an act of Parliament to spread word of a contest beyond the bounds of the experts in the field. Now, the Internet’s looseness with information has enabled contests to become a normal way in which knotty problems can be solved. For example, you might post to your social networking page news about a contest in your field, but that posting will be seen by people outside your field. Even if you lead such a shuttered life that you only have friends within your narrow range of expertise, they undoubtedly straddle some lines: In a world in which everyone is six degrees from everyone else, it’s at the second degree that things start to get really interesting. As news spreads from person to person, it sprays out across far wider networks. This is vital because, as Lakhani’s study of InnoCentive discovered, “the further the problem was from the solvers’ expertise, the more likely they were to solve it.”²⁸ In other words, the Net enables expertise to emerge not only because so many people are connected to it (property #1) but also because those people are different from one another in how they think and what they know (property #2).

InnoCentive is hardly the only place this happens. TopCopder has been running contests for software developers since 2001 and typically has over one hundred of them open simultaneously. In 2006 Netflix offered a million dollars to the best improvement greater than 10 percent in the algorithms it uses to recommend films to subscribers; in 2009 the team known as BellKor’s Pragmatic Chaos won for improving the recommendation algorithm by 10.09 percent.²⁹ And the Sunlight Foundation, a nonpartisan organization dedicated to increasing government transparency, ran the “Apps for America” contest in 2009 to encourage people to come up with innovative ways to extract value from the mass of agency data the Obama administration is making public. In 2010, it ran the contest again, with $15,000 going to the winner.

But contests are not the only way that nuggets of expertise can be found in the vast pebble field of the crowd. New York Times technology columnist David Pogue occasionally throws out a question to the over 1 million people following him on Twitter. In January 2009, to demonstrate the power of Twitter to an audience of 1,000 at a Las Vegas event, he famously asked his followers for a cure for hiccoughs. In less than fifteen seconds, hundreds of answers started rolling in, some serious, many not.³⁰ But he frequently asks more significant questions as well. For example, in January 2011, Pogue used Twitter to research Verizon’s iPhone coverage in New York City by asking his followers to tweet dead spots.³¹ He has also polled his followers for far more specific expertise: Anyone know how to save tweets to a text file? Any ideas for managing the ecological consequences of disposable gadgets? What were some of the old computers designed for use in kitchens? Now, Pogue can get his questions answered so quickly and richly because he has 1.3 million Twitter followers, but there have been sites on the Web for over a decade where people who are not famous can pose a question and receive answers from anyone in the crowd. Some medical-advice sites flag the answers from authenticated health care providers, while others rely upon either the crowd thumbs-upping or -downing answers or on the reputation the answerer has earned at the site. Quora, a site that began building a following in 2011, creates a sophisticated social network and reputation system to evaluate the probable worth of the answers volunteered there. In all these cases, the network comes up with answers not for money or to win a contest but because answering questions is a rich social activity that has a wide variety of rewards. The Net’s social nature in this case drives it as an information network.

My metaphor for this type of networked expertise as the finding of gems in a vast pebble field is not exactly right. Something else is happening. John Davis was not an expert in cleaning up oil spills until someone asked a question to which his own off-topic expertise suggested a particularly good answer. That probably would not have happened if there were no Internet and if the Oil Spill Recovery Institute had decided instead to post its challenge in its printed newsletter. Only because the question broke out of its tight circle of experts did John Davis become an expert in moving freezing-cold oil. The Net’s inability to fence in information—its proneness to “information spills”—let the right pieces connect, so a new idea was born. In cases like these, the Net, as a place that connects lots of people who are different from one another, is not only finding expertise but also generating it.

3. The Internet is like most oatmeal: sticky and lumpy

When a large food company wanted to know how to extend the shelf life of its mashed potatoes, they contacted YourEncore, a network of retired Procter & Gamble employees. The company engaged two retired experts in food microbiology who visited the processing plant, reformulated the product, and made recommendations for changes in production, sanitation, and quality control.

While architecturally the Net connects all nodes equally, socially it consists of billions of sub-networks. YourEncore is one of these: a social network of people joined by membership, mailing lists, and other network tools. Without these sub-networks, the Net would be just a large, flat resource. Instead, clusters naturally form, and it is these lumps of people, pages, and tools that bring the Net most of its value as a place for information, communication, and sociality.

