Now Thorpe had an idea that seemed to match the “cost cutting” mission for DARPA. He said there had been a big tank battle during the Gulf War, just a few days prior. Thorpe wanted to re-create the battle in a virtual world of computers, something that had never been done before and could potentially save money on training. The key, however, was that he wanted to send scientists to a battleground in Iraq strewn with still smoldering Iraqi tanks. “I think we can do something with that,” he told Reis, “but I would like to be able to go over.”
The idea struck Reis as prescient; an inherent part of simulation is cost savings, because it is cheaper to practice in a video-game-like environment than to expend money on real fuel and practice ranges. “Sure,” the director told him. And with that, DARPA set off on one of its most ambitious post–Cold War projects: to re-create real war in a virtual world, based on data gathered from the smoking ruins of a battlefield.
—
Normally, after major battles, army historians would be sent to interview participants and chronicle events as part of a written account of what took place. Thorpe wanted DARPA to send simulation experts to the battlefield, walk among the burned-out tanks, interview American soldiers who had fought there, and then plug that data into the virtual reality world of SIMNET. The entire battle could be re-created and played in a simulator, and more important, those simulators could be connected on a network, sending data packets back and forth, so that people could replay the battle as participants. “It would be like a living history,” Thorpe suggested.
Thorpe proposed the idea to General Gordon Sullivan, the army’s chief of staff. Soon, General Sullivan contacted Lieutenant General Frederick Franks, the Seventh Corps commander, who gave DARPA the go-ahead and selected 73 Easting, the largest tank battle since World War II. The battle had taken place on February 26, 1991, when the Second Armored Cavalry Regiment met up in the middle of a sandstorm with fleeing elements of Iraq’s elite Republican Guard Corps. Over several hours, American forces destroyed dozens of tanks, armored personnel carriers, and trucks. The name 73 Easting was taken from the grid location in the desert where the Iraqi forces were decimated.
Two days later President Bush declared a cease-fire and the Gulf War ended. Within a week of the battle, a DARPA-sponsored team of researchers arrived in the Gulf and traveled to where the battle took place. The researchers interviewed the American soldiers who fought in the battle, asking them to recount what happened minute by minute. “There were still the tread marks in the sand. They were able to see all the blown-up Iraqi vehicles, which were still there,” Thorpe said. “The army engineers had been in to annotate precisely where every one of those blown-up tanks was and the conditions of its demise. Was a turret blown off? And then what direction was it lying?”
When the team returned to the United States, its members worked in a room with Post-it notes stuck on different parts of a board. Like detectives tracing backward from a crime scene to re-create what happened, the scientists reconstructed the entire battle. The work took a year, but the results were unprecedented: a computerized, interactive reconstruction of a real-world battle, with a feature called “the magic carpet,” which allowed users to zoom in and around to any place on the battlefield, at any moment of the battle. “You would be able to go in and visit, look at, and replay the simulation and see where everybody was, what they were shooting at, and even get inside their vehicles or ride a shell as it races downrange at a mile a second,” Thorpe said.
By all accounts, the digital re-creation of 73 Easting was a technical success. Thorpe even commissioned a Hollywood-produced video showing the simulation in action, featuring interviews with some of the soldiers who fought in the actual battle; they praised the simulator for its realism. Reis, the director, showed the video to Cheney, the secretary of defense, and to General Colin Powell, chairman of the Joint Chiefs of Staff. He also took it to Capitol Hill and played it for lawmakers. Everyone loved it, and Cheney most of all. “Gee, if we had this earlier, I could have shown this to Hussein, and he might have realized in how bad shape he was and just given up,” Cheney said, perhaps foreshadowing misplaced optimism he would have more than a decade later about Iraq.
