Figure 1International Business Machines Corporation (IBM). “Smarter Planet” advertisement campaign, 2008. Detail. From Wall Street Journal, 17 November 2008.

p. 107 On November 6, 2008, still in the immediate aftermath of the worldwide economic crisis initiated by the U.S. subprime mortgage market collapse, then chair of IBM Sam Palmisano delivered a speech at the Council on Foreign Relations in New York City. The council is one of the foremost think tanks in the United States, its membership comprising senior figures in government, the intelligence community (including the Central Intelligence Agency), business leaders, financiers, lawyers, and the media. Yet Palmisano was not there to discuss the fate of the global economy. Rather, he introduced his corporation’s vision of the future in a talk titled “A Smarter Planet.” In glowing terms, Palmisano laid out a vision of fiber-optic cables, high-bandwidth infrastructure, seamless supply-chain and logistical capacity, a clean environment, and eternal economic growth, all of which were to be the preconditions for a “smart” planet. IBM, he argued, would lead the globe to the next frontier, a network beyond social networks and mere Twitter chats. This future world would come into being through the integration of human beings and machines into a seamless “Internet of things” that would generate the data necessary for organizing production and labor, enhancing marketing, facilitating democracy and prosperity, and—perhaps most important—for enabling a mode of automated, and seemingly apolitical, decision- making that would guarantee the survival of the human species in the face of pressing environmental challenges. In Palmisano’s talk, “smartness” named the interweaving of dynamic, emergent computational networks with the goal of producing a more resilient human species—that is, a species able to absorb and survive environmental, economic, and security crises by means of perpetually optimizing and adapting technologies.¹

Palmisano’s speech was notable less for its content, which to a degree was an amalgamation of existing claims about increased bandwidth, complexity, and ecological salvation, than for the way in which its economic context and its planetary terminology made explicit a hitherto tacit political promise that has attended the rise of “smart” technologies. Though IBM had capitalized for decades on terms associated with intelligence and thought—its earlier trademarked corporate slogan was “Think”—smart was by 2008 an adjective attached to many kinds of computer-mediated technologies and places, including phones, houses, cars, classrooms, bombs, chips, and cities. Palmisano’s “smarter planet” tagline drew on aspects of these earlier invocations of smartness, and especially the notion that p. 108 smartness required an extended infrastructure that produced an environment able to automate many human processes and respond in real time to human choices. His speech also underscored that smartness demanded an ongoing penetration of computing into infrastructure to mediate daily perceptions of life. (Smartphones, for example, are part of a discourse in which the world is imagined as networked, interactive, and constantly accessible through technological interfaces, their touch screens enabled by an infrastructure of satellite networks, server farms, and cellular towers, among many other structures that facilitate the regular accessing of services, goods, and spatial location data.) But as Palmisano’s speech made clear, these infrastructures now demanded an “infrastructural imaginary”—an orienting telos about what smartness is and does. This imaginary redefined no less than the relationships among technology, human sense perception, and cognition. With this extension of smartness to both the planet and the mind, what had been a corporate tagline became a governing project able to individuate a citizen and produce a global polity. Figure 2International Business Machines Corporation (IBM). “Smarter Planet” advertisement campaign, 2008. Detail. From Wall Street Journal, 17 November 2008.

This new vision of smartness is inextricably tied to the language of crisis, whether a financial, ecological, or security event. But where others might see the growing precariousness of human populations as best countered by conscious planning and regulation, advocates of smartness instead see opportunities to decentralize agency and intelligence by distributing it among objects, networks, and life forms. They predict that environmentally extended smartness will take the place of deliberative planning, allowing resilience in a perpetually transforming world. Palmisano proposed “infus[ing] intelligence into decision making” itself.² What Palmisano presented in 2008 as the mandate of a single corporation is central to much contem- p. 109 porary design and engineering thinking more generally.

We call these promises about computation, complexity, integration, ecology, and crisis “the smartness mandate.” We use this phrase to mark the fact that the assumptions and goals of “smart” technologies are widely accepted in global polity discussions and that they have encouraged the creation of novel infrastructures that organize environmental policy, energy policy, supply chains, the distribution of food and medicine, finance, and security policies. The smartness mandate draws on multiple and intersecting discourses, including ecology, evolutionary biology, computer science, and economics. Binding and bridging these discourses are technologies, instruments, apparatuses, processes, and architectures. These experimental networks of responsive machines, computer mainframes, political bodies, sensing devices, and spatial zones lend durable and material form to smartness, often allowing for its expansion and innovation with relative autonomy from its designers and champions. Figure 3Apple Computer. “Think Different” advertisement campaign, 1998. Detail. From Time, 2 February 1998.

