The Stack On Software and Sovereignty

1. Platform Wars

For The Stack, Cloud Polis represents an extremely complex design problem, because while it may determine so much, it guarantees little about the predictability of outcomes. Renewal and innovation or totalitarian dystopia: neither is assured or foreclosed. While we can survey the landscape of Cloud platform empires as they exist today, we can ’ t presume that the current arrangement of stakeholders and their positions will continue without dramatic unforeseen disruptions to the status quo . It is also possible (even probable) that the most decisive and geopolitical Cloud formations are still some years off. While Google, the specific company headquartered in Mountain View, California, is its most obvious exemplar of today ’ s platforms, my argument concerns something far less immediate than the product and service road map of any single corporation. The figure of The Stack is that of a politico-geographic-technological framework that does not yet exist, and may very well never exist, but serves as a conceptual-technical structure to think with and against as we compose what does emerge. For this, what makes the Cloud layer of The Stack so important is that today each of the major USbased ventures — Facebook, Apple, Google, Amazon (plus Walmart, Oracle, IBM, FedEx, UPS, Microsoft, Goldman Sachs, Cloudera, Dropbox, and others) — embody and enact, each one differently, a prototype (a “ prototypology ” ) for our Cloud geopolitical futures. The architecture of their brands and their software platforms is not only representative of geopolitical interests; each one in its way is a geopolitical model. To be clear, the more relevant questions for us are less how they might extend US superjurisdiction than how each constitutes different and incomplete options for what emerges around it and in spite of it. They perform this role both as privately held corporations and also, as Umberto Eco pointed out years before (1994) in his prescient satirical essay, “ The Holy War: Mac vs. DOS, ” even as exemplars of alternative techno-theological programs. ⁴⁶ In sequence, we ’ ll examine the models posed by Facebook, Apple, Amazon, and perhaps the most significant for this stage of the argument, Google. ⁴⁷

124

Cloud Layer

1. Facebook

Facebook ’ s Cloud Polis is built directly on its Users’ personal lives and their interest in each other ’ s personal lives. Its reserve currency is what the theory of symbolic interactionism in sociology calls the “ presentation of self-identity. ” ⁴⁸ In the reconstruction of the social networks that link the social-psychological capital of hundreds of millions of people, Facebook represents a voluntary and highly limited simulation of human culture, and for this, it is a singular achievement. Whereas Wikipedia, for example, publicly automates topical consensus, Facebook captures that social capital and guards it from strangers, displaying it through limited interfacial prisms. In turn, it leverages that capture into metacapital to be sold to microtargeting advertisers. Facebook ’ s Cloud model is rendered in the cumulative captured graph of the lives, likes, and allegiances of its active users, who spend more time on Facebook than any other site, and who have amassed the largest single repository of mechanical images in human history on its servers. ⁴⁹ Gabriel Tarde ’ s nineteenth-century dream, for a sociology based on diagramming society bit by bit from every microencounter until emergent patterns come into view, is suggested, if not also partially realized, by social network platforms such as these. ⁵⁰ Just as for Google ’ s search algorithms, for Facebook ’ s social graph simulation, the index is the innovation. If the paradigmatic cultural energy of the nineteenth and twentieth centuries was the invention of new media technologies and the exploration of these as vehicles to drive avant-garde forms of expression, then twenty-first century invention focuses instead on the scanning, archiving, cataloging, sorting, visualization, cutting and pasting, sensing, and serving rationalization and capitalization of archived reservoirs of content. Further, “ new content, ” like the billions of images uploaded to

Cloud Layer

125

Facebook or its Instagram, are already archived, socialized, and disseminated in near real-time. For such services, the archive is the primary channel of communication; the index is the medium of the message.

