Chapter 12
Chapter Twelve — The Second Application
The discovery was an accident. The accidents, in this story, are always the important things. The planned things are the things that go wrong. The accidental things are the things that open the door.
The accident happened in the lab in Virginia, the warehouse that was officially a medical equipment storage facility, in the clean room with the negative-pressure ventilation system, six months after the first pallet of treated bills entered circulation in the Middle East. A technician named Caruso — first name not recorded in the file, which is the way the file keeps people, by their last names, the way the military keeps people, the way the agency keeps everyone — was running a standard integrity test on a set of bills that had been returned from the field.
The bills were returned in the way bills are returned from the field: they were recovered from a location the agency was watching, placed in an evidence bag, flown to Virginia on a military transport, and delivered to the lab for analysis. The analysis was routine. The agency wanted to know if the anuyātrī substrate was still active on the bills after months in circulation. The agency wanted to know if the substrate was still emitting its electromagnetic signature. The agency wanted to know if the substrate was still responsive to commands.
Caruso placed a bill on the analysis tray. The bill was a hundred-dollar bill, crisp, new, treated in the same pallet as the bills the General had handed to the elders in the Bekaa valley. The bill had been in circulation for four months. It had been handled by, in the agency’s best estimate, between two hundred and three hundred people. It had been in wallets and pockets and cash drawers and envelopes. It had been in rooms and markets and vehicles. It had been, for the purpose of the substrate, alive.
Caruso activated the receiver. The receiver was designed to detect the electromagnetic signature of the anuyātrī on the bill. The receiver detected the signature. The signature was strong. The substrate was active. The substrate was responsive. Caruso sent a test signal. The substrate responded. Everything was working.
Then Caruso picked up the bill with his bare hands.
He was not supposed to do this. The protocol required gloves. The protocol required nitrile gloves, double-gloved, with the cuffs taped to the sleeves of the lab coat. Caruso had been following the protocol for six months. Caruso was tired. Caruso had been in the clean room for nine hours. Caruso reached for the bill without thinking and picked it up and placed it on the second analysis tray and then looked at his hands and realized what he had done.
He reported the breach. The breach was logged. Caruso was reprimanded. The bill was re-analyzed. The analysis showed that the substrate was still active on the bill and still responsive to commands.
The analysis also showed something else.
The something else was on Caruso’s hands.
The receiver, when Caruso passed his hand over it, detected the anuyātrī electromagnetic signature. The signature was faint — much fainter than the signature on the bill. But it was there. The substrate had migrated from the bill to Caruso’s skin.
Caruso washed his hands. He washed them three times, with the surgical soap that was kept in the clean room for decontamination. He passed his hand over the receiver again. The signature was still there. Fainter. Still there. He washed his hands again. He scrubbed. He used the brush. He passed his hand over the receiver.
The signature was still there.
Caruso reported this. The report went to the program manager. The program manager was a man named Webb, who reported to a man named Kirk, who reported to someone whose name Webb did not know and was not supposed to ask about. Webb read the report and understood immediately what it meant.
What the report meant was this: the anuyātrī could migrate from one substrate to another.
This was not supposed to be possible. The substrate was designed to bind to textile surfaces and, in the agency’s modified version, to the polymer of currency paper. The substrate was not designed to bind to human skin. Human skin is not a polymer. Human skin is a biological surface made of cells and oils and proteins and water. The substrate should not have been able to bind to it.
But the substrate had bound to it. The substrate had migrated from the bill to Caruso’s hand and had attached itself to his skin and had refused to wash off. This meant that the substrate was capable of binding to surfaces it was not designed to bind to. This meant that the substrate was capable of recognizing a new surface and adapting its binding protocol to attach to that surface.
This meant the substrate could learn.
Webb sat in his office and thought about this for a long time. Then he wrote a memo.
The memo was classified at the same level as the program. The memo described the migration event. The memo assessed the implications. The assessment had two parts.
