The Watchers (48 page)

ADVISE's makers still hadn't achieved a system of systems, but that was their ultimate goal. Steve Dennis told the group that day that he was more interested in actual products than data “methodologies.” The department had real work to accomplish. The meeting attendees debated whether there were any privacy constraints on information that the Homeland Security Department already had access to, such as immigration records. Toward the end of the discussion someone noted that Poindexter's old TIA network was testing components for the ADVISE system. That was helpful, since the team's approach had been to use “real data.” The participants also discussed whether they should engage a red team that ADVISE had put together to help come up with ideas for what kinds of data they should be examining.

Had Poindexter been in the room he might have offered some pointers, since he'd been down all these roads already. ADVISE was the conceptual successor to total information awareness. But it was based on a failed technology, one that the NSA had already tried. The incestuous plot line would have made even the most jaded soap opera writer blush.

Poindexter had trouble keeping up with all the ersatz programs that rode his wake. He was familiar with ADVISE only in passing. He'd heard people mention it but had never looked deeply into the idea. And he was aware of others. The team representing the NSA at the meeting worked on a program run by a technologist named Art Becker, who had started up his own program that looked a lot like Genoa. Becker called his version “Geneva.” Poindexter had heard it never amounted to much.

Just as ideas got recycled in the government research community, they also tended not to stay secret for very long. Poindexter had lit a match that even before this government meeting had spread into an untamable fire. It was only a matter of time before it spread beyond the Beltway.

 

On Wall Street, knowledge was power too. And for as long as anyone could remember, the kind of knowledge that could turn information into money was the province of a special few.

Traders, brokers, and dealers placed their bets on who was up and who was on the way down based on information that they swapped within a closed, elite, and jealously controlled network. Rumors, tips, and insider intelligence were the currency of the would-be Gordon Gekkos and Sherman McCoys. To augment their privileged sources they had high-priced newsletter subscriptions, Bloomberg computer terminals, and a chattering class of fellow pros to help them keep ahead of the masses.

And then one day that network was replaced by a new one. The Internet gave rise to a class of day traders, stock aficionados, and amateur investors. A proliferation of financial news sites rivaled what the rulers of the Street had been trading in person. Suddenly, and unceremoniously, the Masters of the Universe had been supplanted by bloggers.

Joseph Kennedy supposedly had warned that it was time to get out of the market when the shoeshine boy started offering him stock tips. But more than seventy years after the great market crash, Jeff Stewart sensed that this new whirlwind of corporate information was something to be embraced.

Stewart, a serial entrepreneur working in Manhattan, thought that if he could detect subtle changes in the market in real time, he might make a fortune. Conceptually, the idea was straightforward. Vacuum up corporate press releases, earnings reports, newswire stories, anything on the Internet that gave even the slightest hint of change in the status of a particular company. Then use that formerly privileged information to guide an investment strategy. But not for the long term. Just for the day. Maybe even for the minute, if making gains meant reacting on a moment's notice to even the slightest event that could spell big change in a traded company's fortunes.

Stewart wanted to build total information awareness for Wall Street. He was well aware of Poindexter's concept, having read about it in the press. Technology made the idea doable. The Internet provided the data, so Stewart just had to build the system to vacuum it up and analyze it.

He might see an uptick in blog chatter about a particular product—rumors of a soon-to-be-released new version of the iPod, or an environmentalist backlash against paper products, like Kleenex tissues. If those signals could be tied back to a company, and thus to its stock, an investor could hedge his bets. He might be able to know whether the stock would rise or fall
before
the rest of the market.

That kind of information was golden. And if obtained in an instant, it could mean the difference between huge losses and huge gains. Stewart needed an early-warning system for arbitrage instead of terrorism. Really, were they all that different?

In 2004 he started talking to potential customers in New York, particularly hedge fund managers, who were crawling all over the city at the time. Stewart couldn't walk into a Starbucks without striking up a conversation with one of them, and so he started asking, would you be willing to pay twelve thousand dollars a year for this kind of early-warning and detection system?

He wanted to know if this was a reasonable fee. After a string of them answered yes, Stewart made hedge fund managers his exclusive target customer. He knew they would pay handsomely for anything that gave them the thinnest edge on their competitors.

When the time came to actually build the system, Stewart knew whom to ask. The government. And more specifically, researchers who had worked for the intelligence and national security community, who would know if something like this was feasible or had already been built.

In late 2004 Stewart met with scientists at the Lawrence Livermore lab. He'd been steered to them through his own connections in the New York venture capital set, which were never more than a degree removed from experts in academia and R&D. When Stewart met with the scientists he could see that they had the answers to his questions on the tips of their tongues. Clearly, he thought, they'd already been researching this question of how to harness penetrating insights from massive data. He listened to them, and then he incorporated their ideas into his product.

Stewart called his system Monitor110. He designed a prototype to gather all information about any changes in a company or product and then to display those changes as they happened. It was a hedge fund manager's dream.

Stewart learned that he wasn't the only one interested in monitoring huge data streams. From his conversations with people in the government, he understood that a big department was also interested in taking this approach—over at the Homeland Security Department. They called it “ADVISE.”

It didn't surprise him. Stewart knew that the community of people who specialized in this kind of information exploitation was so small that their ideas were bound to cross-pollinate the government and the corporate world. Looking back, he figured that the scientists at the labs, the ones who had contributed to his system, eventually sold their ideas to the government.

Stewart was correct.

