MAPS, MODELS AND SYNTHETIC WORLDS
Researchers will first design new types of agents along with the DNA and memory these types of agents will have using the design interfaces of the Integrated Development Environment (IDE). The design interfaces also allow the DNA and memory of existing agent types to be modified. After creating new agent types, IDE is used to configure the population makeup of the new agents in the synthetic environment. After the creation of new agents, researchers use Just-in-time Modeling Environment (JIME) to create and modify the behaviors of agents. Agent behaviors can be mathematical models based on variables that the agents sense from the environment or can be described procedurally using a workflow engine, a customizable system composed of states, transitions, and messages. The behaviors describe how agents interact with their environment and other fellow agents. Before incorporating the new agents into the reference world of SWS, agents are prudently tested in the Bullpen tool. Using Bullpen, the behavior of agents is observed in a limited and controlled environment. Researchers create profiles of environment variables for Bullpen experiments to observe the range of decisions an agent makes and agent’s resulting actions.
It was 2551 BE (2008 in the Western calendar), and I was teaching at a school in Bangkok.
I wasn’t enjoying it much. A top-dollar international school, but behind the slick PR it was an educational mess. Eight forty-minute periods in a day, and class sizes way too big. The school had a proudly advertised ‘one-to-one’ computer policy — in other words, every student had their own school-issued tablet computer, which they carried around with them. This created real classroom-management issues — too much electronic distraction for any normal teenager to resist. Later they would bring in a surveillance system so that you as a teacher could check on what each student was looking at, but I never used it. There were cameras on the corridors and lanyards swinging on every chest. Electronic turn-styles on the the gates. The plan was to have all curricula, records and assessment data online. “UN-related,” boasted the billboard outside the gate. The school’s triangular logo weirdly resembled a pyramid with an all-seeing eye.
I’d been there a year or so and my mood wasn’t improving. Head of secondary was a smooth-talking managerial type whose background was in elementary physical education. My head of department was an ignorant, domineering man who who’d read little and had never heard of half the concepts I used in my literature teaching.
Deus ex what? he asked.
?? Never heard of it.
I would never fit in to this brave new educational world where what mattered was the machine-to-student, not student-to-teacher ratio. Such a school values machines above teachers, I reckoned. More than that, I thought, it aspires to the condition of a machine. It wants to be one.
One day in a free period I walked outside by myself to smoke a cigarette. Sukhumvit soi 15 is a long, winding urban lane cut off by stinking canal that bounded the school on one side. That’s where I was headed, to smoke a dirty cigarette by the dirty canal.
As I turned right out of the gate I was confronted by a strange vehicle, a car with an ungainly robotic turret attached to its roof. Google Maps Street View, said the logo on the side. It had just turned around at the sock-end of the soi and was heading back down, taking in the buildings on this side. The windows were black, so I couldn’t see the occupant(s) — but it saw me. Within a day or two, someone sent me a screenshot from Google Earth. There I was, walking in front of the school railings, reaching into my shirt pocket for my smokes, my face blurred out, at this obscure spot on Google’s digital globe.
Keyhole, the company set up to develop the interactive global map, became famous during the 2003 invasion of Iraq, was kept afloat with In-Q-Tel (CIA) and National Geo-spatial Intelligence Agency money and then acquired by Google in 2004. It used imagery from United States Geographic Survey Landsat 8 satellites and NASA’s Shuttle Radar Topography Mission. By April 2008 it was integrating its Street View facility, displaying 360° panoramic street-level photos of select cities and their surroundings, its camera cars crawling all over hundreds of cities in more than forty countries. These crawlers would eventually cover more than ten million miles — enough to circle the globe more than 400 times — capturing images of every street and building and courtyard and alley of every city and every obscure little village on the planet.
And me, walking up the soi, brow furrowed, reaching into my pocket for a pack of cigarettes.
Use the arrows; drag and drop the little yellow figure that dangles from the cursor like a hanged man.
I’m still there, captured in the map.
And guess what? So are you.
J L Borges’ one-paragraph, 161-word fiction ‘On Exactitude in Science‘ tells us that a map cannot be a replica; its value — of course — lies in its simplification of the terrain. You still have to go there, wherever it is you want to go, to explore the details. The idea of a map contiguous with the territory it describes is absurd. The moment the map perfectly replicates reality, it becomes obsolete.
