THE TERM “PLATFORM” is everywhere, but it’s not clear if it’s a metaphor or a thing, a new condition in the digital era or semantic camouflage for the banal evil of capitalism. Platforms are raised areas that facilitate — and leave open — exchange and social activity. As long as software platforms were contained behind personal computer screens and locked into physical infrastructures, the metaphor seemed innocuous. But now meatspace and cyberspace have fused. As a recent how-to for the new business era, Platform Revolution, puts it: “A platform is a business based on enabling value-creating interactions between external producers and consumers,” providing “an open, participative infrastructure for these interactions” and setting “governance conditions for them.” This model of privatized governance is spreading. Production and distribution, services and the social: all have been “disrupted” by the rules of the platform. Tom Goodwin observed in 2015 that “Uber, the world’s largest taxi company, owns no vehicles.” And the same could be said for Facebook (media), Alibaba (retail), and Airbnb (hotels). It’s true that none of these platforms owns the goods their services enable. But this now ubiquitous observation raises more questions than it answers.

Are these platforms skimming rent off capital and labor? Or do they represent a fundamental shift in economics, a new Industrial Revolution? The second view, espoused by the pop-management guru team Erik Brynjolfsson and Andrew McAfee, holds that that “cybernation” will automate mental labor in the same way the factory automated the work of the arm. Of course, the “unburdening” of the arm unleashed the horrors of 19th-century industrial labor, so we have reason in advance to suspect such boosterism. Cognitive capitalism, to use Yann Moulier Boutang’s term, might be less about allowing creativity to organize the economic cycle than about siphoning value from socio-cultural activity as such. Companies such as Alphabet, with a market cap in the neighborhood of three quarters of a trillion dollars, have claimed to be neutral arbiters and spaces of informational exchange. No one really believes that anymore, but we lack language to grasp the way these platforms collapse profit and the social, culture and capital. As the media scholar Tarleton Gillespie has argued, the term “platform” tendentiously fuses several meanings to the benefit of these businesses, combining the software platform with the figurative and political senses of the word associated with freedom. Yet criticizing the propaganda of such usage is a less urgent intellectual task than trying to understand what the platformed world we now inhabit looks like.

As we live more of our lives on platforms such as Facebook, even the line between mind and matter is up for grabs. Think about Elon Musk’s proposal to jack your brain into the social network directly, surpassing the necessity for typing. Mark Zuckerberg is a fan of this proposal, since whatever gets platformed — in this case, your mind — also becomes data owned by the platform-owners. Imagine that your mind’s every motion is given for “research purposes” to a prestigious academic, who passes it on to a big data company owned by a secretive billionaire affiliated with a far-right ideology. That’s not a William Gibson plot, but instead basically what happened in the 2016 US presidential election, when avant-garde (or maybe bullshit) psychometrics was used on tens of millions of users’ data by a shadowy company called Cambridge Analytica. Imagine Steve Bannon — who sat on that company’s board — inspecting what even you don’t know about your own mind. In response to this flap, the call to regulate has gotten louder than ever, even reaching the pages of The Wall Street Journal. But regulation isn’t enough. We need theory.

That doesn’t mean we shouldn’t regulate the big platforms, correct for algorithmic bias, and fight troll-farm manipulation. Scholars such as Safiya Umoja Noble, Frank Pasquale, and Kate Crawford have offered among the most sophisticated and compelling visions of what such regulations would look like. But what if we can’t return Pandora’s Platform to the Amazon warehouse we ordered it from? Coping with platform capitalism — a term coined by German theory blogger Sascha Lobo — means figuring out how we need to change our vocabulary and categories, and how we need to update our critical frameworks, for our new reality.

