The Age of the Unthinkable (29 page)

Antiretrovirals, by contrast, were handled differently. As a patient, before you ever received a pill you had to agree to
sit through an education program with a friend or family member who guaranteed in writing that you’d stay on the drugs. You
had to attend support-group meetings, disclose the disease to any sex partners, and return immediately to the hospital if
you developed certain symptoms. The pill prep classes included discussions of biology, explanations of the human immune system,
and lessons about how the drugs interacted with each other and with what you ate and drank. This program of disease literacy
had been pioneered, not only by public-health Ph.D.s, but also by unemployed rural women and infected South African city dwellers,
usually patients themselves, partly because they didn’t trust the drug companies or the doctors. They wanted to know what
they were putting into their bodies. But they were also betting that their fellow infected citizens were smart enough to manage
the disease and that most — the vast majority of whom had never completed high school — could be taught science and that such
knowledge would serve not just as context but as motivation. ARV treatment literacy let patients manage far more of their
own care than any outside expert would have guessed possible. And this changed the whole dynamic. ARV patients stuck with
their pills.

The gap between the success of the ARV program and the failure of the TB program revealed a lesson that applies almost everywhere
in our world now: the moment you hand power over to other people, you get an explosion of curiosity, innovation, and effort.
This is a law as true for commodity markets (where it works to disastrous effect when we don’t watch it carefully) as for
disease treatment. And the universality of this law, for good or ill, is what makes it so curious and useful. It’s not wrong
to think of what’s going on in our world as a race between forces that are unthinkably amazing and those that are unthinkably
horrifying. For example, the very same technology that uses genetics to cure heart disease can create lethal biological weapons.
As we move to build a really secure world, it’s time for us to begin investing — heavily — in putting more runners for the
force of good on the track. Moll’s HIV patients are a reminder that, given encouragement and opportunity, people will run
faster than we would ever have imagined.

When you spread power instead of hoarding it, you discover benefits that you couldn’t have imagined in advance — and that
sometimes run contrary to what you might have expected. Military historians have studied the bewildering efficiency of armies
that swarm like bees or ants, highly decentralized groups that bend, adjust, and attack based on a far better sense of local
conditions than any central commander could ever have. This form of warfare flew in the face of centuries of command-and-control
warfare theory, but it was what made German U-boats so effective, even though they were often out of touch with their bases
for days at a time. It’s what made Mongol horse swarms so lethally efficient, and it has marked winning revolutionaries as
different as Mao Zedong and Emiliano Zapata. Making this sort of decentralized approach work was the job of tech geeks who
populate the ranks of groups like Al Qaeda and Hizb’allah. “Hizb’allah uses a swarming approach to deal with Israeli commando
raids in southern Lebanon,” RAND scholars John Arquilla and David Ronfeldt wrote in a landmark paper; “This approach is based
on a general instruction to Hizb’allah’s widely distributed units to converge — like antibodies, it seems — on any intruders
in a given area. No central leadership is required.” Swarming is, of course, the classic immune-system response. It’s what
happens when your blood clots after you slice your finger cutting cucumbers, and it’s what’s going on in your sinuses when
you sneeze.

This kind of self-organization, the ability to pull off an “all hands on deck” reaction, exists in many of the most efficient
and resilient systems in our world. This has been a marked feature of life in an information age, when e-mails, telephone
calls, and text messages have diminished the effects of geography, put people in closer direct contact, and, in the process,
removed the need for much central command and control. Take, for example, what economists call “peer production,” which is
the previously unimagined economic twitch for sharing work that has built Wikipedia, file-sharing systems like BitTorrent
(which now accounts for at least 50 percent of all Internet traffic), or “open source” operating systems like Ubuntu and Linux.
“Peers” can be producing anything from decisions to software, but what matters is that these efforts are largely bottom-up,
which, strangely, makes them more efficient rather than less.
A bunch of chaotic hackers doing as good a job as
Encyclopaedia Britannica? Well, as
Nature
magazine discovered when it compared the accuracy of Wikipedia and
Britannica,
yes.