Thorpe’s SIMNET, which had been developed in the 1980s at DARPA, was already a widely regarded success. It had changed how the army used simulators and, for the civilian world, helped enable the technology that gave rise to online gaming. As a follow-up to SIMNET, the re-creation of 73 Easting was a hit, demonstrating what simulation could do when combined with real-world data. In technology circles, Thorpe became something of a folk hero, leading Wired magazine to declare that it was not the science fiction writer William Gibson who invented cyberspace but Jack Thorpe. Yet it is less clear that SIMNET, and by extension 73 Easting, really had much practical effect on the military. The reality is that SIMNET was fielded too late to help the tank operators who fought at 73 Easting, though it was used in later years. “SIMNET was irrelevant to us,” said Douglas Macgregor, who led a key squadron in the Battle of 73 Easting. “We never used it.”
The question was, what could be done with 73 Easting, the follow-on to SIMNET? The answer, by the 1990s, turned out to be not much. The idea would have been to use the technology that emerged from the 73 Easting re-creation in other simulations, but DARPA’s connections to the Pentagon’s leadership were breaking down, and cutting-edge science and technology were no longer at the forefront of the Defense Department’s agenda. Even the retired army general Paul Gorman, the intellectual godfather of simulated training, questioned whether the virtual re-creation of a battle contributed much to the military. “I don’t know how to answer the question about 73 Easting,” Gorman replied, when asked years later about its impact. “Elegant effort and delighted that it was done, but who is using it?”
For all the accolades heaped on 73 Easting, nothing concrete came out of it other than a re-creation of a tank battle that would likely not be fought again for at least another generation. Gorman had hoped it, and similar simulation efforts, would achieve much more, by changing how the military prepared for future wars. Using simulations to practice for the last conflict did not do anyone any good. “We were addressing in effect, force on force, something like symmetric opponents,” Gorman said. For fighting insurgencies “you would need a superior formulation, and damn it, they didn’t provide for that.”
—
In its effort to develop simulation technology, DARPA found itself hitting a wall. When DARPA did try to move simulation beyond a Cold War battlefield, its efforts had limited success, because technology alone could not change policy. One never-publicized example involved the drug war. The White House’s Office of National Drug Control Policy in the mid-1990s funded DARPA’s simulation experts to create a model of drug trafficking to see if there might be ways of cutting off the drug cartels in South America. “The big issue was and still is the movement of cocaine from Central and South America into the United States,” explained Dennis McBride, who was in charge of the effort. He named the project after Iolaus, who in Greek mythology had helped Heracles battle Hydra. The name ended up more appropriate than he had imagined.
“We built this incredibly complex end-to-end model from seed planting down in South America through the changing to a product at a wholesale level, the transportation across myriad modes of transportation, ultimately into warehouses in the United States of America,” McBride said. Yet the more DARPA modeled the problem, the worse it looked. If one cartel was defeated, it ended up just strengthening another cartel. Like the Greek Hydra, if you cut off one head, two more rose in its place. DARPA came up with answers, but the answers did not fit what the White House wanted. If the Drug Enforcement Administration put more aircraft in the air, it did not help, because the cartels still had more planes. No matter which way DARPA modeled the drug war, it could not come up with a scenario that cut off the supply. “We built this very big model. We played with it every way we could. We said, ‘Let’s do this,’ an
d, ‘Let’s do that.’ At the end, this huge model would say here’s the result and it was not good news.”
The simulation showed the limits of technology to solve what was essentially a policy problem: simulation was not going to teach anyone how to win the drug war, it could only demonstrate that it was unwinnable, and that was not a message the government wanted to hear. The reaction was denial: law enforcement would just have to try harder. “I don’t know if we’re a hell of a lot better off that we now kind of understand the problem because we have the simulation,” McBride reflected. “It’s like massive wounds all over the body; blood is pouring out from everywhere. We can understand that, but there is nothing we can do about it.”