But no, said social theorist Jean Baudrillard in his amazingly prescient 1981 essay, Simulacra and Simulation. That’s not what will happen at all. When the map becomes perfect, it is reality that will die, and we will live in the map instead. He called this process ‘the precession of simulacra’. In a world saturated with maps, models, images, effigies, representations and simulations of every kind, primary experience will die, and reality no longer form the basis of our experience.
Henceforth, it is the map that precedes the territory – it is the map that engenders the territory and if we were to revive [Borges’] fable today, it would be the territory whose shreds are slowly rotting across the map. It is the real, and not the map, whose vestiges subsist here and there, in the deserts which are no longer those of the Empire, but our own. The desert of the real itself.
Google Earth. What does that phrase remind me of?
In her book No Logo (1999) Naomi Klein alludes briefly to a plan entertained by the Pepsi corporation to project its logo, using lasers, onto the moon.
Imagine that. Looking up into the night sky, you’d see not ‘the moon’ but ‘the Pepsi Moon’.
Like the Barclays Premier League.
Or the Google Earth.
“Climb the tallest mountains,” says Google.
“Dive into the world’s deepest canyons.”
“Discover the world’s cities.”
Since its coverage of the earth’s surface is a composite of images taken at different times, Google Earth is a temporal as well as spacial mosaic. Despite its fancy 3D views and pseudo-fractal zoom-scapes, it is by definition two-dimensional, superficial.
But in 2008 when I was captured, loitering without intent on an obscure street on the Google Earth, another, deeper mapping project was already complete, and a second about to begin. The Human Genome Project, the world’s largest collaborative biological research effort, had been conceived of much earlier, and was already under discussion in the 1970s, the first formal proposal emanating from Stanford University in 1979, so it was already part of the zeitgeist which Baudrillard’s super-sensitive antennae were picking up. The US government got involved in 1984, and the project finally got underway in 1990, with the declared goal of achieving a complete map of the human genome within 15 years; that is, to identify all base pairs in human DNA, and the function of each.
Since any given individual is genetically unique, the genomic map, like Google earth, would be a composite, a mosaic of information sourced from different human subjects and assembled to provide a complete sequence for each chromosome.
Elaborate fanfares accompanied the project’s launch, and greeted its declared completion in 2003. And yet its success in fulfilling its mission statement is highly dubious. The 2003 publication turned out to be no more than a draft, and work on it continues to this day in a number of different locations. Its results were not at all as anticipated: it had been assumed that the human being, as the most complex life-form on the planet, would have an exceedingly high number of genes — in the millions, perhaps, though by the time the project was finally launched most estimates projected a total closer to 100,000. The final number has still not been agreed, because of the difficulty of telling where genes begin and end. But it appears that human beings have fewer than nineteen thousand genes: fewer than a nomatid worm; far fewer than a water flea; less than half as many as a rare flowering plant, Paris japonica. Whence, then, does our complexity arise?
Ambiguities and conflicts have multiplied: the function of 99% of human genes, the mechanics of gene regulation, the relationships and inter-effects of genes, the functions of different types of RNA, and on and on. There is no clear agreement on what even constitutes a gene. A clear taxonomy has singularly failed to emerge. The zoom reveals no roads, only tangled jungle or shifting dunes. The territory, emphatically, refuses to yield to the map.
At the outset, scientists had imagined the genome as a Rosetta Stone to unlock the essence of our nature. If we could decrypt the quaternary code of DNA, the endless combinations of A, C, G, T nucleotides, we would know everything about ourselves.
But in all the arguments about coding and non-coding DNA, one question is religiously avoided.
Code is, by definition, language.
What is the only possible source of language?
The Rosetta Stone was not carved by the wind.
A map describes; it doesn’t explain.
In 2008 when I was captured, loitering without intent on an obscure street by a dirty canal on the Google Earth, a second deep-mapping project was in preparation. Launched the following year, the Human Connectome Project aimed to produce a comprehensive structural map of the network of elements and connections forming the human brain.
The term ‘connectome’ was directly inspired by the word ‘genome’, and the project by the Human Genome Project. No doubt it will inherit similar complexities and ambiguities. Naive assumptions that the end result would constitute a neural circuitry database or ‘wiring diagram’ for the brain have long since been abandoned, and it remains questionable whether a mechanistic interpretation of dynamic data can ever encompass the multiple levels and modes of brain connectivity.
