The Stack On Software and Sovereignty

1. Google

In 2011, in a bid against Apple, Microsoft, and others for a cache of Nortel ’ s patents, Google offered what at first seemed like weird, arbitrary sums. It was soon apparent that these were not random, but represented definite and significant mathematical strings — for example, Google bid \$3.14159 billion, or pi times a billion, for one bundle. It had bid numbers that were Brun ’ s constant and the Meissel-Mertens constant, which relate to prime numbers. One can interpret this as an oblique but emphatic symbolic statement by Google that the immutable and universal laws of mathematics, which are by their nature uninterested in the folly of political hierarchies and economic intrigue, will ultimately win out over the hysterical mere numbers of human counting. Google may or may not share this view, of course, but none of the other platforms discussed in this chapter come close to Google in its universal, even cosmic ambition. Not only does the company express a cosmopolitan mission (at least rhetorically), it does so as if the company itself were merely the conduit of a deeper force of quantitative reason. The company ’ s mission statement, “ to organize the world ’ s information and make it universally accessible and useful, ” is in its simplicity about as all-encompassing as you can get. When everything is information, then organizing the information means organizing everything. For example, Google recently purchased the smart thermostat company Nest (as part of a robotics buying spree), and the company still has its license to sell energy. In the longer term, a Google Energy may be a key player in the retail and wholesaling of renewables and the management of both consumer and municipalityfacing smart grids. The company sees the pairing of bits and electrons as part of its vocation in ways that others simply cannot: Google Energy, Google Glass, Google Ideas, Google Car, Google Robotics, Google Earth. Google Space. Google Time. Google AI. Google Grossraum. Google Sovereignty, Google World. For the Google platform model for the Cloud Polis, these are all based on a grand vision encompassing (at least) information cosmopolitanism, search, advertising, physicalized information, and global infrastructure.

Google (and now Alphabet) is a company founded on an algorithm. ⁶² The original PageRank algorithm was Larry Page ’ s attempt to organize the entire World Wide Web according to something like the peer citation models that quantify which academic papers are most influential and relevant. This computational meritocracy is in the service of a universalist mission to not only organize the world ’ s information but to “ make it accessible and useful. ” The most current status report on this mission as a geopolitical framework is the book The New Digital Age: Reshaping the Future of People, Nations and Business by Google executive chairman Eric Schmidt and former State Department official, now Google Ideas director, Jared Cohen. The book pledges allegiance to various contemporary truisms regarding the Internet ’ s ability to flatten and democratize political systems and to deliver market-friendly civil societies. It argues that open data flows puncture local superstitions and will eventually sink capricious authoritarian ideologies under the weight of the transparency and enlightenment that they usher in. Partially reminiscent of the Wilsonian doctrines that drew Schmitt ’ s disdain, and also of Clintonian versions of the end-of-history thesis, for Schmidt and Cohen, friction is equated with antiquated self-serving oppression whereas universal information liquidity is linked with geocultural progress and democracy. Their characterization of Google ’ s role as neutral platform is often more implied than directly outlined, and visionary claims for the digital future are made through and on behalf of the regular people whose lives will be enhanced by the Google platform ’ s evolution. Today, however, Google is also perceived as too linked to the US federal system, and both fairly and unfairly is held in suspicion throughout the world, especially in China and Russia, as a stalking horse for the US State and Defense departments ’ security apparatuses. ⁶³ Any real informational cosmopolitanism is far less likely to come to fruition in any way resembling Google ’ s stated aspirations as long as this is the case. In his slashing review of the book, Julian Assange characterizes it as “ an attempt by Google to position itself as America ’ s geopolitical visionary — the one company that can answer the question, ‘ Where should America go? ’ ” While his cynicism about the chrysalis Google World on tap and what it means for a company hoping to not “ be evil ” is shared by many, Assange ’ s laser focus on the US-centricity of Google overstates the particular. ⁶⁴ Put plainly, Assange ’ s single-mindedness about the United States as the epicenter of hegemony and therefore the key innovator of authoritarian models (an assertion he had no qualms making from own show on Putin ’ s RT television channel) is too idiosyncratic and limited a critique. It is informed by an absolutized (and Oedipalized) metapolitics that crusades for transparency, opacity, and/or privacy as the guarantors of safety and sovereignty for atomized rational actors, and too often also as the core or even sole requirements for human social organization. For Assange, Google equates with sinister schemes of a single state, and according to Schmidt and Cohen ’ s book, Google is but a useful and agnostic enabler of the democratic project. What neither Assange nor Google is able to say is that the more disruptive and even likely future is one in which platforms such as Google undermine the ability of traditional states to control the evolution of the new Cloud polities, irregular and unrecognizable, that spring to life and will in time turn on their state hosts.