This essay illuminates some of the key ways in which the history and logic of the smartness mandate are dynamically embedded in the objects and operations of everyday life— particularly the everyday lives of those living in the wealthier Global North, but ideally, for the advocates of smartness, the lives of every inhabitant of the globe. This approach allows us to consider questions such as, What kinds of assumptions link the “predictive” product suggestions made to a global public by retailers such as Amazon or Netflix with the efforts of South Korean urban-planning firms and Indian economic policy makers to monitor and in real time adapt to the activities of their urban citizenry? What kinds of ambitions permit the migration of statistically based modeling techniques from relatively banal p. 110 consumer applications to regional and transnational strategies of governance? How do smart technologies that enable socially networked applications for smartphones—for example, the Evernote app for distributed multisite and multiuser note taking used by 200 million registered users located primarily in the United States, Europe, Latin America, and Asia—also cultivate new forms of global labor and governmentality, the unity of which resides in coordination via smart platforms rather than, for example, geography or class? Each of these examples relies upon the intermediation of networks and technologies that are designated as “smart,” yet the impetus for innovation and the agents of this smartness often remain obscure.

We see the brief history of smartness as a decisive moment in histories of reason and rationality. In their helpful account of “Cold War rationality,” Paul Erickson and his colleagues argue that in the years following World War II American science, politics, and industry witnessed “the expansion of the domain of rationality at the expense of … reason,” as machinic systems and algorithmic procedures displaced judgment and discretion as ideals of governing rationally.³ Yet at the dawn of the twenty-first century, Cold War rationality has given way to the tyranny of smartness, an eternally emergent program of real-time, short-term calculation that substitutes “demos” (i.e., provisional models) and simulations for those systems of artificial intelligence and professional expertise and calculation imagined by Cold War rationalists. In place of Cold War warring systems based on “rational” processes that could still fall under the control and surveillance of centralized authorities or states, the smartness mandate embraces the ideal of an infinite range of experimental existences, all based on real-time adaptive exchanges among users, environments, and machines. Neither reason nor rationality is understood as a necessary guides for these exchanges, for smartness is presented as a selfregulating process of “optimization” and “resilience” (terms that, as we note below, are themselves moving targets in a recursive system).

Where Cold War rationality was highly suspicious of innovation, the latter is part of the essence of smartness. In place of the self-stabilizing systems and homeostasis that were the orienting ideal of Cold War theorists, smartness assumes perpetual growth and unlimited turmoil. Destruction, crisis, and the absence of architectonic order or rationality are the conditions of possibility for smart growth and optimization. Equally important: whereas Cold War rationality emanated primarily from the conceptual publications of a handful of well-funded think tanks, which tended to understand national populations and everyday culture as masses that need to be guided, smartness pervades cell phones, delivery trucks, and healthcare systems and relies intrinsically on the interactions among, and the individual idiosyncrasies of, millions or even billions of individuals around the planet. Moreover, whereas Cold War rationality was dominated by the thought of the doppelgänger p. 111 rival (e.g., the United States vs. the Soviet Union; the East vs. the West), smartness is not limited to binaries.⁴ Rather, it understands threats as emerging from an environment, which, because it is always more complex than the systems it encompasses, can never be captured in the simple schemas of rivalry or game theory. This, in turn, allows smartness to take on an ecological dimension: the key crisis is no longer simply that emerging from rival political powers or nuclear disaster but is any unforeseeable event that might emerge from an always toocomplex environment.

If smartness is what follows Cold War understandings of reason and rationality, the smartness mandate is the political imperative that smartness be extended to all areas of life. In this sense, the smart mandate is what follows “the shock doctrine,” powerfully described by Naomi Klein and others.⁵ As Klein notes in her book of the same name, the shock doctrine was a set of neoliberal assumptions and techniques that taught policy makers in the 1970s to take advantage of crises to downsize government and deregulate in order to extend the “rationality” of the free market to as many areas of life as possible. The smart mandate, we suggest, is the current instantiation of a new technical logic with equally transformative effects on conceptions and practices of governance, markets, democracy, and even life itself. Yet where the shock doctrine imagined a cadre of experts and advisors deployed to various national polities to liberate markets and free up resources at moments of crisis, the smartness mandate both understands crisis as a normal human condition and extends itself by means of a field of plural agents—environments, machines, populations, data sets—that interact in a complex manner and without recourse to what was earlier understood as reason or intelligence. If the shock doctrine promoted the idea that systems had to be “fixed” so that natural economic relationships could express themselves, the smartness mandate deploys ideas of resilience and practices management without ideals of futurity or clear measures of “success” or “failure.” We describe this imperative to develop and instantiate smartness everywhere as a mandate in order to capture both its political implications—although smartness is presented by its advocates as politically agnostic, it is more accurately viewed as completely reconfiguring the realm of the political—and the premise that smartness is possible only by drawing upon the “collective intelligence” of large populations.

We seek to sketch the deep logic of smartness and its mandate in four sections, each focused on a different aspect. These sections take up the following questions: (1) Where does smartness happen; that is, what kind of space does smartness require? (2) What is the agent of smartness; that is, what, precisely, enacts or possesses smartness? (3) What is the key operation of smartness; that is, what does smartness do? (4) What is the purported result of smartness; that is, at what does it aim? Our answers to these four questions are the following: p. 112

1. The territory of smartness is the zone.
2. The (quasi-)agent of smartness is populations.
3. The key operation of smartness is optimization.
4. Smartness produces resilience.