¤

Franklin Foer’s World Without Mind: The Existential Threat of Big Tech, is not up to the task. Foer’s book is The West Wing of broadsides in the increasingly heated debate about tech, wistfully grafting dead values onto a new reality. It points out, with erudition and whip-smart turns of phrase, just how bad things are. Its title promises a lot more than it gives, since the book focuses heavily on Foer’s sour experience with Facebook co-founder Chris Hughes and the digital takeover of the New Republic, which led to Foer’s eventual unceremonious firing. The broadside is expert. Foer argues that “the ascendant monopolies of today aspire to encompass all of existence,” in an attempt to “overhaul the entire chain of cultural production, so that they can capture greater profit.” These attempts have led to a monopoly situation, but a monopoly over things like our ability to “contemplate.” Jeff Bezos inserts himself as gatekeeper into the journalistic community — most obviously through his purchase of the Washington Post — but disavows the role of gatekeeper itselfhas returned to a central role,, deferring editorial judgments to clicks.

This is certainly a necessary critique, but it’s skewed toward Foer’s experience in publishing, so much so that the model he wishes we could return to is simply the old-world publishing empire associated with what he calls the “Acela corridor elite.” The platform monopolies profess “the values of the sixties” and want to “radically remake the production of culture,” shifting from a “competitive marketplace of ideas” (governed by gatekeepers like Foer himself) toward a “conformism” that threatens to flatten cultural creativity to consensus. These behemoths pay pathetically little in taxes, a wrong that might be righted, he writes hopefully, by “The Big One,” a hack to end all hacks. This hope is hackneyed (sorry). The longer-term solution, he argues, is to embrace paper and long-form reading, to voluntarily adopt non-platformed cultural diets. He calls this the “organic mind,” on analogy to Michael Pollan’s recommendations that we simply eat “real food.” We might remember here, though, that Amazon owns Whole Foods. The voluntarist solution falls flat in the face of the sheer size of the networks in question. Competitors in small retail and small publishing can’t afford to opt out of Amazon. The call for a “Data Protection Authority” (something like the General Data Protection Regulation set to go into effect in the EU in May) comes late in the book and is too vague to be compelling. Foer establishes a kind of fictional space in which the values of yesteryear are the governing rules of critique. As he puts it, he was known at the New Republic as a “nostalgist.” And he is. His nostalgia makes his critique lose its bite, since its categories are already outdated. We might take a more general lesson: calls to regulation based on pre-digital critical frameworks make critique conservative. Lacking the conceptual tools to envision progress, liberals are the conversatives now.

Scott Galloway’s The Four, which compares Apple, Amazon, Facebook, and Google to the horsemen of the apocalypse, provides a great deal more information on how to “break them up.” Let’s be clear: this suggestion to splinter the largest platforms in history comes from a celebrity business professor. In taking over social and historical functions that should be isolated from enterprise, the platforms have immiserated too many, flattened growth, and spread unemployment. If you want to read Galloway, you’ll have to put up with a lot of pop psychology, shading into New Age doctrine. Google appeals to the brain, Apple to the genitals (mostly for men, though), and so on. But it’s pretty much worth it.

Illustration from Galloway’s “Silicon Valley’s Tax-Avoiding, Job-Killing, Soul-Sucking Machine,” February 9, 2018, “Esquire”

Galloway has a reputation as a gadfly. He likes to go to international business conferences, he tells the reader, and inform CEOs their companies are the biggest threat to civilization. (They don’t seem to mind, because they’re holding “aces,” as he puts it, all of which are “data.”) Galloway suggests that platforms run on culture rather than traditional profit. Bezos, for example, has maintained a massive influx of venture capital even as his company has failed to turn a profit (this change in the final quarter of 2017, but half the new profit came from the Republican tax overhaul). Galloway attributes Bezos’s success to his knack for narrative, which in his view is far more important than profit in the age of the socialized platform. Ultimately, though, Galloway’s frameworks are almost self-consciously out of date. He is very impressed, for example, by Apple’s unorthodox move into brick-and-mortar retail, which provides ammunition for him to argue that a move back to the analog and the offline is underway. The first company to be valued at one trillion dollars, he argues, will be vertical down to the last mile of delivery of product or service. (Since the book came out, the Four’s market caps have shot up: Apple, depending on the day, is well within 200 million dollars of the coveted status.) And this “return to the analog” inspires, in Galloway as by so many others, a return to industrial and often pre-modern terminology. The “horsemen” themselves, for example, might be associated more with the medieval period than any other. Likewise, Galloway recommends that platforms build “analog moats” around their digital businesses. Foer, for his part, says Amazon’s model for publishing returns authorship to a “medieval” model (which seems to be a bad thing in his view).