134

Cloud Layer

For their part, Schmidt and Cohen also warn against “ Internet balkanization ” whereby authoritarian state actors (China, Russia, Iran, federations of Sunni states, Texas, and so on) build virtual borders around their citizens, keeping them trapped in a closed bubble of sanctioned concepts, and sheltered from the enlightening waters of the global Internet provided by open platforms, such as Google. Under this deglobalization, the cosmopolitan potential of the Google Cloud model would be subverted and inverted by local “ traditional states ” jealous of their citizens ’ wandering attention. They explain the dynamic of Cloud versus state in the gentlest possible terms for a Western audience. The negative scenarios drawn include alternative DNS (Domain

Cloud Layer

135

Name System) addressing systems allowing any state bloc to contain its own map of the online world, separate from others, and in principle to wipe other locations off their particular map. Such information enclaves, especially those implemented and enforced by traditional state regimes, are a hovering threat to atmospheric universalism, and so Google now also offers its own Google Public DNS, which allows Users behind the wall of a closed-off Internet to tunnel through anyway. ⁶⁵ As these arms races heat up, Schmidt and Cohen speculate about Internet refugees and asylum seekers moving across physical borders, without or without special visas, so that they might access Cloud systems unavailable from where they escaped. The corollary of this proliferation of walled gardens and online enclaves is what the authors call “ virtual states, ” represented by proto-states that hope to realize themselves as distinct communities, and they discuss the examples of formal online polities for a Kurdish state and an independent Chechnya. But the more interesting Cloud polities are not those that simply extrude national or ethnic identity online and reestablish borders there, or those that use online space to make irredentist claims to territorial sovereignty denied them on the physical ground, but those that develop and form without that geographic and historical legacy and for whom the Cloud’ s own geographic situation is the first basis of emergent sovereign imaginaries. Cloud “ citizenship ” is structured around the slippery semantics of User identification and addressing. While Cloud platforms might support, subvert, or sidestep the sovereignties of particular state systems, those same systems may be put to the task of arbitrating some aspect of the final maps of identity and the portions of political subjectivity that they might still contain. Competing platforms, which seek to leverage their core control of self-identity economics as the basis of their platform ’ s effective autonomy, might be less enthusiastic about that particular arrangement. At the same time, identity services are sure to continue to play a decisive role in differentiating privacy and publicity, cozily secure for those who can pay, but for those less fortunate life is lived out in the open, shedding data for whatever may come along to observe them.