Focusing on how the logics and practices of zones, populations, optimization, and resilience are coupled enables us to illuminate not just particular instantiations of smartness—for example, smart cities, grids, or phones—but smartness more generally, as well as its mandate (“every process must become smart!”).

Our analysis draws inspiration from Michel Foucault’s concepts of governmentality and biopolitics, Gilles Deleuze’s brief account of “the control society,” and critical work on immaterial labor. We describe smartness genealogically; that is, as a concept and set of practices that emerged from the coupling of logics and techniques from multiple fields (ecology, computer science, policy, etc.). We also link smartness to the central object of biopolitics—populations—and see smartness as bound up with the key goal of biopolitics: governmentality. And we emphasize the importance of a mode of control based on what Deleuze describes as open-ended modulation rather than the permanent molding of discipline. We also underscore the centrality of data drawn from the everyday activities of large numbers of people. Yet insofar as smartness positions the global environment as the fundamental orienting point for all governance—that is, as the realm of governance that demands all other problems be seen from the perspective of zones, populations, resilience, and optimization—the tools offered by existing concepts of biopolitics, the control society, and immaterial labor take us only part of the way in our account.⁶

Zones

Smartness has to happen somewhere. However, advocates of smartness generally imply or explicitly note that its space is not that of the national territory. Palmisano’s invocation of a smarter planet, for example, emphasizes the extraterritorial space that smartness requires: precisely because smartness aims in part at ecological salvation, its operations cannot be restricted to the limited laws, territory, or populations of a given national polity. So, too, designers of “smart homes” imagine a domestic space freed by intelligent networks from the physical constraints of the home, while the fitness app on a smartphone conditions the training of a single user’s body through iterative calculations correlated with thousands or millions of other users spread across multiple continents.⁷ These activities all occur in space, but the nation-state is neither their obvious nor necessary container, nor is the human body and its related psychological subject their primary focus, target, or even paradigm (e.g., smartness often employs entities such as “swarms” that are never intended to cohere in the manner of a rational or liberal subject). At the same time, though, smartness also depends on p. 113 complicated and often delicate infrastructures—fiber-optic cable networks and communications systems capable of accessing satellite data; server farms that must be maintained at precise temperatures; safe shipping routes—that are invariably located at least in part within national territories and are often subsidized by federal governments. Smartness thus also requires the support of legal systems and policing that protect and maintain these infrastructures, and most of the latter are provided by national states (even if only in the form of subcontracted private security services).⁸

This paradoxical relationship of smartness to national territories is best captured as a mutation of the contemporary form of space known as “zones.” Related to histories of urban planning and development, where zoning has long been an instrument in organizing space, contemporary zones have new properties married to the financial and logistical practices that underpin their global proliferation. In the past two decades, numerous urban historians and media theorists have redefined the zone in terms of its connection to computation, and described the zone as the dominant territorial configuration of the present. As architectural theorist Keller Easterling notes, the zone should be understood as a method of “extrastatecraft” intended to serve as a platform for the operation of a new “software” for governing human activity. Brett Nielsen and Ned Rossiter invoke the figure of the “logistical city” or zone to make the same point about governmentality and computation.⁹

Zones denote not the demise of the state but the production of new forms of territory, the ideal of which is a space of exception to national and often international law. A key example is the so-called free-trade zone. Free-trade zones are a growing phenomenon, stretching from Pudong District in Shanghai to the Cayman Islands, and even the business districts and port facilities of New York State, and are promoted as conduits for the smooth transfer of capital, labor, and technology globally (with smooth defined as a minimum of delay as national borders are crossed). Free-trade zones are in one sense discrete physical spaces, but they also require new networked infrastructures linked through the algorithms that underwrite geographic information systems (GIS) and global positioning systems (GPS) and computerized supply-chain management systems, as well as the standardization of container and shipping architecture and regulatory legal exceptions (to mention just some of the protocols that produce these spaces). Equally important, zones are understood as outside the legal structure of a national territory, even if they technically lie within its space.¹⁰

In using the term zone to describe the space of smartness, our point is not that smartness happens in places such as freetrade zones but that smartness aims to globalize the zonal logic, or mode, of space. This logic of geographic abstraction, detachment, and exemption is exemplified even in a mundane consumer item such as activity monitors—for example, the Fitbit— that link data about the physical activities of a user in one p. 114 jurisdiction with the data of users in other jurisdictions. This logic of abstraction is more fully exemplified by the emergence of so-called smart cities. An organizing principle of the smart city is that civic governance and public taxation will be driven, and perhaps replaced, by automated and ubiquitous data collection. This ideal of a “sensorial” city that serves as a conduit for data gathering and circulation is a primary fantasy enabling smart cities, grids, and networks. Consider, for example, a prototype “greenfield” (i.e., from scratch) smart-city development, such as Songdo in South Korea. This smart city is designed with a massive sensor infrastructure for collecting traffic, environmental, and closed-circuit television (CCTV) data and includes individual smart homes (apartments) with multiple monitors and touch screens for temperature control, entertainment, lighting, and cooking functions. The city’s developers also hope these living spaces will eventually monitor multiple health conditions through home testing. Implementing this business plan, however, will require either significant changes to, or exemptions from, South Korean laws about transferring health information outside of hospitals. Lobbying efforts for this juridical change have been promoted by Cisco Systems (a U.S.-based network infrastructure provider), the Incheon Free Economic Zone (the governing local authority), and Posco (a Korean chaebol involved in construction and steel refining), the three most dominant forces behind Songdo. Figure 4Songdo, South Korea, 2014. Photo: Orit Halpern.

