Burn-In

by P. W. Singer

  He began pacing, his sandals sticking in the wet mud of the game’s simulated grass and loamy earth. “Do you all get that? Standing here, or not, as the case may be, hiding behind these avatars, you have become dependent on what you claim to be combating. You choose the veils afforded by technology, not realizing you are wearing blinders.”

  “Such grandstanding,” Che Guevara said. “What’s your point?”

  “My point? To truly break this dependence, we have to give more, all of us. Much more. As God taught Abraham and his descendants, the faithful only find their conviction with real sacrifice. That’s the wonderful irony of all this. All that we need to do is contained in a book written when faith itself was the most important technology of all.”

  “No Bible talk!” Gandhi yelled.

  “Yes, Moses,” Joan of Arc said. “We agreed, the rule is leave your personal ideologies at home. This movement, this revolution against the future we all fear, fails if we push our own beliefs instead of the common cause.”

  Moses locked his hands behind his back and stared back at her. “There is no we here. We trust each other only so long as we prove ourselves. I hope my actions demonstrate at least that I am done skulking around in the ether.”

  “Just take a breath, Moses,” said Lincoln. “Please. There is time to do this right so that we—”

  “Time for what? Or rather, for whom? I know what you’ve done, trying to profit off of my actions, like you have so many others’ righteous anger.”

  “What . . . I don’t know what you’re talking about,” said Lincoln.

  “Yes, you do.” The Moses avatar pointed a single long finger at each of the other avatars. “In your hearts, each of you knows what I am talking about. The only decision now is whether you will act to be on the right side of history or be ground under by it.”

  At that, he walked toward the blue shimmering wall of water and his figure dissolved.

  Potomac Overlook Neighborhood

  McLean, Virginia

  Jackson Todd pulled off the VR gloves and helmet, slapping them down on the white kitchen countertop, a white quartzite made to look like Carrara marble. They had remodeled the kitchen when they first bought the house. He and Isabella had actually argued back and forth about whether to get the real stuff. It was one of those stupid things that seemed so important at the time. Ultimately, they’d been swayed by the salesman’s pitch that the artificial wouldn’t show the stains of any morning coffees or kids’ fruit juices. He’d been proven right, the white was still gleaming. It showed no evidence of their years together, only the memories that just hung in the air above it.

  Todd ran his index fingers over each eyebrow, up his forehead, and then to the back of his neck, rubbing at the points of pain just where the skull ended. Stretching his neck one way and then the other, he slowly breathed in through his nostrils and out of his mouth. After repeating it nineteen more times, he stood and walked to the other side of the kitchen counter. In bare feet, his steps were careful, still mindful of stepping on the now-absent Legos that had once cluttered the floor.

  Pulling a long wooden match from the drawer, he struck it, lighting the gas burner on the stove and placing the kettle atop the heat. Watched pots never boiled, so as he waited, he stared out into the backyard through the kitchen windows. The grade of the lawn, slightly dipping lower to the right, was barely noticeable in the dark, but it still annoyed him, a defect he had always meant to remedy but had never gotten around to. Over it hung a limp zip line, running from one end of the yard to the other. It was set just high enough to appear dangerous to a young boy, but not so high that it really was. Green lichen covered the rope, unused for years now.

  When the built-in thermometer on the side of the kettle displayed 93 degrees Celsius, he took it off the burner. He then pulled the metal container from the cupboard, softly shaking it three times, side to side, to sift it one last time. He measured out 4 grams of loose-leaf black tea into the stainless steel mesh ball strainer—3 grams made the proper amount, but tonight he gave himself an extra gram for the needed focus.

  He placed the ball in a double-walled glass cup with a slight clink and poured the water over it, stopping when it reached a line carefully etched into the inside of the glass. Letting the tea steep, he walked over to a wooden cabinet in the den that the kitchen connected to and pulled out an old binder. Trailing his fingers over the faded plastic sleeves of compact discs inside, he pulled out the special one, her handwriting on the reflective shine still visible in black permanent ink. He powered up the old compact disc player he’d bought at a church sale, hit “eject,” and loaded the CD in the slide, pushing it back in with his finger, the tray no longer sliding back on its own.

  Eyes closed, he stood in front of the CD player as the song began. The slow guitar strum that he’d heard too many times to count but always made him think of that first slow dance at 1 a.m. in the basement of Princeton’s Colonial Club eating club: Mazzy Star’s “Fade Into You.”

  I want to hold the hand inside you

  I want to take the breath that’s true 22

  His body swayed slightly at the memory of their feet shuffling on a floor sticky from dried beer, two sophomores not knowing that it was the beginning of their lives together.

  As the song ended, four minutes later, he hit the eject button and placed the disc back in the sleeve. He returned to the kitchen and removed the tea strainer, placing it in the sink, a farmhouse-style one they’d also put in during the remodel. She’d wanted white to go with the countertop, but he’d liked the stainless steel more. She’d given in, a little victory she knew meant more to him than her.

  He walked back into his office, careful not to spill the steaming cup of tea.