Google ’ s geopolitical model depends on their core revenue channels, namely search and advertising. Search is conventionally understood as the ability to enter strings of human-understandable language into an index and then receive ranked links to “ web pages ” correlated to their relevance for Users like you. But search, as an epistemological technology, can and does extend to more esoteric realms, expanding “ who ” is searching, the semantics of “ a query, ” and the uncertain discreteness of “ what ” can be searched for in the first place. As computational systems absorb more and more industrial infrastructure at a finer grain of detail, imbuing tiny components with network addressability, nonhuman Users will make far more queries than those by regular humans, and by some measures, they already do. Search shifts from a person-looking-for-information business model toward a subsystem component-querying-external-systems mode of infrastructure. Search becomes about machines querying state conditions of other machines, and these can include Google hardware such as a car, which will search or sense their immediate surroundings and are in turn searched by those same surroundings. ⁶⁶ Perhaps there is no User who can search who cannot also be searched. As search includes many forms of Cloud-based person-to-machine, machineto-machine, and machine-to-person communication, it may also develop deeper and more semantically standardized grammars. Many semantic systems would allow a simple query to execute a script or small program that would return much more than a list of links; “ to search ” becomes more like programming the search engine and its index according to User wants and purposes. When nonhuman hardware executes semantically complex query programs, then search becomes a general platform for an autoprogrammatic physical Cloud. Finally, far beyond Web pages, the landscape of what can be addressed by a given search expands to include any “ thing ” that can be addressed with a discrete numerical signifier, whether it resides on a server in a data center or in the habitats of everyday life. That addressee could be a bit of datum, a physical thing, a mapped relationship between things in time, tangible or intangible (more on this in the Address chapter). The future of search involves all these innovations of the who, how, and what of searching and being searched, and Google ’ s current eminence in this space depends on its ability to outpace (or outacquire) competitors in these domains.

136

Cloud Layer

Google ’ s revenue does not come from the direct provision of retail Cloud services, such as search, but from the monetization of User attention paid to services in the course of engagement. Its Cloud model economics is based on the successful transformation of the web into a massive peer-to-peer advertising platform, and the transformation of advertising into a meshwork of computationally microtargeted points and clicks. By the pairing of supplied and demanded attention at the infinitesimal degrees of individual search results, page views, and keyword bidding, Google absorbs cognitive capital from the collective intellect of its User population, siphoning value centripetally and then pushing it back outward. ⁶⁷ Google ’ s vast suite of Cloud services costs no money for Users to use, but they are not exactly free; the company generates far more money in aggregate by optimizing attention and information that it costs them to provide those services, and in this regard Google ’ s business is exemplary of platform surplus value discussed in the second chapter, and of how the platform derives value proportional to its own use and adoption. Google ’ s essential techniques for this, besides the built-in network effect of its vast User population, are the proprietary algorithms that precisely pair bid-on search terms with optimized advertising and results. Every time anyone makes a search on Google and clicks on a particular result or advertisement, she is also retraining the platform ’ s algorithmic intelligence, a tiny bit each time, helping it to make incrementally more precise predictions of User intention, intuition, desire, and demand. Without minimizing the expertise of Google ’ s scientists and engineers to actually compose these complex systems, we can also say that Google ’ s search and advertising infrastructure draws on the work of its users to optimize which search results will best match the inquiries made by future Users and their interests. Every time we execute a search query, Google is not only working for us; we are also working for Google. The total value of Google ’ s costs to provide those services is always less than the total value of the intellectual and information capital that its User population provides to the platform. The difference between the former and the latter was a gross profit margin of 62.86% for 2015 as of mid-June, according to Google ’ s reporting. Whereas for all advertising economies, the audience is the real product resold by the minute, impression, or click, for mature platform economies such as Google ’ s, the User is the product and the worker cobuilding the platform (of which he or she or it is also a customer). This would not work if Users themselves did not receive real platform value from their participation in this particular platform as opposed to others. From the perspective of the individual User, the personal value of information received must be (or at least seem) greater than value sent, even if at the aggregate level of User populations, the inverse may be true. At once, Google ’ s Cloud economics draws on capitalized cognition, networked value production, and incremental value accumulation and reappropriation. ⁶⁸ In this sense, it is not foremost an apparatus of surveillance or control per se but a medium for the capture and transformation of living, thinking, and knowing into platform value. The latter is not reducible to the former. Its “ organization of the world ’ s information ” strongly depends on the incorporation of the work done for its global index by its Users, and the blending of that work into new aggregates, so as to better train that machine ’ s capacity to simulate, motivate, and mediate Cloud economies with thought-by-thought, gesture-by-gesture granularity. One wonders what the Soviet cyberneticians of the 1950s and 1960s, with their plodding ledgers and “ gosplans, ” would make of such a system if it were stripped of its patina of American hegemony.