Another network of people and resources, CompanyCommand.com, has an entry requirement even more exclusive than YourEncore’s: You have to be a West Point graduate. All graduates of this institution are given command of a company on the ground, but because they had no way of sharing their experiences and learning from one another, a handful of the graduates decided to build a network so they could talk among themselves. About five years later, the US Army embraced the idea fully, creating the Center for the Advancement of Leader Development and Organizational Learning (CALDOL) in 2005 on the West Point campus. What began as a mailing list is now an extensive Web site with discussion areas, multimedia educational materials, and the initial elements of a social networking site. Participants in this large but exclusive network ask practical questions about commanding units, pose challenging problems for group discussion, and talk about the personal issues common among soldiers away from home and sometimes in harm’s way.

This particular network is markedly different from the real-world social network in which the participants are embedded. As soldiers walk around West Point’s grounds, the markings on their shoulders make manifest exactly where they are in the hierarchy. But when participating in the CALDOL online expertise network, those markings are purposefully kept invisible; it’s considered bad social form to pull rank. What counts in the online network is the quality of one’s participation, not the marks on one’s sleeves. As a result, knowledge and expertise are shared and developed with fewer artificial impediments than one often finds in corporate and educational settings.

Social networks such as CompanyCommand and YourEncore add value to the Internet porridge because they comprise people with expertise that can be brought to bear on some particular class of problems. But, as is typical of the Net, nothing is that simple. Even the lumps can have lumps. That’s the case with Expert Labs, the White House–inspired network of members of the American Association for the Advancement of Science, the leading association of scientists in the United States. Expert Labs not only turns the 127,000 members of the AAAS into an expert network, it also opens the network up to anyone who wants to jump in with an idea or suggestion, for the same reasons that InnoCentive doesn’t demand that its entrants jump over a credentials hurdle. The professional scientists in the group provide skilled debunking and fact-checking, but the breakthrough ideas are at least as likely to come from people outside the traditional areas of expertise, as Karim Lakhani’s research into InnoCentive showed. Expert Labs’ first project was a service called ThinkUp that captures the back-and-forth conversations on your social networks at sites such as Facebook and Twitter so that you can slice and dice them, looking for trends, clusters of ideas, and insights. Such networks almost necessarily will include people outside a given scientist’s specialty.

Expert networks need lumps—networked conversations and social relationships among the diversely talented people (property #2) in the Internet crowd (property #1)—because expertise multiplies when it exists between people.

4. The Internet is cumulative

A man is turned back at the Canadian border because the customs agent Googles him and discovers something unsavory. A job applicant is all but hired when someone in the Human Resources department finds inappropriate Facebook photos. Political candidates routinely seek to embarrass their opponents by unearthing some dirt on the Web—information that the opponent had proudly posted back in the day when it seemed more funny than dirty. The Net retains everything we post to it—often out of context and sometimes against our will.

But it’s not just the dumb things that live on after we’ve wised up. There are reasons to think that the persistence of smart postings creates a new type of networked expertise. Indeed, you can see expertise accumulating just about in real time. When a new release of a computer operating system comes out, there will be lots of questions and not enough answers. Within days or weeks, however, it is likely that when something goes wrong with the software, you’ll be able to get an instant answer by searching for the text of the error message thrown onto your screen. As those answers accumulate, the Net gets more expert about the operating system. These days you’re better off waiting a couple of months before installing a new release of software, not because there will be a software fix but because by then the network will have gotten sufficiently expert in the workarounds.

The Net becomes more of an expert not just from the content people create for it but also from the links they—we—draw among the pieces. Linking curates the Net. Yet links are content, too. Indeed, one important type of expertise is being able to run the maze of links. The accumulation of links makes the accumulation of content on the Net ever more usable (because it can be found) and valuable (because a context grows around each piece of content).

The richness of the overall web of content and links enables a healthy mix of types of content-based networked expertise. Some of these collections are carefully vetted, such as medical sites put together by authorized health professionals. Some are process-based, providing functionality that invites people to contribute what they know and to evaluate what others have claimed to know. Some simply enable people to curate content developed elsewhere, linking to articles and debating their merits. And every flavor in between.

