Sensoria. Маккензи Уорк
Чтение книги онлайн.
Читать онлайн книгу Sensoria - Маккензи Уорк страница 13
The image is an effect of an algorithm, which may be a proxy for something, but it may have generated that for which it is a proxy through scripted operations. These operations model in code what it is they are supposed to double, whether it is a population or the subject of an individual photograph. What computation produces as signal out of noise is generated by a probabilistic template of what ought to be there: “Likeness becomes subject to likelihood.”165
The business of producing image and information proxies is of course immensely gameable.166 There’s all kinds of proxy cold war going on, not all of it the fault of “The Russians.” Indeed, “The Russians” are now a proxy stand-in for the whole crazy game of information warfare, fought as often as not with noise. There’s all sorts of actors, acting through stand-ins, duplicates, dupes, sock puppets. It is what the situationists called détournement on an epic scale, producing what elsewhere I called the spectacle of disintegration.167
Global civil war contains shooting wars too, also fought through proxies. These are the military equivalent of shell companies. “The border between private security, private military company, freelance insurgents, armed stand-in, state-hackers and people who just got in the way has become blurry.” It’s not so much a deviation from a norm as the new normal.168 “To state that online proxy politics is reorganizing geopolitics would be similar to stating that burgers tend to reorganize cows.”169
Steyerl: “Not seeing anything intelligible is the new normal. Information is passed on as a set of signals that cannot be picked up by human senses.”170 The critical approach is less about interpreting hidden power structures underneath an orderly culture as it is a practice of questioning the routine habits of apophenia, the selecting of patterns in random data. Your camera’s computer detects what it thinks you want to see in the noisy data captured by its lens. (Kittens!) Your social media service detects what it thinks is acceptable content amid the dick pics. A deadlier version is National Security Agency’s Skynet, trained to find “terrorists” in cellphone data from Pakistan. But were the thousands killed by missile and drone actually terrorists?171 Hard to say, as there’s no empirical test or benchmark for Skynet’s procedures.
Perhaps it’s a matter of finding different patterns, based on different protocols. Steyerl wants to show the connection between the design of death and the design of life. Her emblem is what Harun Farocki called the suicide camera, or what I once called missilecam, the nose-cone camera sending signals of its progress as it approaches its target.172 Steyerl: “the camera was not destroyed in this operation. Instead, it burst into billions of small cameras, tiny lenses embedded into cellphones.”173 Now we are overrun with the fallout of zombie cameras that failed to die.
The camera may once have framed the world as if it were there to be made into a picture for a person. But now humans are just part of a landscape that machines picture for other machines. “If the models for reality increasingly consist of sets of data unintelligible to human vision, the reality created after them might be partly unintelligible for humans too.”174
Who can forget the internet weirdness of Google’s Deep-Dream images, which reveal the presets of machinic vision and yet which managed to visualize the unconscious of circulation, with added cuteness? As Sianne Ngai reminds us, the cute can also have its scary side, as when DeepDream decides that what emerges out of a plate of spaghetti and meatballs is a series of disembodied puppy-heads. It had a habit of recognizing patterns that aren’t there. “It demonstrates a version of corporate animism in which commodities are not only fetishes but morph into franchised chimeras.”175
Humans are an inconvenience for machines. It’s a commonplace to think of work as turning humans into robots, but the humans always seem to remain repulsively mammalian. One way that computation has resolved the human into the world of the machine is through games. Alan Turing’s famous Turing Test was a way of deciding if something is human.176 If we think the way something communicates with us is human, then it is. He based it on a parlor game, involving the guessing of someone’s gender just from notes passed under the door.
John von Neumann tried to formalize the whole problem of decision and decidability with game theory. If we can simplify the stakes and certain tricky concepts such as rationality, utility, and information, then decision can be a science. Following Philip Mirowski, Steyerl argues that the difficulty of human decision-making was resolved in economic theory by taking humans out of the equation.177 If the rationality of humans is a problem for economic theory, replace humans with computation, with rational nonhumans, and let them loose on the world. “It is striking how much reality has been created as a consequence of different iterations of game theory.”178
Steyerl: “The point is that games are not a consequence of computers making the world unreal. On the contrary, games make computers become real. Games are generative fictions.”179 Steyerl borrows my term for this: we live in gamespace.180 “So, regardless of whether humans ever were ‘rational’ in the way game theory assumed, a lot of people have now been trained to understand rationality in this way and to imitate its effects.”181 Gamespace becomes more real, because more rationalizable, than the world it was supposed to model. And we are now all inside it, along with everything else.
For a while, we were all obliged to pass online reverse-Turing tests to prove to machines that we were something like a human. Captcha, which made you write out letters you saw in a fuzzy picture, tested whether you could impersonate a human to a machine, not whether a machine can impersonate a human to a human. This is no longer necessary now that Google has a scripted operation that models what a human is.
Computational models can decide not only whether you are human enough to be online but what you will want to do when you get there and what you will like. These models have an aesthetic dimension, in that they model your taste as a kind of ideal form and serve you with things that are like what it presumes you are like. In everyday aesthetics as in economics, models rule. Whether it’s a fashion model or a financial algorithm, the universe of forms is a sort of Platonist ideal that becomes the world, becomes gamespace.
Only the art world appears to have a solution to this. Attempts have been made to manage art algorithmically, such as ArtRank, which advises investors on which art to buy based on its proprietary metrics. But more usually, the autonomy of the art world banishes the ideal, the model, the beautiful form, from the world and quarantines it in museums. Where once the avant-gardes wanted to unleash the beauty of art on life, we may now count it as good fortune that at least one species of ideal, elegant, beautiful form is kept separate from the world—that of art.
Looking back out at the world from the quarantine of the museum, the task for humans is now to understand how machines picture the world. “Maybe the art history of the twentieth century can be understood as an anticipatory tutorial to help humans decode images made by machines for machines … Mondrian is perhaps an unconscious exercise for humans trying to learn how to see like a machine.”182
Museums are not what they used to be, however. In Benedict Anderson (or before that, in Harold Innis), the space of a nation-state could be regulated by the space-binding media of printed newspapers and also by the time-binding media of the museum.183 But the national museum is now flanked by other phenomena. Consider Freeport art storage, where art remains permanently in