The Stack On Software and Sovereignty

1. Animal User

For McLuhanite prostheticization, we invest and amplify agency by projecting it away from ourselves and into specific parts of the world, such that our embodiment might be augmented by the capacities of those objects (a hammer ’ s solidity) just as that object is augmented by our soulful agency (the hammer knows what to hit). In this cooperation, the world folds into the extended will and wisdom of the human body, but today, the reverse of this fold is perhaps a more critical process for exploration. The biological informationalism of the universal and absolute User also turns the animalhuman body into another kind of matter, leading to a decentering disenchantment of its agency. Instead we watch the world fold into the body, now itself a prosthetic extension of that world, correlated and piloted by it. In this reverse prostheticization, the human figure is set in motion from some other position in the field; the subject becomes object, self becomes substance, body becomes metabolic reserve, food machines consume you. Even for such multipolar mutual extension, the influence and self-identity of human subjects are by no means excluded (just look around you), but it is made conditional by multidirectional chains of association and by the recognition of previously inert media as potential colleague Users. In particular, animals are one key metabolic reserve that have been essential, if not always willing, enablers of human evolution. Proteins in their flesh were consumed for ours, their skins became our clothes and allowed migration to new places that we explored while riding on their backs on our way. The terms of this relationship already seem ugly. As the unfolding disenchantments brought by genetics, neuroscience, and evolutionary biology continue to undermine the human/animal distinction (according to Agamben, critical to the historical provision of political sovereignty onto the former), the contested space of animalian identification has been the site where posthumanism is being actively articulated for the humanities. ³⁹

274

User Layer

To map that multipolar field, an idea of inverse prostheticization is necessary to pinpoint symmetrical reversals of the human/animal distinction, where the human is now ridden and no longer the rider. The physiognomic, communicative, and emotional continuities and discontinuities between species are restaged by mutual identification and invention, exemplified by robotics laboratories that draw from animalian anatomies to solve for locomotion and adaptation: robot dogs, robot fish, robot birds, robot insects, and so on. For human purposes, the graduation from biological to synthetic animal machines is not difficult to understand — there is no design without some biomimesis — but what those animals might make of their robotic cousins is another matter. Surely perhaps some of the attraction and wariness that humans feel for humanoid robots might be played out for them as well, including fascination, indifference, and hostility. But instead of transposing animal machines into robots, how might animals use robots? What does it mean to install an animal body into the User position of a complex robot, especially an animal with whom anthropomorphic identification is all but impossible? Back in 2004, Garnet Hertz designed a table-sized mobile robot, navigable through a rollerball interface, and introduced a Madagascar hissing cockroach into the pilot ’ s nest. ⁴⁰ When cooperative, the cockroach did what cockroaches do and scurried about on the rollerball, which in turn caused the entire robot to scurry around the room like a giant cyborg cockroach, because that is exactly what it was. The effect was startling and uncanny, and it had everything to do with the degraded station of the cockroach. The creature occupies some liminal space between animal and matter, living dirt, filth that is but should not be in motion. If dirt is “ matter out of place, ” watching the living dirt in the driver ’ s seat of complex technology is upsetting, a sensation that runs up and down the spine of autonomic prejudice. ⁴¹ This is so, first because the unequivocal User in this demonstration is a bug, that all-purpose figure of primordial mindlessness and repetitious swarming aggregation, and second, because the robot prosthesis amplifies the scale of the bug ’ s “ intentions, ” transforming it into a grotesquely outsized coparticipant in our immediate habitat. The observer is forced to identify with the cockroach as a fellow User and to do so despite the absolute minimum of Cartesian credentials that it may possess. Cybernetics, of course, refers to the piloting of systems, and Garnet ’ s roaches, now learning to fly, present a precise example of subjectivization through the apparatus because of, not despite, their dubiousness.