Experts, represented and embodied by their individual contributions, are joined by the Net into a networked expertise that can be queried and learned from, that comes up with surprising answers, and that grows faster than anyone can keep up with, because of the Net’s sometimes problematic insistence on retaining everything we put online.

5. The Internet scales indefinitely

Telephones work great one-to-one, they work okay with five people in a conversation, and they don’t work at all one-to-a-million, million-to-one, or million-to-a-million. Television works wonderfully at one-to-100-million but until very recently has not been feasible one-to-one, and still isn’t good at ten-to-ten multiway video conferencing. But the Internet works at every scale, and at many scales it allows for an unprecedented back and forth.

Consider Twitter. Twitter works perfectly if you and your five friends are following one another’s tweets. It also works if you’re Ashton Kutcher and you have millions of people following your every bon mot. It works if you have 100 followers, 10,000 followers, or 6 followers. Twitter works differently at every scale: If you have 6 followers, Twitter is an intimate communication tool, but if you have 1,000,000, it’s a broadcast medium. The only other medium that scales this way is paper. And even when the Net is being used for a 1-to-3,000,000 relationship, it enables a type of interactivity no other medium permits. Ashton Kutcher’s followers may not be able to communicate with Kutcher, but they can communicate with one another about Kutcher. Conversations may not always scale vertically on the Net, but they do scale horizontally.

With a mere 400,000 employees, IBM’s network is not nearly as large as Twitter’s, yet the “jams” the company began in 2001 illustrate the benefits that come from the easy way the Net enables many different types of groups to form. In a jam, all employees are invited to address a strategic question of the sort formerly only pondered by the company’s biggest brains and highest-paid consultants, using everything from email to wikis to collaborative projects. In 2003, the IBM jam updated the company’s core-values statement. In 2006, an “innovation jam” surfaced some big new ideas, five of which—including Smart HealthCare Payment Systems, Intelligent Utility Networks, and the Integrated Mass Transit Information System—have become core to IBM’s “smarter planet” initiative. The sessions tend to be creative, unstructured, and heedless of the corporate hierarchy. Jams have been so successful that companies around the world have adopted them, including Nokia, Eli Lilly, and an urban sustainability group that holds a continuing HabitatJam. Then, of course, startups and Web 2.0 companies began holding “jellies,” which are like jams but bring together multiple smaller companies.³²

Because the Net lets us form expert networks of just about any size and configuration, from twosomes to crowds to massively multiplayer games, the expertise of networks need not be equal to the expertise of its smartest member—or even cumulative. The complex, multiway interactions the Net enables means that networks of experts can be smarter than the sum of their participants.

For example, BellKor’s Pragmatic Chaos was able to win the Netflix prize because the Internet not only made it feasible to assemble experts from around the world but also made it possible for those experts to collaborate. While crowdsourcing can aggregate information—people in every neighborhood of New York City can report on what their local groceries are charging for diapers—networked experts who are talking with one another can build on what they know. We see this all the time on topical mailing lists. Whether the subject is stock picks, knitting, or rocket science (NASA has a social network called SpaceBook), these networks of experts can quickly become the first place received ideas are challenged and the first place new ideas are floated.

That’s why the life of the mind is quickly migrating to these networks of experts.

In May 1967, twelve scholars studying the difficult German philosopher Martin Heidegger gathered at Pennsylvania State University and founded the Heidegger Circle. It met once a year, carefully admitting new members by majority vote, until 1998 when it opened up to all who were interested. In the summer of 2005, the group went online, and in 2008 forums were set up.³³ Even though the forums are not all that active by Web standards, they nevertheless contain more people than the old Circle did, because now all that is required for membership is a payment of $35 ($15 for students and unemployed academics). Judging from how people sign their messages, it seems that most of the participants are professional academics, and the discussions are often about topics of peculiar interest to scholars: What did Heidegger mean by the “transcedens” in a particularly obscure paragraph in one of his early works? Was a particular lecture ever published in German? The forum serves—and expresses—the needs of an inner circle of scholars. Because it’s online and available 24/7, it enables this self-selected group of eighty-eight to get answers and explore topics just not conceivable in the old 1/365 days of the Heidegger Circle.