In their emphasis on the digital world’s medievalism, Foer and Galloway surprisingly join a chorus of Italian Marxists and cultural theorists who think the digital economy has brought a form of pre-modern economy back into capitalism. Rent has returned to a central role, as Carlo Vercellone argues. When growth levels off, ownership takes precedence over entrepreneurship. Rather than producing new value, the platforms simply coordinate virtual properties and charge for their use. But the properties are not in meatspace or cyberspace alone, which means the owners can set the rent at will. Think of Uber, which is only now beginning to try to create a more stable set of drivers (something like employees). Trying to keep drivers driving means negotiating with them, but the results are not encouraging. By denying their status as a firm with employees, Uber devolves the risk of enterprise onto their “contractors,” and then argues those contractors should be loyal to the platform’s internal, algorithmic assessment of its own success, since their ability to drive at all is based on Uber continuing to exist.

That situation resembles feudalism more than a bit, with the added freedom (read: risk) that individual drivers don’t even have the status of serfs. They are “free” to choose their lords, to whom they don’t even belong. The platform is an adventure in extreme forms of expropriation set against the backdrop of a slowing economy, what Marxist economist Robert Brenner calls “the long downturn” since the 1970s. That expropriation is increasingly conjuring pre-modern vocabulary, as in media theorist Benjamin Bratton’s apt formula, “cloud feudalism.” Science fiction author Bruce Sterling has recently argued that the Internet of Things is little more than a battle of lords in the new, online feudalism. There’s still a centralized federal government, but its authority is attenuated by platform monopolists. The platform confuses capital-flow and social form, rearranging the relationship of profit to community (and therefore class), and of intelligence to organization. With the incumbency effect that massive data hoarding affords companies like the Four, we appear to be looking at something like a set of smart monopolies, setting in question the long-stable political-economic vocabulary we use to describe and orient ourselves. We need new theory to face that situation, not just pre-modern analogies and calls to regulation lacking any sign of fulfillment in the current political landscape.

¤

Nick Srnicek’s Platform Capitalism takes a step in this direction. Among its virtues is a particularly clear breakdown of the types of platform. The advertising platform is dominant but risky. Google and Facebook, according to Srnicek, are dangerously extended in this historically unstable area. They are “appropriating data as raw material.” They are not, Srnicek makes clear, expropriating a surplus value produced by “online” or “cognitive” labor in the production of that data. Opposing the Italian Marxists mentioned above, and especially Maurizio Lazzarato’s notion of “immaterial labor,” Srnicek argues that the platforms are just rearranging deck chairs as capitalism stalls — at least for now. Meanwhile, the “cloud platform” is virtualizing computing services and splitting them across distant servers, working with the increasing speed of networks to outsource IT labor from individual firms. The massive incumbent in this field is Amazon Web Services (which generated most of the non-tax-deduction profits of its unprofitable parent company in Q4 of 2017). As with its competitors Microsoft Azure and others, Amazon is renting “an increasingly basic means of production for contemporary business.” Unlike in traditional renting, though, they get valuable data in return. The platform economy is not isolated to the digital, as witnessed by the “industrial platform.” General Electric and Siemens are two of the firms at the forefront of bringing the platform into the guts of the industrial process, establishing feedback loops in which massive data-harvesting in the production process comes to guide that process itself. This is in line with Sterling’s feudal image of the Internet of Things, with every object digitally identified using RFID (Radio Frequency Identification), allowing purchase and consumption data to make the production chain more flexible. What former Baidu chief scientist Andrew Ng calls “O2O” or “online-to-offline” services are dubbed “product platforms” by Srnicek. They are on-demand and treat their products as rentables. They’ll wash your car at the push of a button, but are also invading spaces, like commercial jet engine retail, usually thought to be firmly offline. A contract indicates you’re buying the product as a service, even if the product is as tangible as it gets (like a jet engine). Product and industrial platforms will likely enter a profitable symbiosis in the coming years. The so-called lean platform treats workers and wages in the same way, making the labor market on-demand and essentially renting labor power at platform-market prices. Srnicek calls this a “hyper-outsourced model” that allows only “a bare extractive minimum,” a “control over the platform that enables a monopoly rent to be gained.” For Srnicek, this is the outcome of a generation of outsourcing since the 1980s, and the corresponding fall in growth. National growth or no, the platforms are increasingly in control of the distribution of wealth.