As a model for Cloud Polis, Facebook ’ s social graph relies on at least four techniques: identity and display, network closure, visual communication, and fabricated currencies. Through the profile and the feed, Facebook gives standard form to acts of social display, filtering them through perpetual refinements and updates. Participation in this narrowed Cloud “ public ” means a consistent, performative fashioning, and the “ care of the self ” as a public artifact of its own shadows and connections. Peers like what you say, they share your offerings, and, in time, they subscribe to you, the individual character in your own reality mini-show. The fabrication of the self becomes the primary project of this platform, at least for the User, but in time, this general model could take many different forms, as subjectivity and agency are dispersed into nonlocal networks and assignments. For Facebook today, that means Tarde meets Sartre meets Pavlov: this “ hell of other people ” is drawn as a social prison of timelines, passionate positiontaking, political sentimentality, “ weird tricks ” for this and that, pseudoscience, teen and mid-life narcissism, and stalking, all paid for in microseconds of attention.

The Facebook model is a closed and largely opaque network. How exactly its Users network with one another is based on proprietary graph algorithms and subject only to periodic, largely useless protests and petitions. Spawned in the image of the Ivycovered walls of elite mating grounds, Facebook maintains the reality and illusion of a well-pruned curated lifestream (restricted from other random people, that is, not from Facebook), such that one ’ s social network is not spammed by usurping gatecrashers. In some ways, this closure mimics the enclaved and privatized Cloud itself, monetized mostly by a few early investors. Whenever pressed, it seems that Facebook explicitly and emphatically does not endorse the cosmopolitan vision of information infrastructures professed by Google. However, as default rentier of our social affiliations and wannabe concierge of their capitalization, this conservatism may stunt the platform ’ s long-term evolution. ⁵¹ For one, it pushes Users toward regulatory remedies for perceived mismanagement of what they take to be “ their ” content in a “ public ” forum. It also invites the legal preemption of many techniques for social graph capitalization that Facebook (or any other graph-based Cloud) might wish to pursue and leverage in the future toward further independence. The company ’ s leadership is unloved and popularly perceived as selfish, petulant, and uncommitted (as reflected in the company ’ s IPO problems, the David Fincher movie, a hundred Sean Parker jokes, a smirking Eduardo Saverin defecting to Singapore to avoid paying his taxes, and the contrarian public persona of early funder Peter Thiel). There are, however, many other examples of network closure that deliver very different effects (Apple ’ s own take on the walled garden is discussed below) not the least of which is the model provided by China ’ s portfolio of social media properties like Renren, Sina Weibo, Tencent/QQ, and others.

126

Cloud Layer

All of these are interested first in the extraction of some sort of surplus from a captured social domain over which they exercise a slightly resented monopoly, and each looks at the other — state-sanctioned company versus private holding — as an ambiguous variable in its own strategies. The Facebook model (including its Instagam property) is also of a visual Cloud. It boasts human history ’ s single most prolific consolidation of images. As a primary channel of User communication, here the image overwhelms text as the basic gesture of self-fashioning. Should the Facebook platform bend toward socialized video, then the future of the feed may turn into semi-random cinema, edited by one ’ s network affiliations, and circulated by peers as social currency: real-world Chatroulette as the future of our species ’ intersubjective symbolic thought. ⁵² From the Facebook platform ’ s perspective, the ability to glean information about Users from large-scale artificial analysis of image content also greatly amplifies the scale and scope of the social simulation project. ⁵³ The acquisition of Oculus, the virtual reality hardware company, may extend this into more immersive visual cultures, perhaps operatic platformscale multiplayer online games built from the personal media of Facebook ’ s User base (or perhaps just 4D Farmville hallucinations).

The Facebook model is also based on artificial currencies. The platform has spawned a secondary economy of twitch-and-reward game apps, which feed off the central graph like little symbiots. In the midst of the European financial crisis, Metahaven proposed only half-ironically with its Facestate project that perhaps Facebook credits could replace the euro as a common currency. In this inside-out microeconomics of social capital extraction, we are all World of Warcraft Gold Miners, spamming ourselves to get through the day. ⁵⁴ However, the final form of that currency is still undecided for Facebook (as it is for PayPal, Apple Pay, Google Wallet, bitcoin, and Walmart Banking, for example). How a functional “ social wallet ” might emerge from the consolidation and transparency of our linking and delinking, one to another, generating and paying debt with attention and reputation, is an important component of a truly interconnected Cloud Polis. None of the other digital currency projects is built on the core currency over which Facebook still has privileged position of access: the microeconomies of recognized social debt from which the value of money is primordially derived (at least for humans; HST algorithms are a different story). ⁵⁵ But to date, Facebook ’ s furtive and ill-conceived experiments at the monetization of that capital are based on strategies of rent more than mediation, such as reciprocal likes, selling post promotion, charging users to message each other, and so essentially taxing the graph ’ s own growth. This may be a doubtful recipe for the conversion of public obligation into private money and back again. The magical ontology of money requires a trust so trusting that it requires no deliberation, and while social graph-based platforms may be where new currencies will be sustained in the long run, Facebook may have soiled its own punch, and so perhaps we ’ ll see banks becoming social graph platforms before we see graph platforms becoming banks. Still Facebook is the most widely engaged social media site with well over a billion active users and so its potential for structuring human communication according to its own logics of platform sovereignty remains profound.