  On one wall hung computer science diplomas framed in inexpensive dark wood frames, the kind bought in that purgatory after graduation when cash is short and ambitions are long. The BSE from Princeton beneath the PhD from Georgia Tech, each certifying that Jackson Todd had their blessing to remake the world.

  On the other side of the room, a series of citations and commendations for work at the Defense Advanced Research Projects Agency, including a framed personal letter from the agency director.

  Todd sat down at the desk, moving aside a vinyl-covered copy of the Old Testament bristling with the edges of yellow and blue Post-its. He carefully set the cup of tea on a coaster marked with DARPA’s old motto from its days dominating the world of information: Scientia Est Potentia, Latin for “Knowledge Is Power.” Like so much else, they’d actually gotten it wrong. Back in 1655, Hobbes had explained the true intent of the translation: “The end of knowledge is power . . . Lastly, the scope of all speculation is the performing of some action, or thing to be done.”23

  He would be that action, doing the thing that had to be done.

  The computer monitor on his desk winked awake, its screensaver a 3-D montage of video clips and still images from vacations, a preschool graduation ceremony, and trips to the ice cream shop. Todd moved down through the menu finder and then opened the file that he had saved for just this occasion. He put in his headphones, snaking their connecting cord into his monitor. As a graduate student, he had always coded to music, the rhythms calming his mind and driving him forward. Today, different background noise was needed.

  The file he played instead was a downloaded video of a local TV station’s news broadcast, one of those retrospectives they run on anniversaries and slow news days. It began with security-camera footage from the Metro stop at Ballston Station, the neighborhood a mix of condos and government agency buildings located just over the river from Washington, DC. Rush-hour passengers lined the subway platform, but they were difficult to distinguish individually due to the density of the crowd and grainy footage. Privatization of the DC area’s commuter line had brought sweeping infrastructure upgrades to the Metro’s beleaguered systems, but some corners had been cut to make the bottom line work. The quality of the security cameras was one, the automati
on integration of the track another.

  Moving that feed to a corner of the monitor, Todd sipped his tea and called up a new file. It opened with the logo of Princeton University’s Computer Science Department, then revealed line after line of software code.

  The video footage cut to passengers entering and leaving the Silver Line station at McLean, swirling around one another as they tapped their phones on the turnstiles controlling access to the platforms. That they still used handheld devices dated the video. Off to a side, a mother helped her son through the gates, jostled as she tried to tap again and again but seemed unable to get the system to register. It was a sequence Todd had seen hundreds of times, and relived thousands more, during the past seven years. It was maybe two seconds of footage, but the broadcasts always used it because it humanized the disaster. Was the gate malfunction a premonition, from the system itself, warning her? It was like that for Todd now—he had to make crystal clear for people the signs coming from the systems all around them, telling them to stop before it was too late. She just hadn’t known how to recognize them. Or maybe she had but wasn’t willing to listen to them. He couldn’t blame her. He hadn’t known then either. Now he did.

  Todd’s fingers began to type more furiously, finding their rhythm as the news segment continued playing in the background. The platform cleared, the mother and son dashed across the empty space hand in hand. They didn’t look harried, though. They looked happy by the way they raised their hands high in triumph as they sprinted through the train doors. On their way to meet up with Daddy for lunch at the pizza place just outside his work, where their son would get a cup of ice cream if he was really good.

  Signs that could not be ignored. That was how he was going to save those who were left, he thought, because it was missed signs that had cost him those who mattered the most. Sacrifice was required and he’d made that sacrifice first.

  Meanwhile, the news report continued—a voiceover from a journalist while they showed a camera view from inside the train that had left McLean, onrushing tracks framed by dark haze. The journalist’s voice marked the train’s path as its driverless system moved it toward the city. It stopped at East Falls Church, an elderly couple getting off. “The lucky ones,” the narrator intoned, as the train picked up speed again. The camera view shifted back to the Ballston station, as the crowd there pushed their way onto another train on the line. It slowly pulled out of the station. Before it made its way three cars into the dark tunnel at the end of the station, the screen went black and the narrator paused in silence, perhaps out of respect or perhaps just to let the viewers’ imagination manufacture the horrors of that instant.

  Despite the hot tea, Todd shuddered. But he kept typing, working the ADS code into what it truly needed to be. It had once been an architect of destruction. So it would be again.

  The voiceover narration transitioned from one of hushed reverence to cool analysis of the crash’s cause. A machine error, she said. A flaw in the code that was not foreseen in design, treating an act of code as if it were an Act of God.24 There was no one to blame.

  No, that was not true, Todd thought. We all were. It was that simple point that they had to understand.

  As if the cause of the tragedy could not be questioned further, the story then quickly shifted to the success of the emergency response. A Metro Automatic Train Control shift supervisor explained how the artificial intelligence running the overall train network ensured all the unaffected trains were safely stopped before further accidents and cleared the way for emergency personnel. Then drone footage from an Arlington police quadcopter flashed across the screen as the machine used its onboard cameras to route police and fire vehicles, while overriding traffic lights in real time. Above the station, media drones ghoulishly darted between buildings looking for unique angles, “allowing for much greater transparency and awareness about the everyday heroes and the official response to the tragedy,” the narrator said. Then a shaky video shot of a descending octagon-shaped platform, painted bright white. Eight trauma pods arrayed like spokes under the air ambulance’s platform, looking from the ground like a flying playground merry-go-round.25 Emergency medical technicians rushed the pods belowground, then quickly back up. A flashed image from one of the darting drones revealed the mother and her young boy being slotted in next to each other in the octocopter’s cradles. Adam and Isabella, no longer holding hands, their arms now limp.

