The Stack On Software and Sovereignty

72

Platform and Stack, Model and Machine

The astronomers leave for the Southern Hemisphere, the physicists for the equator, and the geometers leave in order to measure the earth’s meridians. The Bureau of Longitudes is created, universal geography is founded. The world is no longer the empirical domain of time and space. It is the compass of knowledge.

— Michel Serres, Jules Verne¹

And with these, the sense of the world ’ s concreteness, irreducible, immediate, tangible, of something clear and closer to us: of the world, no longer as a journey, having constantly to be remade, not as a race without end, a challenge having constantly to be met, not as the one pretext for a despairing acquisitiveness, nor as the illusion of a conquest, but as the rediscovery of a meaning, the perceiving that the earth is a form of writing, a geography of which we had forgotten that we ourselves are the authors.

— Georges Perec, “ The World ” ²

Molecules don ’ t have passports.

— Carl Sagan ³

The foundational layer within The Stack is the Earth itself. All movement through the lower machine layers draws on the chemistry and the physics of the Earth layer — its energy and minerals, scale and curvature, heat and cold, and so on. There is no planetary-scale computation without a planet, and no computational infrastructure without the transformation of matter into energy and energy into information. But for The Stack at least, what is computation, and how does the computational infrastructure at the Earth layer support the accidental megastructure? “ Computation ” is not only what The Stack is made from; it is also how the megastructure composes, measures, and governs itself. At the Earth layer, algorithms and electrons interweave at landscape scale, driving continental economies; in turn, those landscapes are disciplined by other algorithms hoping to rationalize the enormous energy appetite of the whole. We will examine this recursion from the ground up. The first sections of this chapter consider the materiality of computation in itself, before any artificial computing machines came onto the scene, and will ask if computation was “ discovered ” more than it was invented.

Earth Layer

The Earth layer is also made from the Earth itself, as the terraforming imperative of the Stack megastructure disembowels geological resources toward global conversions. These industrial processes are also as a kind of composition, one for which alternative geoaesthetics may point toward different outcomes. We will also consider how computational infrastructures at the Earth layer extend the planet ’ s capacities to sense and monitor its own energy usage by augmenting its “ skin. ” This is represented by grids that can rationalize energy use, but themselves may demand fatally large amounts of energy to construct. These contradictions contribute to uneven realignments of geopolitical jurisdiction according to the opposed positions of energy-producing geographies that most directly cause climate change and of those most affected by it. For The Stack, the identification, quantification, management, visualization, and provision of energy may serve as the last instance referent of economic value, and it may in time force the evolution of a platform capable of composing and governing such a system. This in turn may put the design of the Earth layer of The Stack in the untenable position of working on behalf of the exceptional “ emergencies ” that most threaten the platform ’ s coherency, such that in the decades to come, the self-amplifying logics of ecological governance demand not only geoengineering, but also incredible computational energy capture-and-distribution megastructures far beyond our current capabilities.

1. Discovering or Inventing Computation?

For the relationship between computation and its terrestrial substrate, the Earth, it is never easy to separate metaphor from physics, and so for my thesis, their conceptual interrelation is perhaps just as important as their material involvements. The Earth layer of The Stack draws from both the conceptual and the material, not by collapsing them into one so much as tracing ever-thickening lines back and forth between them. As I imagine those lines thickening, I become immersed in a photograph of the philosopher Gilles Deleuze, taken during a visit to Big Sur in the mid-1970s, as he sits on the California beach examining its sands and the breathing striations of the silica terrain. ⁴ The world remaking itself in waves, bit by bit and pebble by pebble. In trying to place the image, I also think about how just up the 101 freeway from where Deleuze was sitting, silicon was being repurposed as the physical medium of synthetic computational intelligence, in the areas near Palo Alto in a “ valley ” already having been named in honor of this element. For me these are conjoined, and not just by their geographic proximity: Deleuze on the beach contemplating (we might assume) what he called “ the plane of immanence, ” the field from which all potential forms emerge, and

76

Earth Layer

Intel ’ s initial approximations of microprocessor technology for universal computation, putting a mini-Turing machine on a silicon wafer. ⁵ In different ways and for different ends, both grapple with matter as vibrant, contingent, and mutable, as reproduced in the careful calculation of sets of differences drawn from particular virtual possibilities. At the end of the day, Deleuze ’ s philosophy is more about chemistry than computation, continuities more than discrete digitalizations, but his philosophical imagery of worlds appearing from the multiplication of imminent processes and generic diagrams, on oscillations of the physical and the virtual, is not unfamiliar to the projects of information realism. In the words of the late Friedrich Kittler, “ Silicon is nature! Silicon is nature calculating itself. If you leave out the part of engineers who write little structures on silicon you see one part of matter calculating the rest of matter. ” ⁶