Cloud Layer

127

1. Apple

Apple has assumed a mantle from Disney for preeminence in mass-scale, closed-loop experience design. ⁵⁶ By comparison with the extractive micromanagement style of Facebook, Apple ’ s closed world is ruled with luxury carrots more than with behaviorist sticks. Inside, its public expects total design into which (or onto which) Users can invest individual desires for creative self-idealization. Apple ’ s origin story is now as deeply ingrained in the American myth of the prodigal hero-entrepreneur as that of Peter Parker bitten by a radioactive spider. While the company ’ s roots extend back into mid-1970s Northern California hippie hacker culture, in important ways the Apple weltanschauung was not crystallized until the airing of Lee Clow and Ridley Scott ’ s Superbowl TV ad in and of the year 1984. ⁵⁷ Here the driving theologic dichotomy of the brand is established, cleaving the line between Apple (individual, color, youth, cool, iconoclast) and IBM (mass, monochrome, old, awkward, hierarchical), a creed equally appealing to 1960s counterculture and its boomer aftermath, as it is to the John Wayne wing of the American Right. An LSD-eating Buddhist and his gentle programmer pal set in motion what would become Rush Limbaugh ’ s favorite company. ⁵⁸ The story of the company and its brand have provided privileged archetypes to postindustrial capitalism: a populist concept, rejection by the old guard, a near-death experience, the return of the True Idea, blockbuster appliances, an actual death and hagiographic reverence, the passing of the doctrine, and so on.

Apple has also done more than any other company to make the rhetorics (and sometimes reality) of design into a central problematic for digital culture, far beyond the relentless cognitive utilitarianism of a traditional human-computer interaction (HCI) seeking to annihilate all ambiguity. The company turned computation into a mass medium of self-articulation: a main ingredient for any socially and economically sustainable Cloud Polis. Beyond the Cloud (or computing even), Apple also exemplified the power of brand as a core competency for publicly traded companies. It turned brand into content, encouraging customers to speculate on the brand experience in the third person, not only as something they themselves experience but such that their performative discussions about the brand experience are a core aspect of the experience. For Apple, this is founded on the idea that Apple uniquely “ gets it, ” and if you get Apple, then you are perceptive enough to get it, too. This conspiracy of the illuminated first drew its fundamental distinction against corporate cybernetics in that 1984 ad, and then against Microsoft and the (supposedly) crass, artless boredom of corporate IT departments and the malignant malevolence of Bill Gates. But today the terms of Apple ’ s Manichean doctrine are less clear and less confident. Apple ’ s light is now cast in contrast with what darkness? Is it Apple versus Google or Samsung, or what? The space is too fragmented; others are doing newer, more colorful and imaginative things. The software experience is now best recommended for its seamlessness, tastefulness, and predictability. These are hardly the watchwords of revolution. To paraphrase and invert the referent of the 1984 ad, the Apple model for Cloud Polis may be a walled “ garden of pure ideology, ” but it is backed up by inscrutable attention to physical detail in the design and manufacture of its hardware lines that allow its User to touch a tactile cloud.

