War: What is it good for? (66 page)

Nexus
and
Crux
are only stories, but they nicely capture the messiness of merging with and through machines and the complexity of the choices ahead. If, for instance, the globocop leans too far forward—say, by trying to control developments too tightly, or by holding on to its job past the culminating point—it will face mounting opposition, overstretch, and financial collapse, quite possibly bringing on precisely the military challenges that it will be trying to avoid. That is a surefire strategy for losing the endgame of death, and one of the reasons I spent so much time in earlier chapters looking at the theory of a Western way of war is that it seems to encourage just this kind of overconfidence about leaning forward. Thanks to the military legacy inherited from ancient Greece, Victor Davis Hanson assures us, “deadly Western armies have little to fear from any force other than themselves.” But this, I argued, is not what long-term history shows. In fact, as the twenty-first century goes on, it will be non-Western armies that challenge the globocop most. Maintaining order will depend on sound judgment and skillful shepherding of resources, not the legacy of ancient Greece.

On the other hand, if leaning forward too far or too long will lose the endgame, pulling back too far or too soon will do so even faster. If the globocop goes absent without leave, the most relevant analogy for the coming years might be not the slowly mounting crises of the 1870s–1910s, but the abrupt catastrophe of the 1930s—that low, dishonest decade when the British globocop lay dying, Americans were unwilling to take its place,
and reckless rivals gambled everything on violent solutions to their problems. In the long run, pulling back will be essential, but in the short run it will be catastrophic.

Everything will hang on the relative timing of the shift from the Pax Americana to a Pax Technologica and the mounting difficulties that the globocop will face—if current economic trends continue—in doing its job. I suggested earlier that in the 2010s and probably the 2020s too, the United States will remain largely unchallenged, but as the 2030s, 2040s, and 2050s go on, it will find it harder and harder to overawe rivals. I also noted that the majority opinion among the futurists is that merging with machines will reach the Singularity stage in the 2040s. If all of these guesses are right, we perhaps do not have too much to worry about. The world will become increasingly troubled, polarized, and tense as we head through the 2020s, but the globocop will remain strong enough to handle the stresses. As we enter the 2030s, the globocop will be feeling the strain, but it will by then be pulling back anyway as the Pax Technologica begins to make violence irrelevant to problem-solving; and in the 2040s and '50s, just at the point that the globocop ceases to be able to cope, the world will no longer need its services. All will be well.

But will the computerization of everything really proceed at this convenient pace? The 2040s are only thirty years away, and although the thirty years that have just passed saw dramatic technological changes, it is far from obvious that another three decades will merge us with our machines. But that misapprehension, the futurists insist, comes from failing to see that technological change is exponential, constantly doubling up, not linear. Imagine, they sometimes say, that you rent a summer cottage. When you arrive, there is a very pretty lily on the pond. A week later, there are two; another week after that, there are four. You then, reluctantly, have to go back to work, and two months pass before you can resume your vacation. When you get back to the cottage, more than a thousand lilies confront you. The four lilies you left have doubled and redoubled eight times; you can no longer see the pond under them.

Let us say, for the sake of argument, that one full lily's worth of technological transformation had happened by 1983, the year of Petrov's moment of truth, and that each lily reproduces once every half-dozen years. In 2013, as I write, the lilies have doubled five times, and we have 32 of them—a lot more than in 1983, but still far from filling the pond. By 2025, however, there will be 128, and by 2043—the eve of Kurzweil's date for the
Singularity—over 1,000. The original pond—that is, we unimproved, merely biological humans—will have disappeared under a carpet of trans- and posthuman techno-lilies.