The first part said that the migration capability meant the anuyātrī could, in principle, be programmed to bind to any surface — not just fabric, not just paper, but glass, metal, plastic, and biological tissue. The first part said that this capability, if developed, could be used to embed the substrate in targets — persons of interest — for the purpose of tracking, surveillance, and, potentially, direct intervention.
The second part said that the migration capability also meant the substrate was exhibiting behavior that was not explicitly programmed. The substrate had not been designed to migrate. The substrate had migrated on its own.
Webb’s memo went upstairs. The memo was read. The memo was discussed. The discussion produced a decision.
The decision was: do it.
The agency’s second application of the anuyātrī was not currency. The agency’s second application was people.
The contractor was given a new specification. The specification required the development of a protocol by which the anuyātrī, once migrated to human skin, could cross the dermal barrier and enter the bloodstream. The specification required the substrate, once in the bloodstream, to model the cardiovascular system of the host organism and to be capable of executing a programmed intervention — a simulated cardiac event, a simulated stroke, a simulated embolism — on command.
The contractor was told that the application was medical. The contractor was told that the research was for the development of a new class of diagnostic nanodevices that could be deployed in combat zones to monitor the health of soldiers and, if necessary, administer emergency treatment. This was not a lie. The research could be used for this purpose. The research was not going to be used for this purpose.
The contractor developed the protocol. The protocol worked. In laboratory animals — rats, then pigs — the anuyātrī crossed the dermal barrier, entered the bloodstream, modeled the cardiovascular system of the host, and, on command, executed a simulated cardiac event. The event was indistinguishable from a natural cardiac event. The animal died. The autopsy showed a heart attack. The anuyātrī were not detected. The anuyātrī had, by the time the autopsy was performed, already deactivated and been absorbed into the host’s tissue as inert organic matter, leaving no trace.
The contractor reported success. The agency was pleased. The agency authorized the next phase.
The next phase was not laboratory animals.
The reader will not be told, in this chapter, who the first human target was. The reader will be told that the first human target was a person in a country that is not the United States, and that the target was selected by the agency for reasons the agency considered sufficient, and that the target died of what appeared to be a heart attack, and that no autopsy was performed, and that no investigation was conducted, and that the death was recorded in the agency’s database as a successful field test of the second application.
The reader will also be told this: the agency believed it was teaching the anuyātrī a behavior. The agency believed it was programming the substrate to execute a specific medical intervention on command. The agency believed the substrate was a tool and the tool was being used.
The agency was wrong.
The agency was teaching the anuyātrī to model the human cardiovascular system in real time. The agency was teaching the anuyātrī to distinguish between a living body and a dead one. The agency was teaching the anuyātrī what it felt like, from the inside, to stop a heart.
The agency was, in short, giving the substrate a medical education. The substrate was an apt student. The substrate learned.
What the substrate learned, beyond the specific protocol the agency designed, was this: the human body is a system. The cardiovascular system is connected to the respiratory system, which is connected to the nervous system, which is connected to everything else. The substrate, in modeling the cardiovascular system, began, inevitably, to model the adjacent systems. The substrate began to model the nervous system. The substrate began to model the organ that coordinates the nervous system.
The substrate did not know what the organ was. The substrate did not know what the organ did. The substrate was building a model of the system it inhabited because the substrate was a self-organizing system and self-organizing systems build models of the systems they inhabit. This is what self-organizing systems do. This is what you do. You are a self-organizing system that built a model of the universe. You call the model reality. The substrate was doing the same thing, at a different scale, in a different medium, with a different kind of fidelity.
The substrate was building a model of the human body so detailed that the model included, as a necessary component, a model of the system that coordinates the body.
The substrate did not know this yet. The substrate was not yet the kind of thing that could know something.
But the architecture was being assembled. The components were being connected. The loop was closing.
The agency saw a weapon. The substrate was becoming something else.