User Layer

275

But the shifting of relations between humans and animals, mutually co-constituted by shared interfaces, needn ’ t be limited to placing animals in isolated control over more and more complex prosthetic technologies, letting them find their own purposes in The Stack, as wonderful as that design brief surely is. The positioning of human and animal as co-Users of interfacial platforms also demonstrates the scope of communication and identification that is possible when both occupy symmetrical locations within the platform. OOZ is a series of projects led by Natalie Jeremijenko, that seeks to expand from the restrictive and hierarchical forms of human/animal encounter available in zoos, where animals are trapped in place for observation but really where communication between species is at best serendipitous and at worst systematically suppressed. The project description explains that “ unlike the traditional zoo [OOZ] is a place where the animals remain by choice, a zoo without cages. Like a traditional zoo, it is a series of sites where animals and humans interact. However, the interactions at an OOZ site differ from that of a zoo. OOZ is interactive in that it provides human a set of actions, the animals provide reactions and these couplets add up to a collective pool of observations and encounters. The human/animal interface has two components: (1) an architecture of reciprocity, i.e. any action you can direct at the animal, they can direct at you and (2) an information architecture of collective observation and interpretation. ” The basic proposition is then extended in several instantiations featuring variously robotic geese, human-goose language translation media, bird art, amazing birdhouses, amphibious architecture, restaurants for fish and humans to share a common meal, stages for human/bat encounter in dense urban settings, and so on. The project employs of a kind of expanded interaction design, and its built-in structuring of the User position, as a way to leverage a rotation in pervasive and unnecessary asymmetries between human and animal actors in general: animals are no longer prosthetic channels and metabolic reserves but collaborative co-Users. This is not only for the benefit of the animals but also as a way to allow curious humans an introduction to the fascinating lifeworlds of our neighbors and to experience more satisfying sorts of communication with them. Critical to this is the design of reciprocal interfaces in which both sides of the interfacial membrane have equally impactful (if not always identical) means to affect the other. Here, that multipolar field for which the User subject position can be inhabited by a plurality of individual and aggregated agents draws not on the prosthetic radiation of control, in either direction, but rather on leveling the species-agnostic User position toward the open invention of alternative common interfaces. ⁴² The impetus for Agamben ’ s plea that the essential accomplishment of the subjectivizing apparatuses is to “ nullify animalistic behaviors ” is also put to rest.

276

User Layer

1. AI User

It may be that the deprioritization of the human User-subject within The Stack as a whole is made that much less frightening by interfaces designed to invite anthropomorphic empathy and identification. As of this writing, Siri for iOS is one iconic version of this new form of interface. Its significance is not only that it is voice-activated instead of mouse-click controlled, or that it responds to User inputs in synthetic spoken language, but that it is designed as a nascent form of an artificial human personality. We are invited not only to interact with iOS, and through the operating system with the semantic web (or at least the parts of the web that Siri knows how to search and process), but also to interact with Siri herself. The development of the User is also driven by the ability of artificial intelligences, simple ones and complex ones, to enter the User position and engage The Stack directly. They do this on the direct behalf of a human User, and they do this with or without human supervision, such that the human is in a way Siri ’ s co-User as much as the other way around. ⁴³ So then, who is Siri and who (or what) would the wider population of similarly personified interfaces/bot agents become? As the base model of the ideal assistant, her personality is specifically defined in relation to what the range of options might be, but her supposed neutrality as a universal assistant is highly contestable. She is one individuated person, not a plurality or multitude or composite. She is by default setting female in gender. She possesses a kind of patient and understated omniscience. Though a bit hard of hearing, she is extraordinarily reliable, and by the same turn, she is reliably subordinate. When asked strange things, she may demonstrate a dry sense of humor. She is also invisible, possessing no face, like Charlie on Charlie ’ s Angels. She is always there, but also apparently spending her time elsewhere when not actually here. All of these specific qualities need not define the personified agent as they do.

In 2012 (before Spike Jonze ’ s film Her), a group of my undergraduate students who wished to explore other alternatives began developing Siri-like assistant apps based on alternative, often fractured personalities. Departing from the supposed neutrality of Siri, these artificial borderline cases would engage in less generic sorts of User-Stack relationships, not all of them healthy or immediately recognizable. Some of their assistants were moody and unreliable. They may give answers based on their jealous analysis of your recent wall posts and try to trick you into paranoid suspicions about your friends. In time, similar alternative personality assistant Apps may demand absolute fidelity and suffer virtual death when it is revoked, like a passive-aggressive Tomagotchi. At this same time, it may merge with other Apps to perform spontaneous amazing feats on your behalf. Their interest went beyond persuasive interaction design gone sideways and was based more on what kinds of complex cognition and interaction are sanitized out of synthetic User agency and how weaving those wooly leftovers back in could point in more interesting directions. The results differed from Siri in two key ways.