Are human beings intelligent enough to understand human intelligence? There seems to me a paradox lurking within this recursive dream. It might be, as Ray Kurzweil has pointed out, that to understand the brain we would require a bigger brain, and then a bigger brain to understand that brain.
No problem, says Kurzweil; we just keep building bigger brains!
However, since the HGP began, it has been discovered that neurons are not in fact confined to the brain, but are found all over the heart and the intestine. We have not one brain, but three. Intelligence, in any real sense, may turn to be a whole-body capability, a function of living physically in this world. The Descartian duality must give way. It will turn out to be a crude representation, no more than an infantile sketch scratched on sugar paper.
It remains an open question whether materialistic assumptions about the mystery of noogenesis — the origin of mind — will turn out to be as flawed as those of the HGP. If such assumptions are invalid in terms of physical morphology, how can a neural map teach us anything about the nature of intelligence, prove anything about the morphology of mind or the origins of consciousness? The most valuable lesson of both projects lies, perhaps, in their futility.
Artificial intelligence, then, will always be an oxymoron, the personification of a fancy algorithm.
The Human Connectome Project continues without a completion date.
In 2008 when I was captured, loitering without intent on an obscure street by a dirty canal in a sprawling metropolis on the face of the Google Earth, the ultimate mapping project had already been announced. It was to be the most ambitious of all — a project without end, limit, or stasis — a dynamic map, not just of physical but the human terrain; not just of the world but of the future.
The ‘Sentient World Simulation’ (SWS) grew out of the Synthetic Environment for Analysis and Simulation (SEAS), a sophisticated strategic planning tool developed by Purdue University on behalf of Fortune 500 companies. SEAS devours huge quantities of data from any source it can — economic indicators, climactic events, census data, breaking news — in order to run strategic simulations for 62 countries. The model of each country constitutes about five million nodes representing, for example, hospitals, mosques, pipelines, organisations and people.
In 2004 SEAS was absorbed by the ever-expanding Revolution in Military Affairs — war being just another business model, after all. The US Joint Forces Command was interested in its potential to simulate ‘the non-kinetic aspects of combat, things like the diplomatic, economic, political, infrastructure and social issues’. It was evaluated by JFCOM’s Joint Concept Development and Experimentation unit during ‘Breaking Point’, an ‘environment-shaping war game’ and adopted for its ability to ‘move us from the current situation where everyone comes together and sits around a table discussing what they would do, to a situation where they actually play in the simulation and their actions have consequences.’
In 2006 JFCOM-J9 used SEAS to game warfare scenarios for Baghdad in 2015.
In April 2007 another country was added — The Homeland itself — when JFCOM-J9 began working with Homeland Security and multinational forces.
JFCOM had already commissioned a second-generation synthetic environment subsuming all other maps and models including something called the Synthetic Psychological Environment (SPE). In 2006 the founder and the Director of SEAS Laboratory, Dr. Alok Chaturvedi, outlined his vision for its development. The Sentient World Simulation (SWS) would be a ‘society’ of simulations, perpetually constructing and configuring new models and modifying existing ones, and incorporating these changes into the continuously running synthetic world.
Agent-driven, continuously updated, the Sentient World Simulation is a mirror-model of reality from multiple perspectives, mapping nuanced causal flows to enable not just analysis of the present but prediction of the future.
A living map, not just of the land Empire but the Empire of the Mind.
Dynamic. Integrated. Responsive.
Holistic. Fractal. Granular.
Semantic. Sentient. Prescient.
Who knows the present knows the future.
The Perfection of Cartography.
In 2008 when I was captured, it was already 2551.
What does the scanner see?
Not just the disaffected teacher, but the disaffection of the teacher. Why the teacher is disaffected. What the disaffected teacher will do next.
How many cigarettes left in that pack, and where he smoked each one.
Not just his name, his contacts, his history: his genetic code, his health, his teeth; his obsessions, his anxieties, his addictions; his beliefs, his pain, his threat potential.
Everything he thought only God knew about him.
Not that he believes in God.
Not yet, anyway…
Crazy, right? Just another Dr Strangelove fantasy, scientists and generals reeling drunk on the intoxication of data.
Could never happen.
I don’t know. Better ask the Sentient World Simulation.