  Todd opened up the Wi-Fi. It was the one connection he had to risk, but he’d limited it by tapping into a repeater set that the kids at Langley High School had set up a mile away, which broadcast out the Fairfax County school system’s 5G mesh network. If somehow someone did track it back, they’d still have only pinpointed it to a connection point hiding within Virginia’s most populous county.

  Flicking back to the window of ADS software code still open on his screen, he selected the now modified subroutine of new version updates, which regularly pushed out patches and upgrades across the system. He accessed the directory of current users, his computer noticeably slowing as it downloaded millions of potential targets.

  Georgetown Neighborhood

  Washington, DC

  There was a slight hum as the FBI Suburban’s electric motor stirred to life, advancing the car 2 feet before shutting off again with a sigh—the typical sound of midafternoon traffic in downtown Washington. Keegan wasn’t sure why the vehicle was heading north on Wisconsin Avenue, when she wanted to be heading south to 695 to cross the Anacostia River. Probably no human in the entire traffic jam knew why. Maybe not even the traffic management system that had steered them all this way knew. Whatever the case, it didn’t matter; the ride along was the point, not the destination.

  Keegan looked over at TAMS, sitting shotgun. There was no set Bureau protocol to where it rode. She could have had the machine lie down in the trunk for all it cared. In a seat, though, it would see what she saw. Plus, they wouldn’t get any dirty looks being in the HOV lanes.

  The only problem was that, if the idea was to teach it like Modi advised, TAMS didn’t seem to be paying attention. Its power-saving mode made it stir to life every time the SUV advanced, then power down again when the car stopped.

  “TAMS, change power-management cycling; constant on,” Keegan said. “OK? No sleeping on the job.”

  “Power-management cycle change confirmed.”

  That was also annoying. Whoever had decided the settings for commands had gone for specificity, but it was going to slow them down. “TAMS, change command settings. When it is just you and I, statement of TAMS is not required. Initial authorization assumed.”

  “Command setting change confirmed.”

  “And you don’t have to confirm with re-statement of the order. Just say, ‘OK.’ Understood?”

  “OK.”

  Keegan laughed. Who says software can’t have a sense of humor?26

  “Alright, good,” she said. After a beat of not knowing what else to say to the machine, she got impatient. “I’m going to drive; this is going too slow.” She flicked the wipers to clear the faint layer of dust on the windshield and then put her vizglasses down on the center-console charging pad.

  The traffic was moving just slightly faster in the left lane, and she spied her target, a six-wheeled UPS delivery truck. She manually nosed her SUV out in front of the truck, and it immediately slammed on its brakes and then released its flock of delivery drones—standard collision protocol to avoid damaging the onboard cargo.

  “Bingo,” Keegan said under her breath, congratulating herself for the double score as the drones circled overhead. Liability policies, risk committees, and shareholders all shaped the vehicle’s code with safety in mind, which, in turn, created a gap for her to exploit. This was something Modi would have better understood if he’d ever spent time fighting people whose entire strategy was to take advantage of the gaps that happened when someone tried to fit society’s rules into an algorithm. Machines were built on assumptions that meant one thing in a la
b and another in a narrow alley reeking of dog shit and gunpowder.

  They drove west, Keegan following the algorithmically determined route, but at least she was now in charge. It took them along residential streets, a well-off neighborhood of larger homes with elaborate fortifications. Some were subtle, like Kevlar webbing over the hedges, while others were as gaudy as you might find in a Moscow suburb, with sensor clusters atop Roman columns.

  What did the robot really know about what that meant? Geolocated images and aggregated data revealed a little about the neighborhood but missed the critical question of why the property owners would go to such lengths.

  Having gotten it out of her system, Keegan set the SUV back on autopilot but began tapping the dash screen as more data appeared. The heads-up display kept flagging diverters concealed behind bushes or under false rocks. The devices pulsed false data into the traffic-management and autonomous vehicle navigation networks, trying to keep a rich neighborhood’s streets largely free of cut-through traffic. Local police looked the other way because everybody did it. The only people who really cared were the share transport companies, who were rumored to be paying task-worker bounties to remove them. But the devices cost no more than a case of wine and so were easily replaced.

  When they finally got back onto the freeway, the SUV hit a massive pothole underneath a bridge and slewed to the right. Keegan slapped her hands back on the wheel for an instant, but the vehicle corrected its course and they continued on. They soon crossed over the Anacostia River, its color a green so dark it could have been government issued. They hadn’t gone a block when they got caught in another snarl of traffic. This one wasn’t created by a machine algorithm, though. Ahead was a checkpoint backed by a black DC police MRAP.27 Burning flares and LIDAR reflectors, arrayed like a spray of red flowers on the black asphalt, diverted traffic.