Once users step into active engagement, the dynamics of the system shift forever: users stop being consumers and become participants.
This pushes the opportunity for innovation to the edges of a network, where users reside, instead of leaving it in the hands
of some slow-moving, committee-oriented, centralized manufacturing center. This was, for instance, how the whole mountain-biking
craze got started, as Eric von Hippel has written in his masterful study
Democratizing Innovation
. A few trail-riding gearheads intent on customizing their own bikes began developing better, more suitable bikes from old-style
street bicycle frames — swapping tips and information with each other as they went and accelerating the creation of a new
sport. Von Hippel calls this “user innovation,” and it resulted in both faster change and products that better suited the
way bikers aspired to ride. It cut, in essence, the distance between “hope” and “product” to nearly zero.

Peer producers behave in ways that traditional economics says is absurd, just as public-health experts might have laughed
at the idea that illiterate women could manage a complex medical training program. For instance, peer producers will work
long hours for no financial reward, only for the sheer pleasure of making a better piece of software or a cooler mountain
bike brake. They will renounce the right to own or profit from what they invent. Two hundred thousand people working together
to create and edit Wikipedia and then giving it away for free (in fact, working so hard precisely because it
is
free) is something no standard approach to economics can explain — just as no standard health-care approach explains the
effectiveness of those lifesaving techniques to fight AIDS. It’s proof that once you give power away all sorts of unplanned
efficiencies — boosts that are invisible to our standard way of seeing — emerge.

Peer production solves one of the great troublesome stumbling blocks to progress: that a few people usually control the most-vital
information. Instead it creates an open system, one not so different from the one South African patients created when they
started mimeographing their own home-made ARV guidelines. And there’s a bonus here — what economists call a “return to scale”
— that is like a dividend you get only when lots of people work together. The more users a centralized system has, the closer
it comes to exhaustion. (Think of a mother with five preadolescent kids or an overloaded telephone switchboard operator.)
But the more users a decentralized system has, the more efficient it becomes, since work can be spread around or picked up
by whoever can do it best and fastest. The average time between vandalism of a Wikipedia page and its repair is now less than
five minutes. The average time to fix an error in that version of the
Encyclopaedia Britannica
you have on your shelves is, well, never. It’s already set in print. (And even if you bought a new edition, you’d have to
wait years between versions.)

There’s another impressively useful trait of these systems that’s vitally important: they are much, much more efficient because
often they manage to put excess capacity to work that would otherwise be wasted — the underutilized engineering skills of
bikers, for instance, or the drawing and teaching skills of AIDS patients. The classic example of this phenomenon was a 2000
NASA project called Clickworkers, which let random users log in to a Web site and mark out craters and channels on photos
of the surface of Mars. This was the sort of geographic work that was usually done by experienced grad students. But thousands
of volunteer workers, often clicking for just a few minutes at a time during a coffee break, were as accurate as and far faster
(and cheaper) than a few grad students. NASA said it wanted to answer two questions: Would the public lend a hand to help
science, and would the results be as good as a traditional approach. The answer was yes on both counts.

Think of how much excess capacity lingers in our global system, from unused ideas to misallocated savings to under-employed
academics or researchers in countries where information is blocked or embargoed. (Frankly, think of how much unused capacity
lingers in your life or mine; we’ll turn to this in a moment.) There is more potential power for action in our world system
now than ever before in history — and also more need for action. What Jean Monnet, one of the most masterful strategists of
the last century, wrote in 1951 is true today: “World peace cannot be safeguarded without the making of efforts proportionate
to the dangers which threaten it.” The more peers we can bring online for the business of saving the world, the easier the
effort will be, and, in a sense, the stronger we’ll each be. As the Yale law professor Yochai Benkler wrote in “Coase’s Penguin,”
an essay that set out the virtues of this fresh economics in 2002: “The advantages of peer production are, then, improved
identification and allocation of human creativity. These advantages appear to have become salient, because human creativity
itself has become salient.” As we’ve seen, our future is a race between good innovation and bad innovation. That’s a sprint
that will be decided purely by our ability to create. It’s a shift so profound that it evokes the ideas of the American philosopher
John David Garcia, who once said that we should reject the notion that increasing human happiness is the most important goal
for society. Far better, he said, to increase human creativity. Happiness will follow.