What makes smart territories unique in a world of zonal territories is the specific mode by which smartness colonizes space through the management of time (and this mode also helps explain why smartness is so successful in promulgating itself globally). As demonstrated by former IBM chair Palmisano’s address to the Council on Foreign Relations, smartness is predicated on an imaginary of “crisis” that is to be managed through a massive increase in sensing devices, which in turn purportedly enable self-organization and constant self-modulating and p. 115 self-updating systems. Smart platforms link zones to crisis via two key operations: (1) a temporal operation, by means of which uncertainty about the future is managed through constant redescription of the present as a “version,” “demo,” or “prototype” of the future; and (2) an operation of self-organization through which earlier discourses about structures and the social are replaced by concerns about infrastructure, a focus on sensor systems, and a fetish for big data and analytics, which purportedly can direct “development” in the absence of clearcut ends or goals.

In this sense, the development of smart cities such as Songdo follows a logic of software development. Every present state of the smart city is understood as a demo or prototype of a future smart city. Every operation in the smart city is understood in terms of testing and updating. Engineers interviewed at the site openly spoke of it as an “experiment” and “test,” admitting that the system did not work but stressing that problems could be fixed in the next instantiation elsewhere in the world.¹¹ As a consequence, there is never a finished product but rather infinitely replicable yet always preliminary, never-to-be-completed versions of these cities around the globe.

This temporal operation is then linked to an ideal of selforganization. Smartness largely refers to computationally and digitally managed systems, from electrical grids to building management systems, that can learn and, in theory, adapt by analyzing data about themselves. Self-organization is thus linked to the operation of optimization. Systems correct themselves automatically by adjusting their own operations. This organization is imagined as being immanent to the physical and informational system at hand—that is, as optimized by computationally collected data rather than by “external” political or social actors. At the heart of the smartness mandate is thus a logic of immanence, by means of which sensor instrumentation adjoined to emerging and often automated methods for the analysis of large data sets allow a dynamic system to detect and direct its continued growth.¹²

One of the key, troubling consequences of demoing and selforganization as the two zonal operations of smartness is that the overarching concept of “crisis” begins to obscure differences among kinds of catastrophes. While every crisis event— for example, the 2008 subprime mortgage collapse or the Tohoku earthquake of 2011—is different, within the demo-logic that underwrites the production of smart and resilient cities these differences can be subsumed under the general concept of crisis and addressed through the same methods (the implications of which must never be fully engaged because we are always “demoing” or “testing” solutions, never actually solving the problem). Whether threatened by terrorism, subprime mortgages, energy shortages, or hurricanes, smartness always responds in essentially the same way. The demo is a form of temporal management that through its practices and discourses evacuates the historical and contextual specificity of individual catastrophes p. 116 and evades ever having to assess or represent the impact of these infrastructures, because no project is ever “finished.” This evacuation of differences, temporalities, and societal structures is what most concerns us in confronting the extraordinary rise of ubiquitous computing and high-tech infrastructures as solutions to political, social, environmental, and historical problems confronting urban design and planning, and as engines for producing new forms of territory and governance.

Populations

If zones are the places in which smartness takes place, populations are the agents—or, more accurately, the enabling medium— of smartness. Smartness is located neither in the source (producer) nor the destination (consumer) of a good such as a smartphone but is the outcome of the algorithmic manipulation of billions of traces left by thousands, millions, or even billions of individual users. Smartness requires these large populations, for they are the medium of the “partial perceptions” within which smartness emerges. Though these populations should be understood as fundamentally biopolitical in nature, it is more helpful first to recognize the extent to which smartness relies on an understanding of population drawn from twentieth-century biological sciences such as evolutionary biology and ecology.