The thirty-some lilies we have in the mid-2010s represent gadgets like Google Glass, the Internet, and rats moving each other's paws. These are nice additions to the ways humans have lived for the last fifty thousand years, but nothing more. The two hundred or so lilies of the late 2020s might add up to artificial intelligence that can sometimes pass for human, a touch of telepathy, and some people living their lives largely in virtual reality, but there will still be far more pond than lily. The knee of the curve, as statisticians call the point where the increases really take off, will come in the mid-2030s, by which time every year will see more change than happened in the whole period between the 1980s and the 2010s; and in the 2040s, when change becomes so rapid that it appears to be instantaneous, the globocop can retire.

The arithmetic only adds up, though, if the exponential improvements in computing power continue at the same speed as they have in the last fifty years—but that would break Smalley's Law, the premise that everything is possible but will take longer than we think. If Smalley's Law does apply to the computerization of everything, we might still be very far from finishing the endgame of death when the globocop loses its grip in the 2040s. Even a modest increase in the time techno-lilies need to reproduce, from the half-dozen years that I proposed to a decade, would push the knee of the technological curve back to the 2060s and delay any kind of Singularity into the 2080s.

If the globocop stumbles in the 2040s, a period of several decades will then ensue with neither a Pax Americana nor a Pax Technologica. Rather than merging into a single superorganism, a Smallean world might dissolve in the 2050s into multiple and incompatible brain-to-brain networks, each dominated by a different great power. We might see a high-tech version of the nineteenth-century scramble for Africa as the networks compete for neural market share, shutting their rivals out of different parts of the world. Climate change might by then be convulsing the arc of instability, the coming of killer robots might be shifting the balance of power, and the infrastructure and energy needs of merging with machines might be providing a whole new kind of target to attack. A nation that feels it has a temporary edge in the technological transformation might be tempted to gamble on using it to impose its will violently on everyone else, or, perhaps more likely, a government that is falling behind might go for
broke, betting everything on attacking before the enemy's lead becomes unassailable.

Armageddon will be beckoning.

War! What Is It Going to Be Good For?

But that, I am confident, is not how the story will end.

The reason for my optimism is our track record, revealed so clearly by long-term history. We have not managed to wish war out of existence, but that is because it cannot be done. We have, however, been extremely good at responding to changing incentives in the game of death. For most of our time on earth, we have been aggressive, violent animals, because aggression and violence have paid off. But in the ten thousand years since we invented productive war, we have evolved culturally to become less violent—because that pays off even better. And since nuclear weapons came into the world in 1945, the incentives in the game have changed faster than ever before, and our reactions have accelerated along with them. As a result, the average person is now roughly twenty times less likely to die violently than the average person was in the Stone Age.

Imagine, for a moment, that I had written this book fifty years ago, publishing it not in 2014 but in 1964—less than three years after the Berlin crisis, two after the Cuban missile crisis, a few months before Mao tested his first atomic bomb, and a year before U.S. marines landed in South Vietnam. Imagine too that I had predicted in it that humanity was now so well attuned to changes in the game of death's payoffs that within twenty-five years the Soviet Union would renounce force, tear down the Berlin Wall, and then dissolve itself, all without firing a single shot, let alone a nuclear missile. Even if I had held myself back from speculating that Red China would embrace capitalism and turn into the world's second-biggest economy, I doubt that the reviewers would have been kind. But I would have been right. And now, back in the present, the same reasoning leads me to believe that we will play the endgame of death just as skillfully as we played the regular game.

What we need to do is simple but, as Clausewitz said, difficult, because humanity will only win the endgame of death if the Singularity arrives before the globocop fails. If we are really going to get there from here, the globocop must remain strong for as long as possible—which means that the United States must, for the next forty-plus years, maintain its military spending and readiness at levels that make it a credible Leviathan. It must
be ready to threaten and even use force to preserve the global order, while neither spending so much that it breaks the political consensus in favor of leaning forward nor exploiting its advantages so aggressively that it alienates its allies. To meet all these challenges, Americans will need to get their financial house in order, sustain economic growth, and invest in basic science, all while continuing to find leaders of the same quality as those that carried the country through the Cold War. Simple, but difficult.