Cloud Layer

137

“ Information ” is not only about the world; it is in the world, and so once Google ’ s platform archive has reached a threshold equilibrium, the company ’ s mission statement suggests that it should retool the world to make it generate more information, so that the world itself can be better organized. In some cases, making information physical is indistinguishable from making physical machines more computational; the impetus for hybridization can begin from either side. In 2013 – 2014, Google went on an extraordinary buying spree, acquiring many of the most innovative robotics companies, many working in areas not directly related to the company ’ s existing services and offerings. Google ’ s new zoo of machines represents a diverse range of physical activities that could be programmed and monitored by its platform; together these companies make robots that can run, jump, carry heavy things, fly, make things, take things apart, and manage the temperature of your house based on your daily routines. It is reported that Google ’ s robotics ventures will be one of the occupants of the Hangar One megastructure at Moffett Field, leased from NASA ’ s Ames Research Center in Mountain View, CA, but with the chief architect of the Android mobile operating system, Andy Rubin, initially in charge of the project, we surmise that its mission is perhaps more to link robots together on a common operating system than to engineer their bodies. Rubin has since left Google, replaced by James Kuffner, but the longer term strategy may be in place: organize and make accessible the robotic phylum. Our anthropocentric predispositions often lead us think of a robot in the singular, as a synthetic person that contains its program just like a human thinks that it contains a will and a personality. Platform-based robotics, however, suggest other avenues of evolution. Today, ROS (robot operating system) is one existing platform of tools, libraries, and frameworks for adding software-based functionality to robots in ways similar to how one would program a software-only application; however; it cannot match what something like a robust “ Android for robots ” could do as part of the Google Cloud Polis model. Robotic mechanisms at varying scales would not only execute programs, but could sense, correlate, and report data on their interactions with the world, feeding the platform ’ s appetite for information to search, organize, and optimize. Each mechanism would also be in essence another User of the Google platform, and like all other Users, it would search and be searched, make use of the information that is there, and generate more information for other Users. Furthermore, as a model User, it is plural; its own intelligence is networked with the platform and with the wider collective of robotic and human Users with whom it collaborates. Should any one mechanism encounter something unknown, it could access a live API (application programming interface) or another more efficient technique and register the intelligence necessary to make decisions and actions. Over time, as each encounters different things and modifies its programming to accommodate them, two robot mechanisms born of the same model begin to diverge and even speciate. (More on this in the concluding chapter.) In the meantime, instead of new forms and behaviors, as platform information becomes more physical, its initial accomplishments may to “ roboticize ” machines, processes, techniques, behaviors, and systems as we already find them. The results are not unlike skeuomorphic interface designs where digital icons are made to resemble everyday objects and so allegorize how human Users (and designers) understand their machine ’ s functions to work. It will take some time for platform robotics to invent new infrastructural systems that are unique to how its capacities can be designed instead of merely automating what already exists.

138

Cloud Layer

Toward this, the ambition of Google ’ s Cloud Polis model already extends into the design and deployment of global infrastructures, and this interest is not a recent addition to the company ’ s founding vision, but in many ways precedes it. When Larry Page was a student at the University of Michigan, he was fascinated with the idea of an Ann Arbor monorail that would make local transportation more efficient. Google ’ s interest in transportation-as-platform extends, as discussed, into the Google driverless car project, now a focus of Sergey Brin ’ s interest and attention (more on these in the User chapter). One corollary of governing the Cloud as a global infrastructure for the integration of multiple economies under the rubric of rationalized information is a reframing of existing infrastructures as local instances for the application of a new information systems engineering program. Some of these (e.g., mass transportation, energy distribution, economic identity, cartography, banking) developed during the industrial era as platform mechanisms for large-scale states that may have been, at that time, among the only actors with sufficient capacity to deploy and defend these acts of sovereign inscription at continental scale. For better or worse, platforms such as Google may be positioned to offer parallel innovation of these same inscriptions now as private Cloudbased services. In that the ongoing effective sovereignty of modern states was to a degree underwritten by their management of local sections of global infrastructure, the realization of viable alternatives to this guarantee, now as core functions of The Stack, further complicates the boundaries between platform as state and state as platform.