Biologists and ecologists often use the term population to describe large collections of individuals with the following characteristics: (1) the individuals differ at least slightly from one another; (2) these differences allow some individuals to be more “successful” vis-à-vis their environment than other individuals; (3) a form of memory enables differences that are successful to appear again in subsequent generations; and, as a consequence, (4) the distribution of differences across the population tends to change over time.¹³ This emphasis on the importance of individual difference for long-term fitness thus distinguishes this use of the term population from more common political uses of the term to describe the individuals who live within a political territory.¹⁴

Smartness takes up a biologically oriented concept of population but repurposes it for nonbiological contexts. Smartness presumes that each individual is distinct not only biologically but in terms of, for example, habits, knowledge, consumer preferences, and that information about these individual differences can usefully be grouped together so that algorithms can locate subgroupings of this data that thrive or falter in the face of specific changes. Though the populations of data drawn from individuals may map onto traditional biological or political divisions, groupings and subgroupings more generally revolve around consumer preferences and are drawn from individuals in widely separated geographical regions and polities. (For example, Netflix’s populations of movie preferences are currently created from users distributed throughout 190 countries.)¹⁵ Moreover, though these data populations are (generally) drawn from human beings, they are best understood as distinct from p. 117 the human populations from which they emerge: these are simply data populations of, for example, preferences, reactions, or abilities. This is true even in the case of information drawn from human bodies located in the same physical space. In the case of the smart city, the information streaming from fitness trackers, smartphones, credit cards, and transport cards is generated by human bodies in close physical proximity to one another, but individual data populations are then agglomerated at different temporalities and scales, depending on the problem being considered (transportation routing, energy use, consumer preferences, etc.). These discrete data populations enable processes to be optimized (i.e., enable “fitness” to be determined), which in turn produces new populations of data and hence a new series of potentialities for what a population is and what potentials these populations can generate.

A key premise of smartness is that while each member of a population is unique, it is also “dumb”—that is, limited in its “perception”—and that smartness emerges as a property of the population only when these limited perspectives are linked via environment-like infrastructures. Returning to the example of the smartphone operating in a smart city, the phone becomes a mechanism for creating data populations that operate without the cognition or even direct command of the subject. (The smartphone, for example, automatically transmits its location and can also transmit other data about how it has been used.) If, in the biological domain, populations enable long-term species survival, then in the cultural domain populations enable smartness, provided the populations are networked together with smart infrastructures. Populations are the perceptual substrate that enables modulating interactions among agents within a system that sustains particular activities. The infrastructures ensure, for example, that “given enough eyeballs, all bugs are shallow” (Linus’s Law); that problems can be “crowdsourced”; and that such a thing as “collective intelligence” exists.¹⁶ The concept of population also allows us to understand better why the zone is the necessary kind of space enabling smartness, for there is often no reason that national borders would parse population differences (in abilities, interests, preferences, or biology) in any way that is meaningful for smartness.

This creation and analysis of data populations is biopolitical in the sense initially outlined by Foucault, but smartness is also a significant mutation in the operation of biopolitics. Foucault stresses that the concept of population was central to the emergence of biopolitics in the late eighteenth century, for it denoted a “collective body” that had its own internal dynamics (of births, deaths, illness, etc.) that were quasi-autonomous in the sense that they could not be commanded or completely prevented by legal structures but could nevertheless be subtly altered through biopolitical regulatory techniques and technologies (e.g., required inoculations; free-market mechanisms).¹⁷ On the one hand, smartness is biopolitical in this same sense, for the members of its populations—populations of movie watchers, p. 118 cell phone users, healthcare purchasers and users, and so on— are assumed to have their own internal dynamics and regularities, and the goal of gathering information about these dynamics is not to discipline individuals into specific behaviors but to find points of leverage within these regularities that can produce more subtle and widespread changes.

On the other hand, the biopolitical dimension of smartness cannot be understood as simply “more of the same,” for four reasons. First, and in keeping with Deleuze’s reflections on the control society, the institutions that gather data about populations are now more likely to be corporations than states.¹⁸ Second (and as a consequence of the first point), smartness’s data populations often concern not those clearly biological events on which Foucault focused, but variables such as attention, consumer choices, and transportation preferences. Third, though the data populations that are the medium of smartness are often drawn from populations of human beings, these data relate differently to individuals than in the case of Foucault’s more health-oriented examples. Data populations themselves often do not need to be (and cannot be) mapped directly back onto discrete human populations: one is often less interested in discrete events that happen infrequently along the individual biographies of a polity (e.g., smallpox infections) than in frequent events that may happen across widely dispersed groups of people (e.g., movie preferences). The analysis of these data populations is then used to create, via smart technologies, an individual and customized “information-environment” around each individual. The aim is not to discipline individuals, in Foucault’s sense, but to extend ever deeper and further the quasi-autonomous dynamics of populations. Fourth, in the case of systems such as high-speed financial trading and derivatives, as well as in the logistical management of automated supply chains, entire data populations are produced and acted on directly through entirely machine-to-machine data gathering, communication, analytics, and action.¹⁹ These new forms of automation and of producing populations mark transformations in both the scale and intensity of the interweaving of algorithmic calculation and life.

Optimization

Smartness emerges when zones link the increasingly fine-grained, quasi-autonomous dynamics of populations for the sake of optimization. This pursuit of “the best”—the fastest route between two points, the most reliable prediction of a product a consumer will like, the least expenditure of energy in a home, the lowest risk and highest return in a financial portfolio—is what justifies the term smartness. Contemporary optimization is a fundamentally quantitative but calculation-intensive operation; it is a matter of finding, given specified constraints, maxima or minima. Locating these limits in population data often requires millions or billions of algorithmic mathematical calculations—hence the role of computers (which run complex p. 119 algorithms at speeds that are effectively “real-time” for human beings), globally distributed sensors (which enable constant global updating of distributed information), and global communications networks (which connect those sensors with that computing power).