128

Cloud Layer

The Apple model of Cloud Polis is based on four primary emphases: hardware/physical touch, App ecology, cult brand/leader, and enclave aesthetics. The iOS, iTunes, App, and third-party content ecosystem is the basis of the company ’ s Cloud platform, but the Apple “ public ” coheres through mutual esteem for the premium hardware that Users carry with them in daily life. As Cloud Polis develops beyond the data-on-screens framework of today, it is an open question what the design brief for Cloud hardware will include, and so it may lead a company such as Apple or Google well beyond the domain of consumer electronics as we conventionally understand them and deeper into the personal effects of the world (as Apple ’ s Watch and Google ’ s Car and Glass projects foreshadow). In this, platform lock-in versus neutral interoperability becomes an even more existential problematic. Apple ’ s initial forays into massive User-centric Cloud services have a spotty track record (think MobileMe), while its audience-centric Cloud services (such as iTunes) bend whole industries toward them and generate fabulous profits. Still, it is at the level of the operating system that Apple ’ s model platform logic coheres, and it is through premium hardware that it is guaranteed. As usually credited to Jony Ive ’ s talent and Steve Jobs ’ s perfectionism, Apple ’ s physical objects ground the Cloud as something you can and want to touch and accompany you. This “ design ” adds dramatically to profit margin per device and underpins other channels of involvement and lock-in, pushing User experience of The Stack toward dictates of affect, flattening and cajoling the megastructure to “ just work. ” Beyond individual touch, the physicality and tactility of Apple ’ s platform are also available as architectural immersion in the global footprint of Apple Stores, where an ideal Apple culture is performed by teams of ideal Apple Users, the youngish, intelligent, helpful, at-ease store staff. In this architectural network of mall-based parishes, one does not merely shop for electronics; one is to be supported in the goal of creative self-realization. The expert experiential design of the object, as well as the inspirational native habitat from which it is adopted at point of sale, cannot be separated from the political coherency of the Apple model of Cloud Polis. It is renewed less by new services than by a dramaturgical cycle of tactile collective participation and self-representation. No future Cloud Polis can afford to overlook these accomplishments. However, despite its total design model, Apple ’ s platform logic is underwritten by its (now much-copied) innovation of Apps and the App Store. Instead of providing every possible software service, product, or function that could be anticipated, it instead administers a centralized market for mini-applications, connecting developers and Users indirectly and ensuring profit for the former and some degree of code quality for the latter. An outgrowth of a hardwareled strategy, Apple ’ s App ecology means letting developers assume the risk and reward for the long tail of software experiences. Even so, these must hew to the platform ’ s core interaction affect or face banishment. Whichever way the Cloud (particularly the mobile Cloud) might evolve, Apple is positioned as a preferred physical “ last millimeter ” of delivery (the feel of “ look and feel ” ), the exact physical point where User and Cloud touch. (There is more on Apps in the Interfaces chapter.)

Cloud Layer

129

Apple ’ s walled garden, as opposed to Facebook ’ s, is utopian in tone, a paradigmatic community unto its own: autoexceptional. Jobs spoke of Apple as “ an Idea, ” wholly unlike the consumer electronics and software companies for which his contempt was always easily forthcoming. The Apple Cloud Polis also rests, then, on the enclave aesthetics of that concept and how it extends into particular choices about the company ’ s relationship to the governance of terra firma. The reader is perhaps familiar with the tale of Jobs ’ s Mercedes SL55, which he leased every few months so that he would never have to affix license plates. Having to do so would have required him to publicly acknowledge, in this small way, that he was personally subject to the general jurisdiction of the state of California and to its administrative control over roads and drivers. This conflation of an ethic of exceptional design work with political exception and exemption is inherent in the Apple model for Cloud Polis and gives wind to the effervescent patriotism among its Users. The ecstasy of revelatory launch events, the hermeneutic interest in the company ’ s history and future product road maps, not to mention hagiographic reverence for founding figures, all mutually strengthen real and imagined bonds between the Apple public and the brand.