Cloud Layer

139

Perhaps the most fundamental of these, however, will prove to be energy and the Cloud layer ’ s governance of the Earth layer of The Stack. To return to the earlier figure of a remote sea-based data center, we can say that is not likely that Google ’ s research exploration of this solution is meant as an experiment with the liquid boundaries of Cloud jurisdiction. The company is not, I assume, looking at this particular project as a way to directly challenge state jurisdiction altogether. In fact, Google ’ s fortune is probably more suited not to secession from state platform space, but to strategic interweaving with states (plural) and to an incremental assumption of traditional functions of the state, as they themselves pivot to the Cloud. However, taken as a design fiction (albeit one under Google ’ s patent protection), the oceanic data center cannot but inspire ideas about new geographies. As indicated, Google ’ s direct interest in such a system is driven by the pursuit of energy efficiencies and how to scale a planetary Cloud without a minimum of carbon- and dollar-intensive electrons. Google already has a subsidiary, Google Energy, that allows it to directly buy and sell federally regulated wholesale electricity, allowing the company to purchase directly from independent producers bypassing local utilities, and then selling excess energy back to the grid. ⁶⁹ It is not difficult to imagine how a User-facing alternate version of Google Energy could become both major smart grid concern and provider of Google-branded or Googleproduced energy, municipal and personal metering services, grid platform software, or a business-to-business and business-to-customer customer/ User relationship management systems. In principle, that combined energy-data platform could make (some) international carbon mitigation agreements more enforceable precisely because states would not have to agree among each other as to how to monitor and measure the flows. Could Google (and corollaries in other parts of the world) just do this on its own and report it? Would state security hacking of energy-data routers and sensors only extend the old realpolitik into the realm of networked electrons? Still, as discussed in the Earth chapter, the membrane between an energy grid that powers the Cloud and the Cloud as a grid that mediates both bits and electrons at once is potentially porous at some key points. Google ’ s energy ventures may suggest a model for the Cloud Polis for which the provision of both of these at once is a compound service and system. ⁷⁰ The interface of two existing systems, an information network and an energy network, may combine to produce a third thing: an energy Cloud drawn by a mixture of urban and regional infrastructures, networked production and storage centers, and on-demand utility nodes. Just as the production and management of infrastructure cohered publics for industrial states, the same may be true for the organization of new publics in their image of the platforms that come to manage them. If so, Google might depoliticize energy in some ways and radically politicize it in others. Recall that the political guarantee of a secure energy source (food) was and is a building block of archaic political loyalty, and is continued by states that did and do provide electricity as a public utility. Regardless of any benefits, the rendering of human-usable electrons according to the logics of platform surplus value is also certain to be attached to at least a few calamities. Nevertheless, as the geopolitics of climate change – related energy production and consumption effects looms larger, the questions of energy provision, dissemination, transparency, monitoring, alliance, and allegiance will (as discussed in the Earth chapter) drive realignments of jurisdictional loyalties around common predicaments, whether we prefer them to do so or not. Going forward, there is no Earth layer without the Cloud layer, and vice versa. Because energy is so essential and its calculative rationalization so attractive, it may drive the determining variables — its planetary limits, its antagonistic territories, its reserve currencies, and its included and excluded populations of Users — not just for Google ’ s model of the Cloud Polis, but for the political geographic alignments of the entire Stack.