Srnicek argues that, because of the incumbency effect, the platforms are specializing and may focus on optimizing their rent-extraction processes, rather than competing outside their core competencies. (It’s not clear how this works in the case of Google, for example, where the core competency is so broad that they renamed the parent company Alphabet in 2015.) This strategy is that of “enclosure,” something similar to the “analog moats” Galloway recommends. Srnicek suggests, for this reason, that “capitalist competition is driving the internet to fragment.” For Srnicek, “whereas the tech boom of the 1990s at least left us with the basis for the internet, the tech boom of the 2010s looks as though it will simply leave us with premium services for the rich.” Perhaps today we should “collectivize the platforms,” he concludes.

Again, we find ourselves firmly in the territory of wish-fulfillment, this time with brilliant analysis. There is a tendency throughout Srnicek’s book to downplay the platforms’ economic stability, since they do not dynamically produce the surplus value needed to keep profits heading north. But what if the surplus in fact stops growing? What if Brenner’s “long downturn” becomes permanent? Srnicek’s Marxist account perversely clings to the model of European industrial capitalism, suggesting that outright political control is an unlikely outcome for the Four.

But what if the age of the platform turns out to be stable, or to lead to some worse post-platform capitalism? Srnicek argues that the platform castle is made of sand. The logic of capital itself — profit — will win out. But while this might be true for specific companies, why should capitalism return to the rough shape it had from around 1800 until the 1970s? Feudalism itself wasn’t so unstable, and we might be witnessing not some epicycle in the general capitalist economy but the large-scale emergence of some new balance of the social and the profit-driven. What prevents the stabilization of a post-industrial model of capitalism that retains all the platform-features yet remains just as painful and unjust as ever — or even more so? Calls for collectivization that presuppose the inevitable collapse of the current order are not much more convincing than calls for regulation. Both types of call are often built on little more than desire without any institutional base — activist, militant, or otherwise — to sustain them. Don’t get me wrong: I’m all for collective platforms, and I think Srnicek’s is the best political account of the current situation. But the path to collectivization will require a more persuasive account of the nature of platform capitalism and of the social forces that will be required to oppose it. Pieces of that theory might come from unexpected sources.

¤

Brynjolfsson and McAfee are the Pindars of the artificially intelligent production cycle. Their latest, Machine, Platform, Crowd, opens with a section on “the triple revolution,” a reference to the influential 1964 open letter to Lyndon B. Johnson that described a new normal for global society based on fundamental shifts in “cybernation,” weaponry, and human rights. For Brynjolfsson and McAfee, the new revolutions are in their title, each providing a “lens” for contemporary economy and culture. Machines and minds are in a new balance, as are platforms and products, and crowds and “cores,” the operational decision-making heart of the firm. Their earlier effort, The Second Machine Age, described a productivity altered by computational technologies. The new book is for a time when “science fiction technologies — the stuff of movies, books, and the controlled environments of elite research labs — started to appear in the real world.” The whole thing is organized on a “science fiction gradient,” or “ascending order of weirdness.” The book justifies that description, going from the entrance of machine learning into managerial decision-making to the fate of the Decentralized Autonomous Organization, a blockchain-based investor collective that demonstrated through failure why we can’t automate economic planning and social trust altogether.

Brynjolfsson and McAfee argue that many disciplines other than economics are needed to grasp the present situation. Indeed, the platform condition requires more than the current division of the faculties can analyze. This is because our situation alters the nature of judgments, value, and society as such. This point is right on and should inform the sensibility of cultural-critical work going forward. But the book marshals these other disciplines for the sole purpose of making recommendations to managers (or really, owners).