As the delamination and re-interweaving of Clouds and contemporary states point toward one possible future of proliferating and overlapping enclaves and exclaves, the Apple totality is a model for an elite sovereign format, walled off from the relative chaos of outside publics. As ever, utopias are closed systems — islands — but where “ Apple ” is finally located is not so clear. Whereas Disney ’ s original plan for EPCOT Center was a utopian community, a real Disney city realized with diminished ambition as Celebration, Florida, the town of Apple, North Carolina, is the home of one of the company ’ s most important data centers, but it is a very unlikely site for residential relocation. Apple ’ s comprehensive attention to the interiority of product experience is well suited to a future featuring nation-sized gated communities wherever they may encircle themselves. For this, the exceptional enclave and the camp work both ways. The dystopian megastructures on which experiential seamlessness depends, like Foxconn ’ s factory cities where Apple objects are assembled, are the necessary mirrored doubles of the Apple polis itself, each cocooned in opposing bubbles yet each interior to the other. While some are drawn into live-work-sleep factory cities, where even suicide is not allowed, others are tucked like happy frequent flyers into a benevolent, tastefully designed compound in the sky (more on this pairing in the City chapter). While the Apple model for Cloud Polis represents a real triumph of expert industrial and interaction design at the level of technological geopolitics, these can ’ t be separated from the cloud enclaves against which they lean.

130

Cloud Layer

1. Amazon

By comparison with Facebook and Apple, Amazon ’ s model for Cloud Polis emphasizes an agora of objects rather than of humans, logistical expertise over tactile affect and self-identity, and white label services instead of brand patriotism. By far the largest online retailer, it has made post-Fordist supply chain compression and retail distribution chain automation into a long-tail mass medium. Through its gargantuan website and its network of warehouses and distribution centers, the Amazon platform centralizes connections between producers and retail consumers at a massive scale. The sum of individual items listed in the cosmic Amazon.com inventory index is measured in the hundreds of millions and is kept in check by the company ’ s own proprietary object identifier, the Amazon Standard Identifier (more in the Address chapter below). The Amazon Cloud model is based not only on expertise in faster-than-military-grade point-to-point logistics and mega-retail, but also the linking of these two with algorithmically driven pricing, recommendation, profiling, and proprietary purchase functions like One-Click and Prime. This gives Amazon visibility into small-volume supply chains similar to Walmart ’ s into mass-volume merchandise. The results are price constriction, delivery time compression, and a general undermining of brick-and-mortar retail economics. Of those discussed here, Amazon ’ s model is the most demonstrative of how what was once called e-commerce has transformed how goods are made and consumed, and how spatial and temporal economies of production, habitation, and labor have been refixed accordingly. But the “ flat world ” initiated by these logistical integrations is far flatter for objects than it is for people, as Amazon has orchestrated an incredibly precise scattering of commodified molecules coming and going.

In his 1950 essay “ The Thing, ” Heidegger lamented what he saw as the “ going ” of the organic, situated thing gathered from natural ingredients and the “ coming ” of the technological object, all placeless, anonymous, and cosmopolitan (I do not share his view on this). Scanning the landscapes of shipping containers, we assume that renewed philosophical interest in the agency of inanimate objects — the globality of material cultures, object-oriented philosophical speculations, assemblages, and networks — cannot be coincidental. These are overly timely reactions to an apparent if also unnamed predominance of object flows and freedoms, as supported by planetary-scale computation. This interest variously valorizes or rejects that predominance and those flows, more or less uncomfortable with the anti-anthropocentrism that rootless objects suggest and, as discussed in more detail in the Interfaces chapter, the post-Internet globalization of inanimate and animate substances presents a challenge to disturbed everyday sense of place and agency. For many, logistical distance and nearness bring not systemic transparency but mystification. Objects somehow appear on the doorstep, their origins opaque to their consumers and Users. Perhaps, from some vast and distant reservoir of a Spinozan ursubstance, a pure clay of stuff, commodity form is given to each one of these iterations, then packed in containers with millions more just like it. When this mass arrives at the shipping port, that sequence of smaller components assembling into larger aggregate states is reversed. ⁵⁹ The packed shipping container is opened; pallets are removed, trucked, tracked, broken down into boxes; and then finally a single-User unit appears for delivery, a pebble peeled off this unseen boulder and delivered to your porch-become-retail interface.