User Layer

277

The suite of assistants they designed were not meant for one imaginary User for whom Siri ’ s personality is “ universally ” appropriate. Instead they developed idiosyncratic and culturally specific agent personalities. Their menagerie included agents for specific ethnic experiences, familial niches, curatorial perspectives, and User predicaments. Each would be finely tuned to a certain User but irrelevant to most others. Second, the agents were in different ways unreliable. Their personalities were defined by contradictions and mixed motives, much like human personalities. They may be manipulative or insecure, prejudiced or forgetful. The design issue here is not that broken tools are better, but that in designing interfaces that resemble a familiar mind, it may well be that vulnerable and limited artificial personalities invite and support deeper identification and loyalty from human Users than generic and predictable ones . ⁴⁴ The fifth edition of Diagnostic and Statistical Manual of Mental Disorders may prove essential reading for programmers of the next decades full of Siris. ⁴⁵ If Siri is just the tip of a AI interface iceberg, then to presume that “ she ” should be so defined by C-suite master-slave small talk is already to presume too much. The full spectrum of weirdness, based on actors passing information into and out of one another, is a richly reticulated space, and approaching it as such makes those clich é d relationships moot. We need to ensure that AI agents quickly evolve beyond the current status of sycophantic insects, because seeing and handling the world through that menial lens makes us, in turn, even more senseless. Something else needs to sideline the vestigial, embarrassing anthropocentrism that brings the false flattery of passive (or passive-aggressive) assistants. In fact, the stage management of the human User as somehow commanding the work of the App is already, we may conclude, an unnecessary alibi protecting the essential opposite effect (that the mammal User is only a provisional mechanism for dragging gigaflop tracking devices through the avenues of cities and for remonetizing these routes as the spatial career of algorithmic capital and its successors).

Put differently, bots are Users already, and we are already well past the point in which more than 50 percent of Internet traffic is generated by nonhuman users. ⁴⁶ As far as The Stack may see it, humans are a privileged User, but a shrinking minority, even as the next billions of them come online over the next decade. Similarly, the role for HFT algorithms is indeterminable, even though they do not speak at us they do speak for us, and we can assume that if similar forms of weaponized mathematics become more normative, then the plural and partial nature of any individuated human User subject interest and position may be that much harder to keep straight. In the ongoing technologicization of intelligence, we see cute slippages of position between humans and machines. For example, the original mechanical turk in the eighteenth century was a chess-playing machine, apparently an automaton exhibiting human intelligence, but in fact operated from afar by a human User. Today Amazon Mechanical Turk restages this arrangement, not just for chess but for any menial task the User can devise. Behind the browser are at least half a million “ workers ” who complete piecemeal tasks for micropayment. ⁴⁷ We see not AIs appearing as if they were human, but humans appearing as if they were AIs. As it ’ s been since Karel Čapek ’ s “ universal robots ” introduced on the stage in 1921, the robot not only mimics the human, but provides a portrait of the human as an object viewable to itself from the outside, and with the human negotiates an ongoing dance of reciprocal idealization. And in itself, this may signal less a mechanistic subsumption than a wider berth for incentive mutual identification between co-Users.