140

Cloud Layer

1. Future Cloud Polis and Platforms

The geopolitical design question for the Cloud layer of The Stack circulates around what forms different Cloud polities will take in the near future and how they will draw on alternative organizational logics in their emergence. To diagram this, it is tempting to extrapolate from existing models and to inflate and linearize their exponential development. Surely to varying degrees the models described above, as abstracted from Facebook, Apple, Amazon, and Google, will persist, normalize, expand, and recombine in some linear fashion. But in their recombinations and in the gaps and failures, unexpected and perhaps unlikable developments are certain to shift the terms by which such models mediate power. In the span of a generation, we may be living with an ecology of competing platforms for which the models that differentiate today ’ s players are mixed differently along with new unknown functions. Only for this reason, and not for making futurist predictions about companies or stocks, the models of the current mix are a good starting point from which to sketch what might arrange alternative Cloud polities. Each of the limited sovereignties that a platform enjoys over a specific domain can be recombined. For example, instead of today ’ s division of organizational acumen as drawn above, perhaps some new Cloud platform will be based on a combination of aspects drawn from Facebook, Amazon, Apple, and Google at once. Perhaps one is based on individuation and display, App ecology, object logistics, and global infrastructure. Perhaps another instead draws on visual communication, hardware and physical touch, third-party cloud services, and search, while a third is based on fabricated currencies, enclave aesthetics, retail virtualization and compression, and cloud-based robotics. To date, the current major platforms have evolved as expressions of specific but not inevitable strategic decisions. There is no reason to assume that the ecology of platforms must be as it is, and very different arrangements are not only possible but inevitable. Today ’ s key Cloud platforms might evolve in the direction of these prototypical mutations, while new platforms consolidate and bifurcate, others emerge from bits and pieces of incomplete smaller players, and still others remain resolutely partial and interstitial, carving unrecognized parcels from the overlaps. It would be worthwhile to formally outline full scenarios for each set of possible recombinations from existing platforms (several others — e.g., Walmart, Alibaba, General Electric, Samsung, Unilever, Siemens, Toyota — could be included here) formalizing the gaps, unexpected synergies and conflicts, and the different agendas that drive them to cohere and diverge. The resulting diagram of this menagerie of chimeric Cloud platforms will likely include some creatures that will in time appear for real and stalk the Earth.

Cloud Layer

141

If the basic ingredients from which those might be combined to develop global proto-state platforms seem dangerously inadequate then my point about the necessity of tuning our design attention here has been made. That designable future of Cloud Polis will not evolve on its own, as if determined by some autonomous space of calculative variables. It will appear in coevolution, sometimes complementary and sometimes violent, with other existing modes of political imagination and organization, including the Westphalian state and its inept adaption to global ecological governance. This is complicated by how the future of the state will also develop through its incorporation of the Cloud as its own field of vision, domain, and jurisdiction, in parallel with the Cloud further incorporating state functions. What, if anything, finally grounds these counterclaims? In one sense, the practical locus of governmental control is over the capacity to structure, police, and, especially, tax flows: flows of people, money, data, energy. Through the control of the interfaces of input and output, be that a port city, a great firewall, or a superjurisdictional seizure, the taxation of flows is both the means by which states execute the right to exact capital and the result of sovereignty. In turn, that sovereignty over flows of value is translated into the right to issue standard guaranteed currencies, through which those flows are measured and which symbolize the sovereignty that legitimizes that claim in the first place. Money is also a way to identify the content of flows through abstracted equivalencies and, through the Address layer, to enforce territorial claims over identification itself. But as Cloud platforms have perhaps more direct means to validate them, competing claims for unforceable identifications of taxable flows may be where critical conflicts for the coevolutions of Cloud governance also rest.