Cloud Layer

131

When you click on the Buy button, you trigger this whole process again. One fewer unit on the shelf means one more is assembled elsewhere to replace it as “ choice ” activates the supply chain as surely as turning a key in a car makes it start. Amazon ’ s is the state-of-the-art model platform for this Cloud Polis of things. Amazon ’ s central demands for that polity ’ s prototypical future are object logistics, retail compression and virtualization, procedural individuation, as well as third-party hosting, the provision of wholesale Cloud hosting at global scale. Amazon ’ s eminent retail position leverages the just-intime global delivery networks born in the 1970s that enabled overnight links between the East and West coasts of the US through central hubs in the Southeast, namely, UPS and FedEx. ⁶⁰ Tracking the growth of planetary-scale computation to synchronize and accelerate flows of data and metadata, these package delivery networks have evolved from expensive supplementary services for high-priority clients into general platforms for the distribution of objects to and fro. Their efficiencies threaten federal postal systems, which offer universal service even to remote locations but comparatively slow delivery options (more on the postal ontology of the state in the Address chapter). Through megaoutlets like Amazon, these delivery platforms not only grease the flow of goods from warehouse to shelf, they virtualize the shelf and alter the relative advantages and disadvantages of local businesses. The velocity of object flow tracks to exponential increases in the computational capacity of individual processors, which are also the switching gateways of real packages coordinated by them in aggregated multiples. As each switch in the landscape of logistical infrastructures increases in its ability to route bits, it also increases its ability to route the real objects that those bits represent. Amazon (and others — e.g., Walmart, Carrefour, Alibaba) aggregate and centralize those routes, feeding on efficiencies of scale, the competitive supervision of these networks, and the leverage over smaller actors in the cycle that these both guarantee.

The Amazon model demonstrates (as does Google ’ s) how the microtargeting of content to individual Users based on previous search-and-click history allows unique terms of engagement with diverse subpublics without needing to put one User on display to the other. This technology of procedural individuation-without-identity is in contrast with Facebook ’ s, which defines the individuation of the User largely through the visual display of identity. For Amazon, the long tail is a model of objective tendencies, not subjectively performed gestures. The platform does not care about your name or who you really are deep down, but only in the likelihood that the next presentation of object X, Y, or Z will motivate your One-Clicking and the subsequent activation of supply chain cascades that ensue. For this, Amazon does not even necessarily have to provide the User-facing front end for Cloud services. Just as UPS and FedEx moved into high-margin logistics consulting businesses, Amazon Web Services is a major provider of large-scale Cloud hosting and e-commerce for third parties (including states and their intelligence agencies). In addition to finding customers for its suppliers, Amazon also rents the pick-axes for the Cloud rush. This superterranean platform-of-platforms (from servers to warehouse to inventory delivery) allows both small and large affiliate actors to engage multiple overlapping and even opposed publics on the same shared hosting infrastructures.

132

Cloud Layer

As Amazon further shrinks the remote order-delivery cycle from a network of nearby distribution hubs and subhubs toward the same-day and instantaneous downloadable delivery of items, the remaining functions of local retail become both more constricted and more open. ⁶¹ On the one hand, ever greater supply chain compression may further transform retail space into purely experiential demonstration theaters, designed to motivate later online purchase, or may push to reprogram them all as publicly accessible storage facilities: a Tesla showroom or Home Depot, and not much in between. In time, 3D printing and other digital fabrication technologies may accelerate and overpower even this process, as your Kindle evolves from an e-book reader to an e-goods fabricator key and license library. Perhaps the Amazon model will lead away from today ’ s warehouse archipelagos dotting rural landscapes connected by cargo trucks and toward a denser network of print stations where furniture, electronics, appliances, clothes, and other products are fabricated on demand for pickup or shuttled to nearby residences (by drones, maybe — and maybe not). Perhaps the same network would also reverse the supply chain, providing deep recycling and upcycling of used goods. A network of Amazon printing stations might not only regurgitate commodities, but also reabsorb used matter as well, slurping up that universal commodity clay, and closing the open trash loop of mass consumption by shifting Users away from the redundant ownership of individual items and toward lighter subscription and service economies. The impacts of this gray-goo macroeconomics are unknown unknowns: all goods made free through object torrents, implosion of the demand base, empty factories in China, shipping ports turned into Museums of Cement, centrally planned provisioning, nanocommunism, nanofeudalism? As ever, platforms at this scale are both cure and poison at once.

Cloud Layer

133