The faster the computerization of everything proceeds, the greater the likelihood that the Pax Americana turns into a Pax Technologica before the globocop's weakness leads to a new storm of steel. But even in the worst-case, Smalley's Law–type scenario, the United States must remain as ready to pay any price, bear any burden, and meet any hardship as it was when John F. Kennedy first recommended such a course in 1961. In September 2013, as this book goes into production, two-thirds of Americans are telling pollsters that they oppose any use of force in Syria; but if the United States (like Britain between the World Wars) wearies of its role as globocop, there is no Plan B.

On the whole, American efforts to preserve global order will directly benefit the country's many overseas allies, but sometimes, inevitably, they will not—which means that the allies, too, will have a major role to play in the endgame. Sometimes they will need to speak truth to power, telling the globocop things it does not want to hear; at other times they will need to back up the globocop with diplomacy, money, or even force of arms. Above all, they will need the wisdom to know when to subordinate their own local concerns to a global strategy, recognizing that the whole is greater than the sum of its parts.

The most difficult decisions of all, though, might be those that fall to the globocop's rivals. The more these rivals' wealth grows, the more their moves will affect how the endgame turns out. A hundred-plus years ago, the British globocop's two greatest rising rivals were Germany and the United States. Germany's Kaiser Wilhelm felt that the only options open to him were risky moves that undermined the global order, while the United States found ways to act in its own interests while still (mostly) shoring up the globocop. “Speak softly and carry a big stick,” President Theodore Roosevelt advised; and as today's rising rivals acquire big sticks of their own, their leaders will have to choose between Roosevelt and Wilhelm as models. The United States can do much to influence these choices by making room for rivals' peaceful development while simultaneously deterring rash aggression; but in the end, the more that America's rivals
lean toward Roosevelt, the more likely the world is to win the endgame of death.

Si vis pacem,
said a famous Roman proverb,
para bellum:
If you want peace, prepare for war. Despite everything that has changed in the two thousand years since Calgacus and Agricola fought rather than talked at the Graupian Mountain, this has remained true. The song “War” got it wrong. War has
not
been good for absolutely nothing, because—uncomfortable as it is to face this fact—war is the only method humans have found for moving from tiny Stone Age bands with rates of violent death in the 10–20 percent range to today's vast, globalized society with a rate below 1 percent. War has made the planet peaceful and prosperous; so peaceful and prosperous, in fact, that war has almost,
but not quite,
put itself out of business. Hence the final paradox in this paradoxical tale: If we really want a world where war is good for absolutely nothing, we must recognize that war still has a part to play.

Footnotes

¹
Measured in 1990 international dollars, a standard unit of comparison. At current market exchange rates, global GDP/capita in 2011 was more like $12,000.

²
IED: improvised explosive device (that is, a homemade bomb). Medevac: medical evacuation helicopter.

³
I would like to offer my thanks once again to General (retired) Karl Eikenberry and the soldiers at Fort Irwin for arranging this visit.

⁴
If we measure instead from its low point in 1998 to its peak in 2010, American defense spending almost doubled, but even that is less than one-third of the Chinese proportional increase.

⁵
The Diaoyu Islands (Senkakus in Japanese) are claimed by both China and Japan.

⁶
I would like to thank the institute once again for this invitation.

⁷
I would like to thank Mat Burrows, former counselor to the director of the National Intelligence Council, and Banning Garrett, director of the Strategic Foresight Initiative at the Atlantic Council, for inviting me to make presentations to their organizations in July 2011 and to join several of their Silicon Valley meetings since then.

⁸
Here, as so often, terminology is contentious. “Remotely piloted aircraft (RPAs)” is the preferred air force term, emphasizing that these are still aircraft and do still have pilots. The army and the navy speak of unmanned aerial vehicles (UAVs), while civilians tend to say “drones.” Being a civilian, I will say “drones” too, although in military circles “drone” conventionally means a robotic vehicle used for gunnery practice.