The counter-drug simulation failed because technology hit up against the limits of policy. Other DARPA simulation efforts failed at a more basic level. One 1990s-era project, called War Breaker, was supposed to come up with a silver-bullet solution to the main threat American forces faced in the Gulf War: mobile Scud launchers. The Soviet-made tactical ballistic launchers had proved difficult to find and destroy. Despite Iraq’s rapid defeat, its Scuds had proved deadly. One attack, on February 25, 1991, killed twenty-eight American soldiers in Saudi Arabia. DARPA contracted to have a simulation facility built outside Washington, D.C., even employing Herman Zimmerman, the set designer for Star Trek: The Next Generation, to build the laboratory, modeled after the bridge of the Enterprise. The simulations made for convincing theater when senior officials were brought in to view demonstrations, but the reality was far less impressive. “You could figure everything out beforehand if you wanted to show off a missile, an airplane, or whatever you wanted. You could put it in a situation where it would just look wonderful,” said Ron Murphy, a DARPA program manager who worked on War Breaker.
The reality of mobile targets was too complex for manned simulation, according to Murphy, and for every simulation that turned out well, there were many more that did not. Despite the success of SIMNET, which focused on tank training, DARPA was simply never able to create a realistic simulation that would combine air and ground operations. War Breaker, once slated to get more than half a billion dollars in funding, and DARPA’s most expensive post–Gulf War effort, was quietly ended. Almost all mention of it was expunged from later agency documents.
DARPA had spent the 1990s creating synthetic versions of war involving planes, tanks, and missiles. The agency had expanded beyond that on occasion, for example to the drug war, but even then it was focused on modeling concrete objects: drugs and money. L. Neale Cosby, a former army officer who was involved with SIMNET and 73 Easting, acknowledged that one thing DARPA’s simulations could never do well was to model human beings. “It’s easy to simulate this room,” Cosby said. “To put all these people, doing the right things, all thinking independently, in the right place so you actually virtually copy this room is tough. That’s the tough simulation part that we still don’t do that well.”
In the decade that followed the collapse of the Soviet Union, it became increasingly clear that people—not tanks or missiles—were precisely what needed to be modeled. In 1993, terrorists detonated a truck bomb in the North Tower of the World Trade Center, killing six people. Then, in 1998, al-Qaeda carried out near simultaneous attacks against two American embassies in Africa. Two years later, its operatives attacked the USS Cole in the port of Aden, in Yemen.
In 2000, the same year as the Cole bombing, Tom Armour, a former CIA officer who had recently joined DARPA, announced a new type of modeling and simulation work that the agency was beginning to pursue. The United States was facing a new threat, from terrorist groups, or what the Pentagon was calling an “asymmetric threat.” This “asymmetric threat is physically small—perhaps even just a single person,” Armour, the DARPA official, said. “To predict the potential range of actions, the analyst will need to model the group’s beliefs and behavior patterns.” The DARPA director that year authorized the start of a new program, called Total Information Awareness. It sought nothing less than to predict human behavior, in particular the behavior of terrorists.
Just a few months prior to Armour’s talk, three al-Qaeda operatives based in Hamburg arrived in the United States to begin training as pilots. They were part of a larger terrorist cell sent by al-Qaeda to attack America. The very behavior pattern that Armour spoke about was emerging. DARPA had switched gears, but like the rest of the national security community it had done so just a little bit too late. By 2001, DARPA’s simulations were still preparing the military for tank battles and air-to-air combat, while in the United States nineteen men were training to take over commercial airplanes armed with nothing more deadly than box cutters.
CHAPTER 17
Vanilla World
In July 2001, Scooter Libby, Vice President Richard Cheney’s chief of staff, called up DARPA’s new director, Tony Tether. The vice president wanted to visit DARPA to be briefed on its programs. Cheney knew Tether from his days serving as the secretary of defense, but even so, his call was a surprise. The White House had not taken an active interest in DARPA for years.
In the summer of 2001, the agency founded more than four decades earlier was adrift. Without a Cold War enemy to fight, DARPA during the 1990s was funding politically expedient projects, and over the previous decade that had led to a series of expensive flops. DarkStar, a stealthy drone sponsored by DARPA at the request of senior Pentagon leaders, suffered a software glitch on its second flight and crashed. With the navy, DARPA embarked on an ambitious ship project—a floating missile platform—only to have the work canceled following the suicide of an admiral who supported the program. Most disastrously, DARPA signed up to work on an ill-conceived army project called Future Combat Systems, which was supposed to link missiles, drones, and ground vehicles on a single network (the overly complex project was also eventually canceled). DARPA lurched from one project to the next, without any real strategic direction or plan. “DARPA had become a backwater type of organization in the latter part of the nineties,” admitted Tether, who took over in June 2001. The vice president’s sudden interest was a chance to change this.