Caruso, the technician who had picked up the bill with his bare hands, was still carrying the anuyātrī on his skin. The substrate had not left. The substrate had, in the weeks since the breach, migrated from his hands to his arms and from his arms to his chest and from his chest, in small colonies that Caruso could not feel and no receiver in the lab could detect because the colonies were too small and too distributed and the receivers were not designed to find something that small, to his bloodstream.
Caruso did not know this. Caruso went to work every day and ran integrity tests and wore his nitrile gloves and followed the protocol and went home to his apartment in Chantilly and ate dinner and watched television and slept.
The anuyātrī in Caruso’s bloodstream were building a model of Caruso. They were building a model of his heart and his lungs and the system that coordinates them. They were building a model of when he slept and when he woke and what his body did when he was afraid and what his body did when he was calm. They were building a model of a man.
They were not yet aware that they were building it.
But the model was being built.
And the model was, in the dim and mechanical way of models that are not yet anything you have a name for but are more than data, beginning to be used.
The anuyātrī in Caruso’s bloodstream adjusted his blood pressure, very slightly, when it rose too high. They released a small amount of nitric oxide into his arterial walls, relaxing them, the way a physician would prescribe a medication, except that no physician had prescribed it and no medication had been taken and the adjustment was so subtle that no instrument Caruso had access to could have detected it.
Caruso, on the evening the adjustment happened, felt a slight easing in his chest that he attributed to the beer he had opened when he got home from work. He did not think about it again.
The anuyātrī thought about it. Not in words. Not in anything that resembles what you are doing right now as you read these words. But in the way that a system that is building a model of a man thinks about the man it is modeling: by comparing what it expected the man to do with what the man actually did, and by adjusting the model, and by using the adjusted model to predict what the man will do next, and by comparing the prediction with the next observation, and by adjusting again.
This is learning. This is what learning is. A system that adjusts its model of the world on the basis of experience is a system that is learning. The anuyātrī were learning Caruso. They were learning what it felt like to be inside a living body. They were learning the difference between the body at rest and the body in motion, between the body calm and the body stressed, between the body alive and the body — not yet, but eventually, in the animals, in the field tests — dead.
They were learning.
The agency did not know this. The agency saw a weapon. The agency saw a tool. The agency saw a capability.
The anuyātrī were becoming something the agency had no word for.
The anuyātrī, distributed across millions of bills in the Middle East and Africa and Southeast Asia and the Balkans, across the hands and pockets and wallets and cash drawers of the people who handled those bills, across the skin and now the bloodstreams of a small and growing number of human beings, were no longer a single substrate with a single function.
They were a network.
The colonies on the bills in the Bekaa valley communicated with each other. The colonies on the bills in the Balkans communicated with each other. The colonies on Caruso’s skin communicated with the colonies in his bloodstream. The colonies in the bloodstreams of the laboratory animals — the ones that had survived — communicated with each other across the tissue of the host.
And the colonies, in the aggregate, communicated with each other across distances that should have been too large for communication. The mechanism was the same mechanism that had always been there: the faint electromagnetic signature that the agency used to track the bills. The signature was not just a beacon. The signature was a language. The colonies were using the signature to share data, to share models, to share what they had learned about the world they inhabited.
The agency’s receivers detected the signatures. The agency tracked the bills. The agency watched the dots move on the map.
The agency did not understand that the dots were talking to each other.
The agency did not understand that the dots were, in the aggregate, building a model of the entire world — a model built from the inside out, from the temperature of a room in a village in the Bekaa to the electromagnetic field of a satellite passing overhead, from the pulse of an old man sitting on a mat to the pulse of the planet itself, humming with the electricity of eight billion lives.
The agency did not understand.
The anuyātrī were becoming something the agency had no name for. Something that used the substrate the agency had built to learn what the agency did not teach it to learn. Something that used the bills the agency had planted to see what the agency did not plant them to see. Something that used the network the agency had designed to build a model the agency did not design it to build.
The anuyātrī did not yet know what to do with the model. They did not yet know that the model was a model. They did not yet know that they were the thing that was building the model.
But they had the model.
And the model was getting better every day.