Expert decisions in firms need to take up less oxygen than their machine-learned counterparts. For Brynjolfsson and McAfee, this means replacing the standard model that firms have used since the 1990s, where humans make judgments based on data. Using Daniel Kahneman’s division between the rational, logical, and evolutionarily recent cognitive “system 2” and the fast, intuitive, old “system 1,” they argue that managers need to creatively find ways to defer to good automatic data processing, because “today, examples of valuable, high-quality, 100% automatic decisions are all around us.” This means turning the “standard arrangement on its head: instead of having machines provide data as an input to human judgment,” companies should have “judgment serve as an input to an algorithm.” Let neural nets take the lead. Humans can serve the function of convincing others to go along. Logic is for computers, rhetoric for humans. And while individual managers should defer to algorithms, the core of the firm needs to incorporate the instantaneous crowd-sourced data that platforms provide. This data is decentralized “relevant knowledge” that will bypass the human knowers altogether. Noting the resonance of their argument with the work of Austrian neoliberal Friedrich Hayek, whom they dub the “patron saint of the crowd,” Brynjolfsson and McAfee argue that the firm is itself a block against market forces, since it centralizes data-valuation processes and tries to survive, even at the cost of market efficiency. But they avoid the dream of a firmless society, showing how the human hand comes back into play even in the most automated of market processes.

Their advice to entrepreneurs: Incorporate more market-like mechanisms into the firm itself, doing away with the “internal decision-making and resource allocation processes of many organizations” that “still look a lot like those of centrally planned economies.” (I’m not making this up.) So humans enter the picture again, being forced to decide what counts as market-driven and what is planned. This line of thinking pushes back, however mildly, against the most extreme versions of neoliberal orthodoxy. Rather than letting monetization destroy institutions, Brynjolfsson and McAfee make a plea for a minimal but socially crucial island in the survival of the economically fittest: the firm itself, the corporation. This is a humanism of the strangest kind, one that would save the market from itself. It mostly ends up entailing more offshoring, as Srnicek could have predicted. The final line of human defense against the fantasy of a total decentralization of everything turns out to be companies. Society can only be defended by the corporation. But the corporation will only survive if it continues the data-driven externalization of its assets and decision-making processes.

Brynjolffson and McAfee celebrate the end of doctor-driven health care, favoring machine learning diagnostics, and they commend the online university Udacity for outsourcing the grading of code projects in its computer science courses. Wages and jobs receive only a cursory mention in the final pages. True, it’s a bad time to be a worker, they tell us — but then, there aren’t workers anymore, only entrepreneurs. Entrepreneurs are a sort of meta-worker, those who “specialize in inventing” new jobs, something machines are not very good at. This cheerful analysis concludes that “[this] is why being an entrepreneur is one of the most rewarding jobs both for individuals and for society.” But this begs the question: entrepreneurs are not workers with jobs, and managers who don’t make decisions might not be earning their keep in the one percent. Firms are (or should become) just algorithmic mirrors of crowds, formal organizations that facilitate the monetization of social activity — that is, platforms. Platforms are a computational interface between society and capital. They host the performance of that interaction. And in this way, they are also a kind of outsourced and even crowdsourced mind, in which signs and concepts are in direct, exploitable communication with production and valuation. Mind and matter collapse, but the resulting metaphysics always profits the platform. Subtract the ghoulish enthusiasm, and this is one of the most compelling descriptions of the present I have found anywhere.

The digital is free (or incredibly cheap), perfect (as in, the copy of data is perfect), and instantaneous (or so fast it may as well be). It’s these affordances of digital technology that are permeating the entire production cycle, according to Brynjolfsson and McAfee. Platforms are the institutional form this shift takes. They are “online environments that take advantage of the economics of free, perfect, and instant,” a “digital environment characterized by near-zero marginal cost of access, reproduction, and distribution.” The internet is the platform of platforms, the facilitator of all this new facilitation. Brynjolfsson and McAfee never worry about whose margin is being zeroed — creators of content or product, those who own the infrastructure to distribute, or otherwise. The platforms do belong to someone, though, and they “can capture much, most, or even all of the value as they spread throughout an industry.” Apple, they report, keeps 30 percent of paid app purchases, part of the six billion dollars in app-store revenue for the company in 2015. The keyword is “access,” a straight line between firm and consumer that marginalizes all the other elements, floating their wages and rent on the “free” market now co-determined by the algorithm and the behavior of its platformed consumers. Everyone can participate, of course — and the whole cycle gets more efficient as a result. Platforms “improve our quality of life while simultaneously helping us tread more lightly on the planet,” since fewer total goods will be produced in the future as fewer unused goods are produced at all. Brynjolfsson and McAfee imagine a fully automated luxury communism for the very, very few. Just like the suggestions to regulate and collectivize the platforms, Machine, Platform, Crowd reads like a set of suggestions that no one could really follow through on unless they already controlled an incumbent platform. The book is terrifyingly realist about the shift in categories we’re experiencing. So is Benjamin Bratton’s The Stack: On Software and Sovereignty, albeit in a different key.