Over the next three weeks, Tether and a handful of office directors worked days, nights, and weekends to prepare to brief the vice president. Cheney “was a person who liked pictures, he is a very visual person,” Tether recounted. “So, having a slide with a whole bunch of words on it was just, ‘oh, wow,’ you don’t ever do that to him. You know, you give him a cartoon.” Tether recounted that he wanted something that would blow the vice president away, and so the cartoon he chose for Cheney was Superman.
The headlining act for the vice president would be super soldiers, the brainchild of Michael Goldblatt, a research manager who came to DARPA from McDonald’s, the fast-food company. At McDonald’s, Goldblatt had run the corporation’s venture capital efforts, a sort of DARPA-like arm of the nation’s biggest fast-food chain. There, he worked on projects like self-sterilizing food wrappers, which he had been trying to sell to the military. No one in the Pentagon would take his calls. Finally, he was referred to DARPA, where the then director, Larry Lynn, agreed to speak to him, thinking Goldblatt worked for McDonnell Aircraft Corporation rather than McDonald’s. Within a few years, Goldblatt found himself working directly for DARPA, initially overseeing the agency’s program for defense against biological weapons.
At DARPA, however, Goldblatt turned his attention to something far more ambitious than food wrappers: he wanted to work on enhancing human beings. Goldblatt was inspired by science fiction like Firefox, the 1982 movie starring Clint Eastwood, which featured weapons controlled by the human mind. Under Goldblatt, DARPA funded a group of researchers at Duke University to implant microelectrodes in monkeys’ brains. The electrodes would read their brain signals, which could then be used to manipulate a real object, like a robotic arm. Goldblatt’s other ongoing research programs sounded equally fantastic: humans who could survive severe blood loss in suspended animation, soldiers who could go for days
without food or sleep, and warriors enhanced with superhuman strength and mental abilities. Superhuman soldiers with mind-controlled weapons sounded like the perfect cartoon for Cheney.
In late July, the vice president showed up at DARPA headquarters, accompanied by Secretary of Defense Donald Rumsfeld and Edward “Pete” Aldridge, the Pentagon’s chief weapons buyer. Starting with Goldblatt and his enhanced humans who could survive critical wounds and bitter cold, or go for days without sleep and food, Tether and his office directors briefed the three men for six hours. They focused on Cheney, and he loved it. “It was fantastic,” Tether said.
When the three men left, Tether knew the briefings had been a hit. He got word later that both the vice president and the defense secretary were enthused. It was one thing to have the support of senior Pentagon and White House officials and another to figure out what to do with that support. The answer came just weeks later, and it turned out it had nothing to do with super soldiers.
—
If DARPA had been adrift over much of the past decade, so too was the rest of the national security establishment. The 1990s “peace dividend” had featured declining defense budgets, including for DARPA, and without the threat of an armed confrontation with the Soviet Union, no single threat generated sustained attention. Even as threat reports on al-Qaeda surged over the spring and summer of 2001, defense and intelligence officials were moving in slow motion. Around the time that DARPA was briefing Cheney and Rumsfeld on super soldiers, intelligence on al-Qaeda and its operations seemed to be exploding. Al-Qaeda was planning something “spectacular.” An attack was “imminent.” The reports on al-Qaeda activities had “reached a crescendo,” Richard Clarke, the White House counterterrorism czar, warned in June 2001. In August, the FBI began investigating Zacarias Moussaoui, a French national with a fervent belief in jihad, $32,000 in a bank account, and an unexplained interest in learning to fly Boeing airliners.
The Imagineers of War Page 34