Though optimization has a history, including techniques of industrial production and sciences of efficiency and fatigue pioneered in the late-nineteenth and early twentieth centuries by Fredrick Winslow Taylor and Frank Gilbreth, its current instantiations radically differ from earlier ones.²⁰ The term optimization appears to have entered common usage in English only following World War II.²¹ Related to emerging techniques such as game-theoretical tools and computers, optimization is a particular form of efficiency measure. To optimize is to find the best relationship between minima and maxima performances of a system. Optimization is not a normative or absolute measure of performance but an internally referential and relative one; it thus mirrors the temporality of the test bed, the version, and the prototype endemic to “smart” cities and zones.

Optimization is the technique by which smartness promulgates the belief that everything—every kind of relationship among human beings, their technologies, and the environments in which they live—can and should be algorithmically managed. Shopping, dating, exercising, the practice of science, the distribution of resources for public schools, the fight against terrorism, the calculation of carbon offsets and credits: these processes can—and must!—be optimized. Optimization fever propels the demand for ever-more sensors—more sites of data collection, whether via mobile device apps, hospital clinic databases, or tracking of website clicks—so that optimization’s realm can perpetually be expanded and optimization itself further optimized. Smart optimization also demands the everincreasing evacuation of private interiority on the part of individuals, for such privacy is now often implicitly understood as an indefensible withholding of information that could be used for optimizing human relations.²²

Smart optimization also presumes a new, fundamentally practical epistemology, for smartness is not focused on determining absolutely correct (i.e., “true”) solutions to optimization problems. The development of calculus in the seventeenth century encouraged the hope that, if one could simply find an equation for a curve that described a system, it would then always be possible in principle to locate absolute, rather than simply local, maxima and minima for that system. However, the problems engaged by smartness—for example, travel mapping, healthcare outcomes, risk portfolios—often have so many variables and dimensions that completely solving them, even in principle, is impossible. As Dan Simon notes, even a problem as apparently simple as determining the most optimal route for a salesperson who needs to visit fifty cities would be impossible if one were to try to calculate all possible solutions. There are 49! (= 6.1 x 1062) possible solutions to this problem, which is p. 120

beyond the capability of contemporary computing: even if one had a trillion computers, each capable of calculating a trillion solutions per second, and these computers had been calculating since the universe began—a total computation time of 15 billion years—they would not yet have come close to calculating all possible routes.²³ In the face of the impossibility of determining absolute maxima or minima for these systems by so-called brute force (i.e., calculating and comparing all possible solutions), contemporary optimization instead involves finding good-enough solutions: maxima and minima that may or may not be absolute but are more likely than other solutions to be close to absolute maxima or minima. The optimizing engineer selects among different algorithmic methods that each produce, in different ways and with different results, good-enough solutions.

In the absence of any way to calculate absolute maxima and minima, the belief that smartness nevertheless locates “best” solutions is supported both technically and analogically. This belief is supported technically in that different optimization algorithms are run on “benchmark” problems—that is, problems that contain numerous local maxima and minima but for which the absolute maximum or minimum is known—to determine how well the algorithms perform on those types of problems.²⁴ If an algorithm runs well on a benchmark problem, then it is presumed to be more likely to run well on similar real-world problems. Figure 5Ackley benchmark function. Image: Gaortizg, Wikimedia Commons.

The belief that smartness finds the best solutions is also often supported by the claim that many contemporary optimization algorithms mimic natural processes, especially computational ideals of biological evolution.²⁵ The algorithm begins with the premise that natural biological evolution automatically solves optimization problems by means of natural populations. The algorithm then seeks to simulate that process by creating populations of candidate solutions, which are mixed with one another (elements of one candidate solution are combined with elements of other candidate solutions) and culled through successive generations to produce increasingly good solutions. David B. Fogel, a consultant for the informatics firm Natural Selection, Inc., which applies computational models to the streamlining of commercial activities, captures this sense of optimization as simply a continuation of nature’s work: “Natural evolution is a population-based optimization process. Simulating this process on a computer results in stochastic optimization techniques that can often outperform classical methods of optimization when applied to difficult real-world problems.”²⁶ Optimization research implements these features (reproduction, mutation, competition, and selection) in computers to find “natural” laws that can govern the organization of industrial or other processes that, when implemented on a broad scale, become the conditions of life itself.

This vision of optimization then justifies the extension and intensification of the zonal logic of smartness. To optimize all p. 121 aspects of existence, smartness must be able to locate its relevant populations (of preferences, events, etc.) wherever they occur. However, this is possible only when every potential data point (i.e., partial perception) on the globe can be directly linked to every other potential data point without interference from specific geographic jurisdictional regimes. This does not mean the withering of geographically based security apparatuses; on the contrary, optimization often requires strengthening these to protect the concrete infrastructures that enable smart networks and to implement optimization protocols. Yet, like the weather or global warming, optimization is not to be restricted to, or fundamentally parsed by, the territories that fund and provide these security apparatuses but must be allowed to operate as a sort of external environmental force.