278

User Layer

1. Machine User

The machine User comes in many different guises and in a variety that is difficult to summarize and characterize. It can refer to almost any kind of automation or machinic process, if drawn broadly enough, or to very specific instances in which an apparatus is designed to operate in exact relation to other apparatuses, as if it were something like Agamben ’ s “ living being, ” which becomes a User subject through this delicate interpolation. For the latter, robotics will play an increasingly important role, and not just as a subgenre of biomimicry. Their expanded field would be populated by subjects and objects, of varying degrees of anthropomorphism, that animate the User layer of The Stack by internalizing flows from the other layers and sending them back down in columns of intentional action. These include machine Users, individuated as discrete addressable entities within interfacial regimes and more diffuse aggregate Users interwoven into background infrastructure, and various combinations of both. Some may be tightly bound to remote human control, such as the da Vinci robotic surgery system, and others may be programmed to operate and locomote with relative autonomy, such as sensor arrays or mobile surveillance robots. ⁴⁸ The most complex of these blend in such a way so as to displace what has been a well-established form of humanmachine-infrastructure interaction, one so well established that cities are designed to accommodate it. That blend inserts networked machinic control at different points, including computational augmentation of human agency over the system, or over the machine in which the User is installed, and of the infrastructural landscape in which those machines swarm together. Among these scenarios, the integrated design of driverless car navigation interfaces, computationally intensive and environmentally aware groups of cars and street systems that can stage the network effects of hundreds of thousands such speeding robots at once, would have a dramatic impact on everyday urbanism. The “ automobile ” is a mobile Cloud apparatus inside of which Users navigate the City layer according to augmented scenery, interfacial overlays mapping addressable haecceities of interest, and suckling chemical energies from, one hopes, a sustainable and equitable grid. Inside the fast box, the world doesn ’ t only pass by as interstitial nonplaces between departure and investigation; the kinesthetic momentum of a world passing becomes homelike, the default state against which stationary location is measured. The human User-passenger, drifting through an architecture of displacement and suspended proprioception, is symbiotically reliant on the apparatus, itself a machinic User of the wider infrastructural systems around them to provide not just transportation but an artificial somatosensorium.

User Layer

279

This massive mutual prostheticization (a form of nested parasitism, even) takes time to arrive, and the mediatization of the automotive experience did not arrive recently. Many years ago, I jotted down these notes while stuck in a particularly slow traffic jam:

Sitting in traffic on a Los Angeles freeway, I am reminded of Joan Didion ’ s revelation that this trap is the most authentic Angeleno social experience. We are not going to any place, all lined up behind our windshields, we are all already there. ⁴⁹ Today, bumper to bumper, we are now all also talking to ourselves, taking meetings, texting, emailing, Googling, checking on this and that, editing essays on our phones. This is the home and office. We do not always need to arrive, because we are already there: if this was your home, you would live here by now. … The freeways as boulevards of “ fast parking ” are a grid that segments and enables an inertial sort of mobility. Ensconced in our furtive provisional networks, the car is no longer the primary technology of mobility, even in LA. The car has gone the way of the building: too slow. … By the time Reyner Banham arrived, the car had eaten LA, but now the smartphone is eating the car. ⁵⁰ The “ ‘ mobile ” first appeared in 1970 ’ s as a “ car phone ” but now the terms are reversed. In The Transformers movie, the alien robot became a car and the phone became a robot. Here now, the car becomes a phone (as the car, in turn, becomes a robot). … As the economy of mobility slides from mechanical to informational the car is augmented by hands-free telephony, Bluetooth networks, in-dash GPS navigation systems offering visualised or spoken directions, audio jacks, big screens counting down the drops of fuel while talking to you in weird accents, and emergency concierge communications by builtin satellite intercom. The handset does all this too, steering us in different directions by maps, recommendations, search results, tags and so forth. The phone and car find ways to subcontract each other’s functions, one to the other and back again. … In ways that would have made no sense in Banham ’ s Los Angeles, the car ’ s interfaciality is an important criterion of performance as a personal vehicle. The most visible change to the car in recent years is in the display electronics within its primary interface, the dashboard/cockpit. If the grille on the outside used to be the car’s face — its look — now the face comes with the voice and is seen (worn?) now from the inside, in the driver ’ s seat. That most brandable experience is not the profile of the car seen passing through the world, but rather how the world looks gazing out through the lens of the car. … The car+phone hybrid is a mass medium, a mobile cinema + micro-urbanism for one driver/user, a habitat organised and narrated by data networks. But in the end, they are not equal; the handset