We owe everything to human creativity. Everything that lasts, that changes our lives, that emerges from what was once unimaginable
has its roots in that initial spark of innovation. What we need now is a way to spread the capacity for peer production of
our own future. We need a global order in which those “irrational” forces of invention and surprise, forces that don’t fit
foreign-policy models — any more than free labor fits economics — can be used to improve the parts of the world in which we
can’t seem to make progress. In short, we need a global version of that distributed ARV approach. Sure, it may seem irrational
that uneducated dying Africans could save themselves. But in practice they do just fine. Our standard models don’t know how
to make sense of these kinds of productivity miracles — which is why those models will never be able to save us.

As I sat with Moll in Tugela Ferry one afternoon in 2006, talking about XDR-TB, I was reminded of a friend of mine, who was
building robots long before that was cool (assuming, of course, that it
has
actually become cool). In the early 1990s my friend was trying, and failing, to build a machine that would vacuum floors
without human input. This problem occupied her for years, and it was interesting because solving it would open up all sorts
of other possibilities, such as how you might robot-clear a minefield. The challenge was how to teach the robots about the
millions of different rooms they would need to vacuum. Old-style artificial intelligence suggested this would involve finding
a way to jam a map of every room in every house into the machine — the checklist approach. You’d need to be able to observe
every room, the way those doctors observed every TB patient. It was impossible. But then my friend had a breakthrough: what
if the robot could learn the map of the room by itself, without any preprogramming? It could do this by scooting around and
gently bumping into walls, chairs, and sofas. Such a machine would be kind of stupid, actually. It would require only the
barest amount of intelligence. But even in that state it would work in a way that the smartest centralized system never could.
Once she and her engineers made this leap, they were able to develop a self-directed vacuum cleaner, the Roomba, which became
a huge commercial success. The lesson of Roomba was that a slow-thinking, wall-bumping, but
empowered
vacuum cleaner was smarter than a room full of Ph.D.s. If only those South African health officials had understood this, a
great tragedy might have been avoided. And, of course, suppressing the devastation of AIDS in Africa has major foreign-policy
implications: The genius of what happened in Tugela Ferry wasn’t that it happened in Tugela Ferry. It was that it could happen
anywhere.

In a way, what I am about to propose as the last part of our deep-security immune system, the revolutionary spread of power,
will probably seem inevitable at this point. But it’s worth stopping for a moment to remark how far we have come from that
older hat-on-the-floor picture of our world at the start of this book. We’ve now left behind the idea that only states matter,
that the futures of nations should be determined only by looking at leaders in rooms, that we can make policy as if we were
making a cake: just add democracy or capitalism and get prosperity. We’ve changed how we look at the world, begun to obsess
about resilience, started to see threats as systems, not objects. We’ve seen practical ideas about how such deep security
might work in practice, changing how we create Middle East peace, for instance, or how we confront Iran, nuclear proliferation,
and complex nations such as China. And now, as a last adjustment, we will move to the extreme opposite of the old way of looking.
Instead of worrying about big objects (states) and expecting them to be predictable, instead of obsessing about heads of state
or terrorist leaders, what I want to propose is that we focus our attention also on the very smallest parts of the system,
on
people,
and bet that the one thing we know for sure is that we can’t predict what they’ll do. In other words, the last step to deep
security in a world of unthinkable granular surprise is to push — as hard as possible — for even more unthinkable granular
surprise.