Monitor110 earned some favorable media attention as one example of a new, innovative approach to managing information. The financial trade press was abuzz. But inside the intelligence community people knew that Stewart's creation had the same core that the Livermore lab had sold to Homeland Security to build ADVISE. The BAG was at the heart of Monitor110. The government's bad idea had made its way to Wall Street.

They knew about it within the National Security Agency, where technicians referred to “ADVISE/Monitor110,” as if it were the same system. The agency had discarded the BAG when it failed to produce much more than hair balls. That helped spur the move to in-memory databases.

Recycled ideas. Recycled baggage. Again, the surveillance narrative had reached new heights of absurd complexity. A failed technology to track down terrorists, which had been ditched by the NSA. Then it was repackaged and sold to the Homeland Security Department, whose core mission was to protect the country against terrorists. Then it leaked out into the private sector, where it was adopted by Wall Street moneymen. But the technology died there too, overwhelmed by the proliferation of noisy data on the Internet, particularly junk spam messages about companies and stocks that ended up clogging Monitor110's output. Years later one would have to wonder if it would have been smart enough to detect the pending meltdown of the global economy, which was set off in part by opaque transactions by the very companies that would have purchased such a system.

Monitor110 wasn't built on the BAG alone. Stewart talked to dozens of technologists, at the labs, in academia, and in the private sector. But to those who knew the BAG's sordid tale, Monitor110 was just another casualty of a long struggle. Stewart was reaching for the same, elusive ring as the NSA and Poindexter.

Monitor110 shut down, and its creator moved on to new pursuits. By the end of 2006 Poindexter had engaged in new ventures too, with Saffron, the associative memory technology making gains in Iraq. He'd eventually pitch it to civilian agencies in government too, including the IRS for tax fraud detection.

Poindexter maintained his close ties with high officials in the administration, particularly those like Fran Townsend, who were on the front lines of the terror war. But he couldn't have predicted that one of his oldest allies, and great fans, was about to take over that war, and to change its course for years to come.

 

 

 

 

At the age of sixty-three, with a distinguished Navy career behind him and the rewards of a seven-figure salary ahead, Mike McConnell never expected to find himself in the Oval Office. And yet that's where he spent almost every morning now, as the newly minted director of National Intelligence. In February 2007 the onetime NSA director and retired admiral was sworn in as the president's new spy master, the man George Bush would depend on to bring some professionalism and calm back to the cantankerous intelligence community. Bush was convinced that certain career employees, particularly at the CIA, had tried to sabotage his reelection bid in 2004. Unnamed intelligence officials had leaked accounts of “cooked” intelligence on Iraqi weapons programs and claimed that the White House had relied on dubious reporting to make the case for war. The latter turned out to be true, but the intelligence community was hardly vindicated, since its analysts had relied on those sources in the first place. Porter Goss's leak hunt at the CIA and the firing of Mary McCarthy only deepened the distrust between the president and his spies, and it reinforced the feeling among many career intelligence officials that the Bush administration had politicized intelligence to a dangerous degree.

McConnell was widely seen as a professional and a nonpartisan. But he had been reluctant to return to the fold, having been out of government more than ten years now. The NSA was his last post, and he'd done better than most retired intelligence officers in the private sector. McConnell had twice turned down offers to join the Bush administration—once in July 2006, when director of national intelligence John Negroponte offered him the deputy slot, and then again in September, when McConnell was offered the top job. It was well-known that while the DNI was supposed to be the new chief executive of the spy agencies, the defense secretary still controlled most of the intelligence budget. The statute creating the DNI hadn't vested the office with enough legal leverage to overrule the Pentagon.

Had McConnell decided to return in 2006 he would have had to contend with Donald Rumsfeld, who was presiding over an internecine war with the civilian intelligence agencies. Rumsfeld, long distrustful of the CIA, was setting up a covert human intelligence apparatus that reported through the Defense Department chain of command and not to the agency. The CIA had always run foreign espionage operations, and stories abounded about hard-headed, inexperienced Pentagon spies running amok because they weren't coordinating with CIA station chiefs in foreign capitals. Negroponte had been unable to halt Rumsfeld's advances.

But then, in November 2006, Rumsfeld's star finally burned out. Republicans were dealt a “thumpin' ” in the midterm elections, as Bush called it, and Democrats took control of both houses of Congress for the first time in a dozen years. Voters had rebuked the Iraq war, and by extension, Rumsfeld's leadership. He was out. The administration announced it was putting up Robert Gates, an ex-CIA director, to replace him.

On December 23, as McConnell was preparing for the Christmas holiday, his secretary walked into his spacious corner office at Booz Allen Hamilton, about twenty miles outside the capital. “The vice president's on the phone,” she said.

“The vice president of what?” McConnell asked.

“The vice president of the United States.”

McConnell jumped up and grabbed the phone. “Mr. Vice President! Mike McConnell here.”

Dick Cheney got right to the point. “Mike, the president and I want you to consider a nomination to be the next DNI.”

Cheney had known McConnell from the first Bush presidency, when then captain McConnell was the intelligence officer for the chairman of the Joint Chiefs of Staff, Colin Powell, and unexpectedly found himself helping to fight a war. When Saddam Hussein amassed forces on Iraq's border with Kuwait, McConnell predicted the invasion almost a day early. But he was an intelligence expert and unfamiliar with ground warfare tactics. He'd never expected to be engaged in combat operations. Nevertheless, McConnell was such a quick study that Powell eventually put him in charge of daily press briefings during Operation Desert Storm. After the war Powell and Cheney, who was the defense secretary, supported McConnell for director of the NSA. It was a three-star position, and McConnell had been promoted only recently to be a one-star admiral. McConnell's patrons ensured that he was bumped up in rank.