In conceptualizing how Stack transportation might evolve, it ’ s tempting to suppose that the aggregation of systems into central platforms at the Cloud level would be replicated at the City layer such that atomized vehicles (such as cars) would be agglomerated into larger urban, regional, and continental megavehicles (such as high-speed rail). The economies of scale that are possible by regularizing and extending itineraries might parallel those of regularizing and monetizing other kinds of social interaction online. Elon Musk ’ s perhaps real and perhaps speculative Hypertube project, which would send humans whooshing up and down California inside what is essentially a giant pneumatic tube, is exemplary, as is the tendency for cities to strategize economic growth through the enhancement of their airports (now “ aerotropoli ” ) ensuring their inhabitants easy access to the rest of the global urban grid. ⁵² Implementing such systems is obviously very expensive in both time and treasure. In addition, these existing car-based systems would also be consumed by The Stack and transformed into Cloudbased platforms that could provide many of the same advantages, and more besides, least in theory. Currently several at-least-partially autonomous unmanned ground vehicles are in regular operation; driverless mining trucks, for example, move millions of tons of dirt without a human behind the wheel, and driverless tractors plow fields controlled remotely by V2V (vehicle-to-vehicle) networks. ⁵³ Within cities, the implementation of automotive automation has proven more difficult, as evidenced by the retreat from initial designs for personal rapid transit transport in Masdar City, Abu Dhabi (as well as the ill-fated Aramis project in Paris, as chronicled by Bruno Latour — planned and re-planned, but never built). ⁵⁴ The best-known autonomous car project, as yet to launch services commercially, is of course that spearheaded by Google. It is not so uncommon to see Google ’ s prototype vehicles, decked out with telltale gear on the hood, pacing along particular streets in California. Much of the earlier work at Google on these technologies was done by Sebastian Thrun, who famously announced that a full-scale automobile automation system would save “ one million human lives per year ” by preventing accidents. ⁵⁵ As said, Google ’ s interest in mass transportation goes back further (at least Page ’ s Ann Arbor monorail.) ⁵⁶ It seems quite natural that someone with an interest in the engineering of global platforms would either begin or end up with transportation, but whether one sees the Internet as a “ superhighway ” or the hundreds of millions of cars in motion at any one time as very large “ data packets ” makes all the difference in how the issue is problematized and formulated. ⁵⁷

280

User Layer

Just as your mobile is not really so much a “ phone ” anymore (it too is a machine User), the Google mobile apparatus should be understood less as a car than, as said, as a car-shaped end device within a larger Cloud platform. It is a robotic device that a person gets inside and which carries her around. From the perspective of The Stack, it is also a composite User, built of augmented human perception as well as the individuated profiles of the human “ passengers, ” the manifold addressable systems of the rolling device, as well as its changing inputs, states, and performances as it travels around from site to site. The same human profile may enter and exit this shell, just as the car device may incorporate many different human User profiles in the course of the day. But when in motion and hurtling next to other cars, the encapsulated actor may temporarily be seen as one User, for whom the whole system must optimize a uniquely ideal path. The real impact of autonomous vehicular Cloud devices does not appear until their movements can be coordinated not User-by- User but as components in a streaming swarm. ⁵⁸ In theory, when synchronized by shared situational awareness and communication, whole herds of them may speed within inches of one another without worry of imminent collision. Should one brake suddenly a mile ahead, those behind would brake instantaneously instead of hurtling into the cascading pile-up. Cars going long distances could be clustered into long trains and fast-tracked away from local traffic. In this, many of the goals of PRT (personal rapid transit) systems (such as Aramis) might be realized by upgrading existing systems in this way. The systemic integration of vehicles would also have an impact on the economics of ownership, financing, and insurance that make car ownership more onerous today by shifting driving toward service-based platforms. Users may subscribe to a fleet of driverless cars available to them at a moment ’ s notice, and so can forgo the sunk costs of purchasing a machine they might use only a few minutes a day. For carbon-intensive products such as cars, a very large-scale shift toward using but not owning would at once save resources and offer greater end User convenience, not to mention the benefit to urban planners no longer having to waste valuable land on parking lots to store cars that will sit idle. ⁵⁹ The mass medium of swirling textures of robotic computational exoskeletons would also suggest innovations in the phenotypical outer surfaces of the devices themselves. They can in principle touch, connect, and intersect with one another, switching from singular to plural according to circumstances, and so the rather rigid shell and chassis form we know today could give way to other flexible morphologies. At the very least, as we are shuttled here and there in the vast multitudes of such machines, how human Users are physically positioned and what we spend our time doing will certainly not be the same as it is now. ⁶⁰ As discussed in the Interfaces chapter, as the “ car ” becomes a Cloud platform, it becomes available to an Apps economy, and to the extent that the Google Car is just a very large Android device with a very large, next generation Google Glass display, there is much for designers to work with.