142

Cloud Layer

Today some states seek to reconcretize their authority over the otherwise untaxed flows of value through the Cloud and have turned to forums like the UN International Communications Union to redefine global governance of the Cloud accordingly. ⁷¹ Some simply want more global representation in standards and policy bodies dominated by the United States, other want to meter and tax Google for sending data over their lines, others to build tighter control systems into the apparatus making it easier for them to police troublemakers, and others even seek to shift the assignment of neutral infrastructural processes, such as the allocation of IP addresses, to individual nation-states for reasons that are unclear even to them. ⁷² Instead of resolving the contradictions posed by the sovereign Cloud platforms by putting everything back in the bottle of the state ’ s borders, some propositions would immediately exacerbate pockets of friction by fraying points of mutual incompatibility toward kaleidoscopic delirium and structural vacuum. A further decentralization of Internet governance is not a bad idea in itself, but basing governance instead on the privileges of local fiefdoms and seizures, and on making blunt old-school patriarchal authoritarianism more convenient, is in fact a very bad idea. The Stack suggests no teleology or necessary outcome, and the specter of Cloud feudalism becoming the normal twenty-first-century macroeconomics is a real possibility, but the state pushback against today ’ s version of planetary-scale computation should not be seen as extrinsic, or even contradictory, to the logics of Stack geopolitics. As already suggested, a political and cultural momentum toward generalized secession, not one-worldism, seems to drive what we might see as the nomos of the Cloud. As a response to the uncertainties of the Cloud, polities consolidate into enclaves, new and old. Some of these are based on gestures like a reinscription of original lines of ethnic division and self-location, and some are based on platform sovereignties provided or demanded by regular participation in and identification with particular Cloud services, and some are based on both at once. Libertarian dreams of atomic secession and Eurasianist dreams of atavistic empire are two sides of the same coin. ⁷³ Just as fundamentalism is a persistent symptom of globalization, sometimes conservative localism is a symptom of planetary systemization; as such, it succeeds less as countermeasure against hegemony than as a culture laying siege to itself.

At some point, these conflicts may resolve into stable, if also contradictory, geopolitical architectures. Cloud Polis models may recognize some consensus(es) for the superimposition of one sovereignty on the other, sharing the same spot on alternative maps. Or perhaps, even if without mutual recognition, the “ Google ” and “ China ” models of Cloud Polis, for example, may still coexist in some as-yet-undetermined modus vivendi, the terms of which change as quickly as unintended effects of technical maneuvers ripple back and forth through their planes of jurisdiction. But even so, we expect that the map that is supposed to mark the sites of disputation between the two would itself be disputed. The two may weigh the scale of a demilitarized zone very differently, and the drawing of a real line on the ground may or may not be registered as formal demarcation regardless of how much it might actually structure actual flows. Any functional monopoly over the crucial symbolic lexicon of geography structures the kind of economics that can play out on it, and the Google Maps Nicaragua – Costa Rica conflict is telling, if only as a geopolitical allegory, of a novel shift between states and cloud platforms in establishing functional authority over that geographic symbolization. But China is no Costa Rica, and as India, Japan, and every other neighbor can attest, it is not shy about redrawing its own version of territorial reality according to an ambitious geographic self-image. Google ’ s ultimate geographic strategy is less obvious and strategically articulated; indeed, it neither needs to be nor should be. Instead of claiming and occupying an exclusive sovereign territory, Google absorbs existing spaces into its purview by capturing and consolidating images of all territory at various scales, from street to satellite and back, and rendering them into the platform ’ s comprehensive interface of rationalized space. Territory absorbed by Google is not one loop among others on a zero-sum horizontal plane, but a universal layer beneath the particular claims of any one entity in a thickened geographical stack. Google ’ s mapping ambition therefore is foundational and inclusive of territory-in-total, and thereby it may be seen as interfering with an absolute last instance of sovereign identification that states might otherwise enjoy. Its descriptive reason extends beyond land masses, even deep into the Amazon river basin to oceanic underwater canyons, and beyond the planet, including Mars and Earth ’ s moon, and offering them up through its own versions of the Blue Marble Earth, now interfacial regimes for the Google Earth publics. Does the Cloud Polis model suggested by Google ’ s geographic reach recast political territory in a recognizable way, and if so, where are its binding limits? Are they established by the interfaces between Google and the state, or do they extend “ beyond the line ” ? On the one hand, the superterranean spaces of the Cloud layer are a geometric mismatch with political geography as conventionally understood, but they are also coterminous with a return and ascendance of new city-state networks, and so with the fissiparous unbinding of one-world architectures comes a general resorting of mobile subjects into new infrastructural cosmopolities, enclaves, and exclaves, including ancient nationalisms and futuristic alliances. ⁷⁴

Cloud Layer

143