The Watchers (22 page)

 

 

 

 

Poindexter worked all day at his desk, convinced that the morning's events were an opening salvo, not a final shot. Outside, the country braced for a follow-up attack. The Federal Aviation Administration grounded all air traffic in the United States. The stock market closed. Soldiers and policemen patrolled the streets of New York and Washington, keeping close watch on subway trains and government buildings. Initial estimates put the death toll at the Pentagon as high as eight hundred. In New York, where the casualties at the World Trade Center were yet unfathomable, a Navy aircraft carrier and warships took up defensive positions off Long Island. America was at war.

As he sat in the nearly empty office, Poindexter cleared his mind, pushing aside the noise that surrounded him even now in such a quiet space. What signal had been missed? What clue? What moment? But more important, how could he catch it next time? What did he have to do now to preempt that second wave he was certain crested just beyond his line of sight?

He worked the puzzle over all day and at home that night. It waited for him when he woke up the next morning and when he piled into his car again, the radio turned off. Poindexter had just driven out of his quiet, tree-lined subdivision when the answer hit him.

He pulled his car over to the side of the road and grabbed his cellphone, scrolling through his contacts list. He found the number of the one man he knew was knocking around in the same headspace that morning.

“That's funny,” Brian Sharkey told his old friend when he rang. “I was just thinking about calling you.”

Sharkey and Poindexter had been waiting for this day. And this conversation, where they'd commiserate over lost chances. Poindexter vented about how slowly the intelligence community had accepted the Genoa program, and how few tools he and Sharkey had developed had actually been implemented. Poindexter thought DARPA management had been too stingy, allocating a mere $40 million over the past five years to advance and sustain his research.

Poindexter had plenty of reasons to condemn those he'd tried to persuade. But at that moment he saw a window, and he decided to jump through it.

“We need to talk to Tony,” he said.

Tony Tether, the new director of DARPA, had been on the job only three months. Poindexter didn't know him well, but Sharkey did. Poindexter wanted a chance to get in front of Tether with an idea that had been gelling in his brain that morning but that both he and Sharkey had been nurturing for some time.

They called it “total information awareness.” Sharkey had introduced the phrase two years earlier in Denver during a speech at the annual DARPATech conference. The event brought out all the big names in the military R&D world, and Sharkey had told them what they already knew: The world was awash in information, and every day valuable intelligence was lost in the deluge. The sources of information seemed to expand at uncontrollable rates. It was clear that without some intervention the momentum of data would outstrip the government's ability to keep up with it.

Sharkey had imagined a comprehensive system to collect all this diffused data about real world events and then assimilate them in a process of “collective reasoning,” which would be conducted by human analysts linked together through their computers. This was the heart of Genoa, but the concept of total information awareness took things a step further. This system would collect information too and vet it against known types of crises. If it could be trained in what signals to look for, then the system might identify an impending crisis early and increase the chances of preempting it.

It was a fanciful notion, one that, Sharkey admitted, tested the boundaries between human reasoning and computer automation. Sharkey was still a DARPA employee, and he reminded the crowd that total information awareness was a concept, not an active program.

Poindexter wanted to change that. It was time to bring TIA to life.

He told Sharkey that DARPA needed a new program office, a place to harness Genoa and a slew of other technologies into a large system, like the one that he'd outlined two years earlier. Poindexter could think of several tools under research that would fit logically into this new portfolio. There was one aimed at reaping rich data harvests from large information sources and making meaningful connections. Another area that Poindexter thought received far too little funding was automated translation, computer programs designed to convert foreign languages, whether written or spoken, into English. Also, he was keen on a program that DARPA had neglected that could use facial recognition technology from a distance, and in crowds.

These ideas were force multipliers—each would make the other stronger when used in concert. Poindexter wanted to bind them all. To create a superstructure, a “system of systems.” A single, automated design to discover information, to mold and shape it. It would see what human eyes could not. And perhaps to finally, truly, understand the world around them.

Though Poindexter had imagined such a grand apparatus from his earliest days in the White House, the technology to build it didn't exist then. Times had changed. Now the physical constraints were the least of his concerns. People had to change. That was the hardest part.

Fortunately for Poindexter, people were scared now. And that meant they'd be ready to take risks. Sharkey agreed to call Tether and arrange a meeting. Poindexter went back to his office and spent the rest of the day sketching the first draft of his plan.

 

In all his years of thinking about preempting terrorism, Poindexter had fixed his eyes abroad. Even at the height of the terror wave in the eighties he had never contemplated a massive assault on U.S. soil. Terrorism was a foreign problem, and more specifically, a foreign intelligence problem. Vast and nebulous though his work was, those factors gave it a useful frame. It told him, and told the government, where to look, what kinds of information to gather.

Whatever signals pointed to the catastrophic events of September 11, they were most certainly not to be found exclusively abroad. Poindexter reasoned that unless these men had slipped into the United States only days before their mission—and that seemed implausible—they would have left many traces here. They made phone calls. Probably sent e-mails. They bought merchandise with credit cards. They made plane reservations. Perhaps they'd lived here a long time, leased apartments, maybe bought property. They would have gone to grocery stores, meandered through shopping malls, dined at restaurants. People would have seen them. But more important, and more dependably, computers would have seen them and kept tiny memories of the encounter. Everywhere they stepped these men left digital footprints, the permanent markings of where they'd been, what they'd done, and what they might do next. These were their signals.