Meanwhile, a professor in Costa Rica set up a Facebook page for people interested in Heidegger.³⁴ It has 1,400 subscribers and is lively with discussion. Clearly not everyone there is an expert, and not everyone focuses full-time on Heidegger. But the discussion is animated, continuous, and unconstrained. It spills out into links to articles that have their own discussion threads. For example, one member links to an article in the Chronicle of Higher Education³⁵ and writes: “Thought you might find this interesting.... A new book bashing Heidegger for his Nazi period.... Even more interesting is the almost 100 comments defending Heidegger and/or ripping the article’s author.” Among those hundred comments are deep reflections, studied fact-based arguments, juvenile baitings, and links to yet more discussions and debates. The Heidegger Circle of the 1970s was a closed circle. The Heidegger network of the Web is always connecting to the next new site or idea.

The old Circle had advantages. It was an enclave where you could get productive work done by being with those who shared your basic assumptions. The information exchanged was quite likely to be accurate, the ideas well-considered. And it was an honor to be admitted. But the wall around the club insulated the members from criticism and from outsider points of view that could have helped them. The new network is lively and sprawling, but it contains people who know much less than they think, and can grind important topics down into moist dust.

We don’t have to choose between them. Both have value. The Circle is a lump of qualified, sober experts. The Facebook page is a big, throbbing lump of people who want to talk about Heidegger for whatever reason. The two together form a loosely connected network of people who care about Heidegger. The participants collectively know more, they find answers faster, their curiosity is more stimulated, they are made aware of more facets of their topic, and they are involved in more discussions about those facets. The multi-way nature of the Net enables smart experts to be smarter than ever, although it’s clear that the Net can also enable us to go down wrong paths with ever more certainty.

Nevertheless, this is a significant and palpable change. Those who suffered from Smartest Guy in the Room syndrome are learning that the rules have changed. When an expert network is functioning at its best, the smartest person in the room is the room itself.

Expertise That Looks Like a Network

The Challenger Commission remains a model for approaching a particular type of problem: Use the power and prestige of the office of the presidency to pull together an elite team of the very best in their fields, fund them heavily, relieve them of some of their ordinary duties so they can devote sufficient time to the project, and give them all the resources a Presidential Commission needs.

But it is not a model that scales. The various forms of networked expertise we have just looked at, each reflecting an essential property of the Net, are enabling new models of expertise to emerge—even within quite traditional sectors.

“Mitre’s role is providing expertise to our government clients,” says Michal Cenkl, director of Innovation and Technology.³⁶ To do so, the MITRE Corporation, with a Defense Department pedigree going back to its founding in 1958, certainly could take the Presidential Commission route, bringing together a tight circle of highly competent and highly credentialed experts to produce a final report. Instead, explains Cenkl, “[t]he products we’re delivering to our sponsors, in some cases, have evolved from formal written publications—MITRE tech reports—to formats that are more timely and interactive, such as email exchanges, in-person briefings, presentations and discussions, which are much harder to codify and capture.” Why? For the sake of speed, but also because the materials that are “harder to codify” represent an interactive dialogue, have a shared context at a specific point in time, and are characterized by a richness of ideas, information, and knowledge. Codifying them would in fact lose some of that context and richness.

“True enough, we have internal experts we can draw on,” says Cenkl, “but we’ve also realized that we need to change the way we present ourselves. It’s not necessarily that the MITRE person is the smartest person in the room. We’ve decided that the model needs to evolve so that we become the brokers of expertise. Our value is that we understand the government’s problems really, really well and we can bring the entire community to bear.”

Jean Tatalias, MITRE’s director of Knowledge Management, says: “We have an environment in which everyone can publish,”³⁷ instead of designating people as experts and rationing their time. MITRE’s internal search engine lets users search for nondesignated experts who nevertheless have been contributing to discussions. “You can see what forums are discussing the topic and what forum to put your question out to,” she says. “No one is anointed as the only credentialed expert in a field.”