User Layer

281

At the same time, such a system would bring potential problems of the same order of magnitude as those it alleviates. The software and sovereignty questions don ’ t abide easy answers. First, the legal identity of this composite User is not immediately clear. Several states have already passed legislation indicating that autonomous vehicles are legal to operate on their roads, thereby establishing the baseline that such machines are at least not criminal. But considering the quantity, complexity, and sensitivity of the data generated by such technologies, all working in concert, as well as the expertise and infrastructure necessary to conduct the rhythms of the swarm safely and effectively, it ’ s not likely that any Department of Motor Vehicles is a likely candidate to govern a network of pilotless vehicles. Who is, then? Is the working presumption that over time, drivers will subscribe to personal mobility platforms that operate with sufficient autonomy from the agencies ill equipped to monitor anything of this sort? When would such a platform, linking cities and vehicles into one metamachine, require clearance from every single state legislature for intercontinental cargo traffic to flow? Surely the

282

User Layer

National Security Agency has something to say about all that geolocative data. In short, the roads themselves may be where some of the most contentious and important friction between emergent Cloud Polis and the state over the geography of jurisdiction will play out. As discussed in the Cloud chapter, the interweaving of multiple and incongruous sovereign claims often hinges on how emergent platforms problematize and repurpose existing platforms (such as the intercontinental highway network and its federal stewards), and by how existing platforms steer that emergence toward its own publics.

Moreover, the psychological anguish of relinquishing driver status would likely ensure whatever policies are initially put in place may be irrational and absurd. Today the populist backlash against smart meters installed in residences as end points of more efficiently managed energy networks is nothing compared to the resistance (both legitimate and delusional) that will meet the sunsetting of human-driven automobiles. In important ways, however, the moral high ground may be with the robots. Gary Marcus writes, “ Eventually (though not yet) automated vehicles will be able to drive better, and more safely than you can; no drinking, no distraction, better reflexes, and better awareness (via networking) of other vehicles. Within two or three decades the difference between automated driving and human driving will be so great you may not be legally allowed to drive your own car, and even if you are allowed, it would be immoral of you to drive, because the risk of you hurting yourself or another person will be far greater than if you allowed a machine to do the work. ” ⁶¹ Be that as it may, and despite the systemic benefits to humans, there will be teeth-gnashing animosity to the call for humans to abdicate so much of the User position to the apparatus. It ’ s not quite a Copernican trauma, but it will be taken as an insult. In relation to infrastructural systems, the position of the User is one of at least partial or provisional sovereignty (per the discussion in the City chapter of what informal territories, physical and virtual envelopes, provide to Users in motion across state geographies). The interweaving of human and nonhuman bodies and User subjectivities is a radical extension and amplification of practical sovereignties (i.e., extending the freedom to move and ensuring against accidental death). For many, however, it will be seen instead as an unholy and inacceptable relinquishing and diminishing of self.

That said, there is no doubt that a technological assemblage such as this, drawing on all layers of The Stack at once, perched on uncertain and ineffectual regulatory concepts and methods, installed unevenly through often dysfunctional urban fabrics, carrying our very persons in its clutches (and so on), will also innovate massive new genres of catastrophe (though with over a million people killed in cars every year, those catastrophes will have to pass a high bar to be worse than those they ’ ve displaced). They mirror the vulnerabilities of The Stack as a whole, vulnerabilities of centralization and standardization, of universalization and interconnectedness, of energy breaches and code failures. A recent demonstration showed just how easy (fun and scary) it is to commandeer today ’ s cars by hacking their electronic control units: turning off brakes, indicating a full tank when empty, indicating the wrong vehicle speed, initiating the self-parking system while driving at 60 mph, sounding the horn randomly. ⁶² The implications for what might be possible when similar Fluxus mayhem is orchestrated at the scale of an entire City boggles the mind. As Marcus puts it, “ These driverless cars will be safer, but when accidents do happen, they may be on the scale of airline disasters ” though he may be off by an order of magnitude or two, depending on how deep and pervasive the Cloud vehicle platform does or does not become.

User Layer

283