Poindexter had a theory. Just as submarines emitted a unique sound signature in the ocean, terrorists emitted a unique transaction signature in the information space. If he could pick it out, then he could find the goblins hiding in that noisy sea.

The idea departed radically from every other counterterrorism program to date. The intelligence and law enforcement agencies had always hunted for terrorists by starting with a known target. They'd get a name, or a phone number, and then scan their databases for any mention of the quarry. Working step by step they could link the target to other people and places, or to known events. This was shoe-leather investigation. Excavating a terrorist network one data point at a time.

Technology had sped up the process, but the end results were limited. As Kleinsmith had learned in Able Danger, sometimes the mass of links and connections was so dense, so crisscrossed, that it was unintelligible. But terrorists were hiding in that mass. As investigators located the 9/11 hijackers' electronic transactions, surely they would find overlooked clues, dots unconnected. But for present purposes, the CIA, the FBI, everyone had obviously been blind. How was the intelligence community expected to find men who it didn't know existed? That, Poindexter thought, was the crux of the new problem. Terrorists were not about to signal their intentions. The government would have to find them first. But in order to find one man's digital footprints, he decided, you had to look at everyone's. The innocent and the unknown guilty. Total information awareness.

Instead of asking airlines to raise a red flag whenever a suspected terrorist purchased a ticket, Poindexter wanted to look at all passenger reservations for anomalies. There were patterns of transactions he thought might indicate terrorist activity. People buying one-way airline tickets who also entered the country together, or near the same time. Maybe they stayed in the same hotel together, or shared a credit card account. Maybe they rented a car together. He decided that the best way to find these phantoms was to look for evidence of their activity first—and decide later whether the people in question were truly innocent.

Before September 11, he would never have dared propose such an idea, much less have any confidence that anyone would back it. But the old method of tracking known targets had obviously failed. Now pattern-based searching of everything, of everyone, of all that was known or knowable, seemed to him the only logical choice.

 

Sharkey called Poindexter back. Tether was willing to meet and hear about his idea. Given their personal history, Poindexter thought that Sharkey should make the first, broad pitch. He arranged for the two of them to have lunch at a quiet restaurant a few blocks from DARPA headquarters. Poindexter had spent most of September 12 crafting a set of talking points, enough to give Tether the essential elements of the plan.

Tether liked what he heard. All across the government agencies were revving up for a new kind of war. Previously “risk averse” organizations, as critics liked to think of the CIA and others, were now desperate for ideas. Tether would back Poindexter's program, he said, but on one condition: Either he or Sharkey had to come back to government to run it. This was far too complicated a task to leave to anyone other than the experts, and clearly, they were Poindexter and Sharkey.

Sharkey left the luncheon and relayed the bad news to his friend. Neither of them wanted to return to public life. Since his speech in Denver, Sharkey had left DARPA and joined a boutique technology and consulting firm, Hicks & Associates. He enjoyed private-sector work, both for the intellectual freedom and the financial reward. Hicks's parent company, Science Applications International Corporation, was one of the largest employee-owned companies in the country, and Sharkey had amassed considerable stock options. The company's business would boom as government agencies planned to spend billions on new technology for counterterrorism. If Sharkey returned to public service, he might have to give up his options. He'd certainly have to take a substantial pay cut.

Sharkey was a younger man than Poindexter. He had different obligations. He just wasn't prepared to come back, and Poindexter didn't want to force him. Besides, Poindexter was the mastermind. He could see the system from end to end. And he knew how the bureaucracy worked. Few DARPA program managers had ever been national security adviser to the president. Poindexter's return was the agency's gain.

All that was true, but Poindexter had his own reasons to turn the job down. Along with his résumé came his baggage. How big a distraction might that be? He and Linda weighed the options together. He had the bureaucratic expertise, the technological fluency, the vision. He also still had the highest level of security clearance, owing to his work at Syntek.

He also weighed the financial concerns. Poindexter was not a rich man. But his children were grown. He owned his home. The boat was the biggest expense, yet a manageable one. There was really only one question: Was John Poindexter his plan's best asset or its worst liability?

Though he'd been a master of casuistry in his dealings with Congress, Poindexter had always been honest with himself. His association with an ambitious and controversial counterterrorism program would cause trouble. He had no doubts about that. The moment he stepped back onto the public stage he could become a lightning rod atop his provocative venture.

Poindexter had long thought that if he had it to do over again, he would have devised a public relations strategy for the “Iran-Contra business,” as he preferred to call it. He had warned his cohort then that if the convoluted operation ever became public, they'd have a very hard time explaining it.

Washington was a town of institutional memories. Many of his oldest allies in government were back in power, in senior positions within the Bush administration. And his oldest foes had remained in Congress. Poindexter wasn't sure now how, or if, he could avoid the lightning. But he figured it would come eventually.

He thought about it for a few days, then picked up the phone and called Sharkey.
Okay. Let's do it.