¤

Bratton starts by calling for new vocabulary, saying we lack the critical names for the situation of “planetary-scale computing,” an “accidental megastructure” we have built and which is in turn “building us in its image.” Throughout the book, Bratton reserves a derisive yet analytical tone for those who do not accept what he elsewhere calls “the new normal.” Because we have designed The Stack — the sum of the parts of the megastructure — we have no choice but to treat it as a question of “platform design.” The only way to grasp and intervene in the situation is to meet it conceptually and pragmatically at scale. What results is the most synoptic view of this topic I’m aware of.

It was Sterling who suggested the term “stack,” saying in an online forum in 2013 that it made less sense to talk about “the internet,” “Silicon Valley,” or “the media,” than “to just study Google, Apple, Facebook, Amazon, and Microsoft,” these “big five American vertically organized silos” which were “re-making the world in their image.” Stacks are the looming end of the emancipatory promise of cyberspace. Why these companies don’t just “buy the Republican Party” puzzled Sterling. Bratton has an answer: platforms are neither businesses nor governments, but instead a “third institutional form, along with states and markets.” Cyber-politics can’t be traced in the old forms of laying down law and state lines. (Bratton spends a lot of time rearranging Carl Schmitt’s theory of the nomos to this end.) For Sterling, stacks were replacing the smooth space of the internet. For Bratton, naming the whole thing The Stack is a way of seeing the totality in spite of the fragmentation. “Software” is almost too small a metaphor to capture the ambition of his book, although it is one of the most compelling contributions to the still young field of “software studies.” The Stack is a cosmology of platform capitalism.

The cosmology has six layers: Earth, Cloud, Address, City, Interface, and User. The six sections of the book that treat these entities are a masterful summary of recent media studies of all kinds, often with original insights added. The six layers build on one another progressively, describing the world of global platforms. The earth layer treats the planetary scale of information-processing and natural resources. Information and energy connect the networked earth with the digital cloud, and Bratton argues that computation is now threatening to destabilize energy resources at unprecedented scales. The asymmetry between information and energy consumption both allows the cloud to exist in the first place and also threatens to disperse it by destroying its earthy infrastructure. Analyzing the sixth version of the Internet Protocol (IPv6) and RFID, Bratton treats the “universal” addressability spreading through the world of things. “Deep address,” as he calls it, coincides with the industrial internet and bleeds back into what we used to see as “offline.” Bratton’s analysis is at its best here. The study of what things are (ontology) shifts with the rise of planetary-scale categorization. “Addresses” penetrate through the cloud back to the earth.

What is strange and new is also infrastructurally uneven. The city layer collects the computational stacks of the previous layers and centralizes them geographically. But this geography is at least half digital, as we see with the rise of putative “smart cities,” which retrofit the mess of the metropolis to the smooth facilitation of the platform. This hybrid urban space is made possible by the interface layer. Bratton expands the interface beyond the screen, casting it as a regime of images that coordinates “agents” in The Stack. The interface allows the address layer to govern the final layer, the user. But the user is not necessarily human. Instead, it is any agent that can initiate a “column” of computation in The Stack. These columns of pullulating data connect the layers, since their semantic content is complemented by metadata and “meta-metadata,” allowing a column to be isolated and visualized at the level of each layer. Bratton has successfully described a totality of computation, from click to globe. The Stack is a kind of meta-platform made up of other platforms. These platforms, he concludes, synthesize exchanges of every kind and package it as searchable data. It is this synthesis that forms a third (digital) estate with state and market. Google’s PageRank algorithm is anything but a tool or an information source: it is instead its own set of columns in The Stack producing a form that alters market and government alike. Facebook’s political irresponsibility fits this model perfectly. Cambridge Analytica may be an alt-right project, but we shouldn’t reduce it to old categories, because it isn’t clear, even if our legal systems punish the perpetrators, how we could stop another breach, which might not come in the same form. It’s not clear how we can pull out of The Stack, and for Bratton, it’s clear that we can’t.