Resilience

If smartness happens through zones, if its operations require populations, and if it aims most fundamentally at optimization, what is the telos of smartness itself? That is, at what does smartness aim, and why is smartness understood as a virtue? The answer is that smartness enables resilience. This is its goal and raison d’être. The logic of resilience is peculiar in that it aims not precisely at a future that is “better” in any absolute sense but at a smart infrastructure that can absorb constant shocks while maintaining functionality and organization. Following the work of Bruce Braun and Stephanie Wakefield, we describe resilience as a state of permanent management that does without guiding ideals of progress, change, or improvement.²⁷

The term resilience plays important, though differing, roles in multiple fields. These include engineering and material sciences— since the nineteenth century, the “modulus of resilience” has measured the capacity of materials such as woods and metals to return to their original shape after impact—as well as ecology, psychology, sociology, geography, business, and public policy, in which resilience names ways in which ecosystems, p. 122 individuals, communities, corporations, and states respond to stress, adversity, and rapid change.²⁸ However, the understanding of resilience most crucial to smartness and the smartness doctrine was first forged in ecology in the 1970s, especially in the work of C.S. Holling, who established a key distinction between “stability” and “resilience.” Working from a systems perspective and interested in the question of how human beings could best manage elements of ecosystems of commercial interest (e.g., salmon, wood), Holling developed the concept of resilience to contest the premise that ecosystems were most healthy when they returned quickly to an equilibrium state after being disturbed (and in this sense his paper critiqued then current industry practices).

Holling defines stability as the ability of a system that had been perturbed to return to a state of equilibrium, but he argued that stable systems were often unable to compensate for significant, swift environmental changes. As Holling writes, the “stability view [of ecosystem management] emphasizes the equilibrium, the maintenance of a predictable world, and the harvesting of nature’s excess production with as little fluctuation as possible,” yet this approach that “assures a stable maximum sustained yield of a renewable resource might so change [the conditions of that system] … that a chance and rare event that previously could be absorbed can trigger a sudden dramatic change and loss of structural integrity of the system.”²⁹ Resilience, by contrast, denotes for Holling the capacity of a system to change in periods of intense external perturbation and thus to persist over longer time periods. The concept of resilience encourages a management approach to ecosystems that “would emphasize the need to keep options open, the need to view events in a regional rather than a local context, and the need to emphasize heterogeneity.” Resilience is thus linked to concepts of crisis and states of exception; that is, it is a virtue when crises and states of exception are assumed to be either quasi-constant or the most relevant states. Holling also underscores that the movement from stability to resilience depends upon an epistemological shift: “Flowing from this would be not the presumption of sufficient knowledge, but the recognition of our ignorance: not the assumption that future events are expected, but that they will be unexpected.”³⁰

Smartness abstracts the concept of resilience from a systems approach to ecology and turns it into an all-purpose epistemology and value, positing resilience as a more general strategy for managing perpetual uncertainty and encouraging the premise that the world is indeed so complex that unexpected events are the norm. Smartness enables this generalization of resilience in part because it abstracts the concept of populations from the specifically biological sense employed by Holling. Smartness sees populations of preferences, traits, and algorithmic solutions, as well as populations of individual organisms. Resilience also functions in the discourse of smartness to collapse the distinction between emergence (something new) and p. 123 emergency (something new that threatens), and does so to produce a world where any change can be technically managed and assimilated while maintaining the ongoing survival of the system rather than of individuals or even particular groups. The focus of smartness is thus the management of the relationships between different populations of data, some of which can be culled and sacrificed for systemic maintenance.³¹ Planned obsolescence and preemptive destruction combine here to encourage the introduction of more computation into the environment—and emphasize as well that resilience of the human species may necessitate the sacrifice of “suboptimal” populations.

The discourse of resilience effectively erases the differences among past, present, and future. Time is not understood through an historical or progressive schema but through the schemas of repetition and recursion (the same shocks and the same methods are repeated again and again), even as these repetitions and recursions produce constantly differing territories. This is a self-referential difference, measured or understood only in relation to the many other versions of smartness (e.g., earlier smart cities), which all tend to be built by the same corporate and national assemblages.