In fact, “[t]he forums don’t necessarily come to consensus. There’s a desire to use all the expertise available but not a pressure to drive to the right answer for all circumstances,” adds Les Holtzblatt, chief Work Practices architect.³⁸ Why? Because networks of interacting experts are smarter than the accumulation of individual expert opinions whether we’re using a simple mailing list or a more highly structured knowledge management community. This is a fundamental distinction not only in the way expertise is derived but also in its nature: MITRE, which is in the expertise business, finds it often delivers more value to its clients when it involves them in a network of experts who have differing opinions.

Of course, experts have always profited by being in communication with one another. But now that expertise is embedded in—and enabled by—a digital network, it is shedding properties of the old medium and taking on properties of the new one:

 

Expertise was topic-based. Books focus on specific topics because they have to fit between covers. So, in a book-based world, knowledge looks like something that divides into masterable domains. On the Net, topics don’t divide up neatly. They connect messily. While people of course still develop deep expertise, the networking of those experts better reflects the overall truth that topic boundaries are often the result of the boundaries of paper.

 

Expertise’s value was the certainty of its conclusions. Books get to speak once. After they’re published, it’s expensive for the authors to change their minds. So, books try to nail things down. But because the multitude of people on the Internet are different in their interests and abilities, a network of experts is of many minds about just about everything. The value of a network of experts can be in opening things up, not simply coming to unshakable conclusions.

 

Expertise was often opaque. While experts’ reports usually tell us how their conclusions were derived, and typically include supporting data, we don’t expect to be able to go back very far in the experts’ thinking: The reports and their included data have been our stopping points. Networked expertise puts in links to sources—and even to contradictions—as a matter of course. A simple search is likely to turn up contextualizing information about the expert and about the information the expert is relying on.

 

Expertise was one-way. Books are the original form of broadcasting, a one-to-many medium: The reader can write a big red “No!!!” in the margins, but the author will never know about it. The Net, on the other hand, is multi-way. Any expert who thinks she will talk and we will simply listen has underestimated the Net. We will comment on her site, and if she doesn’t permit comments, we will angrily note that fact on blogs, on Twitter, on Facebook. This multi-way interactivity can make a network of experts more creative, and more responsive to the multitude of ideas and opinions in the world. It can, of course, also create and propagate misinterpretations of the expert’s ideas.

 

Experts were a special class. Relatively few people get to publish books. To do so, you have to pass through editorial filters. Because those filters have generally done a good job, getting a book published both required and bestowed credentials. It might, for example, have gotten you tenure. On the Net, we find expertise emerging from contexts that at first seem unrelated. The person who figures out how to clean up an oil spill may be a cement expert. On the Net, everyone is potentially an expert in something—it all depends on the questions being asked.

 

Expertise preferred to speak in a single voice. Books have authors and editors who ensure the content is self-consistent. Even anthologies have editors who ensure that there is appropriate consistency among the contributors, at least in terms of content and tone. Experts, too, have been self-consistent; they get embarrassed if caught contradicting themselves. Networked expertise is more like a raucous market of ideas, knowledge, and authority.

This transition from expertise modeled on books to expertise modeled on networks is uncomfortable, especially now as we live through the messy transition. We know the value of traditional expertise. We can see a new type emerging that offers different values. From credentialed to uncredentialed. From certitude to ambivalence. From consistency to plenitude. From the opacity conferred by authority to a constant demand for transparency. From contained and knowable to linked and unmasterable.

Most of all, expertise is moving from being a property of individual experts to becoming a property of the Net. Not all networks raise the level of connective expertise, of course. Some networks are in fact dumber—and more insistent about their dumb views—than its smartest members are. It all depends on the network’s particular constellation of the two basic Net properties: The Net is connective, and it connects pieces that are different from one another. Connectedness is nothing new for experts, although the scope, scale, and transparency of Net connectedness certainly matter. It’s the connectedness of a trillion differences that is really shaking up expertise. We paid the big-time experts all that money because we expected them to take the differences buzzing around our brains like gnats and blow them away once and for all. But networked expertise draws its strength from those differences in connection. Indeed, on the Net, the measure of one’s strength as an expert often is not that you have the final word on some topic but that you have the first word. And from that first word—whether it’s on a blog post, a tweet, or a sheet of old-fashioned white paper—spin out a million gnats of difference, buzzing across the linked world, unsettled and unsettling.

For networked knowledge necessarily contains differences.