The paradox of efficiency means that as the web tightens it grows more vulnerable to small disturbances—disruptions and delays that can cascade through the system for days. For example: American Flight 1128, inbound from Mexico, is now forty-four minutes late, and the computers are deciding whether to delay some of the connecting flights those passengers will be racing toward. This, too, will be a real-time decision based on complex modeling. The computer will know how many people are how many minutes late for each flight. It will consider the distance to the gate, the time before the next available flight to the same city, the likelihood of new delays at the other end. It will consider the passengers, too—if they have paid for first-class tickets, they will be more likely to find the gate waiting open for them. Pilots often accuse Nason and his computers of being overly fixated on time. “They ask, how can you close the door on a passenger running from three gates down?” says Nason. “Well, there are 130 people on this airplane looking at their watches.”

The paradox of efficiency means that as the web tightens it grows more vulnerable to small disturbances—disruptions and delays that can cascade through the system for days. For example: American Flight 1128, inbound from Mexico, is now forty-four minutes late, and the computers are deciding whether to delay some of the connecting flights those passengers will be racing toward. This, too, will be a real-time decision based on complex modeling. The computer will know how many people are how many minutes late for each flight. It will consider the distance to the gate, the time before the next available flight to the same city, the likelihood of new delays at the other end. It will consider the passengers, too—if they have paid for first-class tickets, they will be more likely to find the gate waiting open for them. Pilots often accuse Nason and his computers of being overly fixated on time. “They ask, how can you close the door on a passenger running from three gates down?” says Nason. “Well, there are 130 people on this airplane looking at their watches.”