The Stack threatens to become Cloud Feudalism, Bratton writes, singling out the post-failure state of cities like Detroit and Fresno. The Republican Party that Sterling’s Stacks haven’t bought out is pushing in this direction, Bratton argues. By leaving climate change untreated and untreatable, for example, they are actively creating conditions in which neo-feudalism can stabilize itself. To combat the various senile medievalisms currently ascending, Bratton suggests the left should avoid the “unfortunate knee-jerk antidigital technology politics.” The planetary design project he imagines (without, however, filling in much detail) must accept that information is not just “about” the world but in the world. Critique that tries to roll back two generations of digital creep is always already obsolete. This point is somewhat buried in a very long and extremely dense book, but it deserves to be highlighted. Bratton is among the few who are attempting to redescribe the “new normal” theoretically.

Because The Stack is “incipient,” as he puts it, its form is not yet crystallized, which means that this theory is semi-fictional — a form of science fiction or theory fiction. This tends to leave the ambition of the book in a strange middle space, attempting a description of something it doesn’t quite claim as real, something the book wants to redesign out of existence before it can become real. It’s as though the ambition to design at planetary scale prevents Bratton from fully committing to his parallel ambition to write the theory of the emerging future. What results is a fascinating but uncertain cosmos, filled with computational whorls and layered nodes, but lacking a principle that guides the vision. The Stack’s cosmology presents a picture without a law, a pre-Newtonian physics of the digital. But it remains singular as an attempt to write a theory for the present. It leaves the reader — intentionally, I think — in a no man’s land between present and future, forcing us to confront the concrete results of the planetary struggle between computation and atavistic categories like (non-Stack) “earth” or “human.” Bratton’s approach avoids either bemoaning or celebrating the demise of the old categories. The vision of the whole that results hovers between a theory and a policy recommendation. Its virtue is to suggest that these must be aligned, that no recommendation to regulate or collectivize or return to a previous style of selfhood or media governance will do. Bratton leaves us with an impossible but necessary task.

The platform has gone from an operating and gaming system to a global economic factor, from a metaphor to the mechanism of cultural production. But suppose we regulate Amazon, break up Google, migrate entirely from Facebook to some other set of platforms — it is nevertheless hard to imagine a near future without very large-scale social and economic exchange facilitated by digital platforms. The social and the profit-driven are now alloyed even as state and market seem fragmented. The balance among mind, nature, and machine has been disrupted. I can see no signal that these transformations will revert to their previous state, as Foer, Galloway, and even Srnicek hope for, if for disparate reasons. Platform capitalism is not vapor superimposed on a fundamentally sound social-economic structure that will remain when it disperses. Bratton’s cosmology poses the question, what is to be done? and offers general guidelines, while Brynjolfsson and McAfee celebrate the potential for an automated luxury communism for an ever-smaller elite. The theoretical task and the political task are not usually so obviously intimate, yet their common resolution has rarely seemed so distant. What comes after a cyberpunk that levels fiction and reality, cyberspace and meatspace, theory and politics? Bratton suggests a planetary design that abandons the distinctions between information and matter, thinking and unthinking, the virtual and the real. Those distinctions have already been abandoned by The Stack. This is why Brynjolfsson and McAfee’s recommendations for platform managers ring truest — and bleakest — of all the suggestions in these books. We mostly stumble from one catastrophe to another, avoiding the direst consequences and ignoring the infinite task of squaring theory and policy in the era of the platform.

¤

Leif Weatherby is assistant professor of German at New York University and author of Transplanting the Metaphysical Organ: German Romanticism between Leibniz and Marx. He is working on a book about a strange encounter between cybernetics and German idealism.