The collapse of emergence into emergency also links resilience to financialization through derivation, as the highly leveraged complex of Songdo demonstrates.³² The links that resilience establishes among emergency, financialization, and derivatives are also exemplified by New York City, which, after the devastation of Hurricane Sandy in 2012, adopted the slogan “Fix and Fortify.” This slogan underscores an acceptance of future shock as a necessary reality of urban existence, while at the same time leaving the precise nature of these shocks unspecified (though they are often implied to include terrorism as well as environmental devastation). The naturalization of this state is vividly demonstrated by the irony that the real destruction of New York had earlier been imagined as an opportunity for innovation, design thinking, and real estate speculation. In 2010, shortly before a real hurricane hit New York, the Museum of Modern Art (MoMA) and P.S.1 Contemporary Art Center ran a design competition and exhibition titled Rising Currents, which challenged the city’s premier architecture and urban design firms to design for a city ravaged by the elevated sea levels produced by global warming:

MoMA and P.S.1 Contemporary Art Center joined forces to address one of the most urgent challenges facing the nation’s largest city: sea-level rise resulting from global climate change. Though the national debate on infrastructure is currently focused on “shovel-ready” projects that will stimulate the economy, we now have an important opportunity to foster new research and fresh thinking about the use of New York City’s harbor and coastline. As in past economic recessions, construction has slowed p. 124 dramatically in New York, and much of the city’s remarkable pool of architectural talent is available to focus on innovation.³³ A clearer statement of the relationship of urban planners to crisis is difficult to imagine: Planning must simply assume and assimilate future, unknowable shocks, and these shocks may come in any form. This stunning statement turns economic tragedy, the unemployment of most architects, and the imagined coming environmental apocalypse into an opportunity for speculation— technically, aesthetically, and economically. This is a quite literal transformation of emergency into emergence that creates a model for managing perceived and real risks to the population and infrastructure of the territory not by “solving” the problem but by absorbing shocks and modulating the way environment is managed. New York in the present becomes a mere demo for the postcatastrophe New York, and the differential between these two New Yorks is the site of financial, engineering, and architectural interest and speculation.

This relationship of resilience to the logic of demos and derivatives is illuminated by the distinction between risk and uncertainty first laid out in the 1920s by the economist Frank Knight. According to Knight, uncertainty, unlike risk, has no clearly defined endpoints or values.³⁴ Uncertainty offers no clear-cut terminal events. If the Cold War was about nuclear testing and simulation as a way to avoid an unthinkable but nonetheless predictable event—nuclear war—the formula has now been changed. We live in a world of fundamental uncertainty, which can only ever be partially and provisionally captured through discrete risks. When uncertainty, rather than risk, is understood as the fundamental context, “tests” can no longer be understood primarily as a simulation of life; rather, the test bed makes human life itself an experiment for technological futures. Uncertainty thus embeds itself in our technologies, both those of architecture and of finance. In financial markets, for example, risks that are never fully accounted for are continually “swapped,” “derived,” and “leveraged” in the hope that circulation will defer any need to represent risk, and in infrastructure, engineering, and computing we do the same.³⁵

As future risk is transformed into uncertainty, smart and ubiquitous computing infrastructures become the language and practice by which to imagine and create our future. Instead of looking for utopian answers to our questions regarding the future, we focus on quantitative and algorithmic methods and on logistics—on how to move things rather than on questions of where they should end up. Resilience as the goal of smart infrastructures of ubiquitous computing and logistics becomes the dominant method for engaging with possible urban collapse and crisis (as well as the collapse of other kinds of infrastructure, such as those of transport, energy, and finance). Smartness thus becomes the organizing concept for an emerging form of technical rationality whose major goal is management of an p. 125 uncertain future through a constant deferral of future results; for perpetual and unending evaluation through a continuous mode of self-referential data collection; and for the construction of forms of financial instrumentation and accounting that no longer engage (or even need to engage with), alienate, or translate what capital extracts from history, geology, or life.

Smartness and Critique

Smartness is both a reality and an imaginary, and this comingling underwrites both its logic and the magic of its popularity. As a consequence, though, the critique of smartness cannot simply be a matter of revealing the inequities produced by its current instantiations. Critique is itself already central to smartness, in the sense that perpetual optimization requires perpetual dissatisfaction with the present and the premise that things can always be better. The advocates of smartness can always plausibly claim (and likely also believe) that the next demo will be more inclusive, equitable, and just. The critique of smartness thus needs to confront directly the terrible but necessary complexity of thinking and acting within earthly scale—and even extraplanetary scale—technical systems.

This means in part stressing, as we have done here, that the smartness mandate transforms conditions of environmental degradation, inequality and injustice, mass extinctions, wars, and other forms of violence by means of the demand to understand the present as a demo oriented toward the future, and by necessitating a single form of response—increased penetration of computation into the environment—for all crises. On the other hand, not only the agency and transformative capacities of the smart technical systems but the deep appeal of this approach to managing an extraordinarily complex and ecologically fragile world are impossible to deny. None of us are eager to abandon our cell phones or computers. Moreover, the epistemology of partial truths, incomplete perspectives, and uncertainty with which Holling sought to critique capitalist understandings of environments and ecologies still holds a weak messianic potential for revising older modern forms of knowledge and for building new forms of affiliation, agency, and politics grounded in uncertainty, rather than objectivity and surety, and in this way keeping us open to plural forms of life and thought. However, insofar as smartness separates critique from conscious, collective, human reflection—that is, insofar as smartness seeks to steer communities algorithmically, in registers operating below consciousness and human discourse—critiquing smartness will in part be a matter of excavating and rethinking each of its central concepts and practices (zones, populations, optimization, and resilience), as well as the temporal logic that emerges from the particular way in which smartness combines these concepts and practices.