The shimmering idea that the world is composed not of given forms on a fixed stage but of an atomic field of flux and churn is ancient. The idea precedes our ability to mathematize the hypothesis experimentally and predates by millennia the engineering of machines that can simulate a calculation of discrete bits of information as if they are those atoms. Close your eyes and visualize dust motes floating and falling in the white light of a projector. See them just barely touch or miss one another. This swirling and tumbling through the void is also, given some poetic license, one model of elemental computation. Around 1 A.D., Lucretius called these atomic bits primordia or seminarerum, and for the Epicurean philosophical tradition, this flux is ontological and the basis of their own information theory avant la lettre. ⁷ It says that what seems to be naive observation as solid figures and grounds, withdrawn into themselves and oriented as objects, are but clusters of bits that have fallen into one another over time, and will in more time fall apart and again into other things, conjugating or calculating themselves again and again. The name for the force of collision that causes their downward arcs to tumble into assemblage is translated from the Latin as swerve. Atomic bits swerve, as if by accident, and in their accumulation, the entropy of the noisy void gives way to the negentropic formulation of the world and its temporal orderliness: from this calculation, forms form. Lucretius called this economy of entanglement between atoms, located by their fluid communication in flight, the clinamen, and it has been the source of considerable philosophical and literary rumination (including Marx ’ s doctoral dissertation).

Today, enjoying a vantage point that includes contemporary atomic physics, we see the clinamen less as a spontaneous lurch of some thing from its track (the universe as the eventual archive of these accumulated deviations) than as interlocking fields of stochastic probabilities structuring emergent order in this way or that. The details evolve, but the idea of calculative emergence persists. The basic innovations are well known. In twelfth century Majorca, Ramon Llull described logical machines, influencing Gottfried Leibniz, who developed a predictive calculus and a biliteral alphabet that, drawing on the I Ching, allowed for the formal reduction of any complex symbolic expression to a sequence of discrete binary states (zero and one, on and off). Later, the formalization of logic within the philosophy mathematics (from Pierre-Simon Laplace, to Gottlob Frege, Georg Cantor, David Hilbert, and so many others) helped to introduce, inform, and ultimately disprove a version of the Enlightenment as the expression of universal deterministic processes (of both thought and physics). In 1936, with his now-famous paper, “ On Computable Numbers, with an Application to the Entscheidungsproblem,” a very young Alan Turing at once introduced the theoretical basis of modern computing and demonstrated the limits of what could and could not ever be calculated and computed by a universal technology. Turing envisioned his famous “ machine ” according to the tools of his time to involve an infinite amount of “ tape ” divided into cells that can store symbols, moved along a stationary read-write “ head ” that can alter those symbols, a “ state register ” that can map the current arrangement of symbols along the tape, and a “ table ” of instructions that tells the machine to rewrite or erase the symbol and to move the “ head, ” assuming a new state for the “ register ” to map. The Church-Turing thesis (developed through the 1940s and 1950s) would demonstrate that Turing ’ s “ machine ” not only could simulate algorithms, but that a universal Turing machine, containing all possible such machines, could, in theory, calculate all logical problems that are in fact computable (a limit that Turing ’ s paper sought to identify). The philosophical implications are thorny and paradoxical. At the same moment that Turing demonstrates the mechanical basis for synthetic logic by machines (suggesting real artificial intelligence), he partially delinks the correlation between philosophical thought and machinic calculation. The implications continue to play out in contemporary debates from robotics to neuroscience to the philosophy of physics, as has Turing ’ s later conceptualization of “ thinking machines, ” verified by their ability to convincingly simulate the performance of human-to-human interaction, the so-called Turing test. ⁸ In the decades since Turing ’ s logic machine, computation-in-theory became computers-in-practice, and the digitalization of formal systems into mechanical systems and then back again, has become a predominant economic imperative. Through several interlocking modernities, the cal culation of discrete states of flux and form would become more than a way to describe matter and change in the abstract, but also a set of standard techniques to strategically refashion them as well. Computability moves from a universal logic to a generic technology (and so contemporary claims that this passage is reversible are both predictable and problematic). Although the twentieth century invented computers, it did not invent computation so much as it discovered it as a general force, and offered some initial basic tools to work with it more directly. We are, like everything else, also its product.

Earth Layer

77

This conceptual shift is important to how we hope to consider reforming The Stack. One of Turing ’ s signal achievements is to show that an artificial “ machine ” could approach, and even approximate, the scope of natural computation, as defined in a particular way. His innovation was the specific pairing of formal logic with industrial technology that was, even after Charles Babbage and Ada Lovelace ’ s Victorian-era calculating machines, by no means obvious in its implications. For measuring the significance of that pairing in relation to The Stack, it is important to distinguish the limits of formal computation, on the one hand, from what the limits of actual computational technologies can really do, on the other. These are two very different kinds of limits. While Turing ’ s hypothetical machine demonstrated the mathematical limits of formal computability, it also demonstrated that any problems that could be captured and expressed symbolically through a reduction to rational integers (which likely describes the vast plurality of things and events in the world as representable by intelligent creatures) could be simulated and solved by a machine engineered to do so, given enough time, materials, and energy. Anything expressed as computable information, regardless of the natural appearance, linguistic identity, or economic value, could be processed by a universal information machine programmed to do so and physically capable of running through enough operations. A strong computationalist philosophical position may also extrapolate from this that natural systems can be (and so must be) reducible to information and computational processes. ⁹ Problems arise when the notion that things are formally equivalent by their shared computability slides into the claim that they are therefore ontologically equivalent, or even culturally and economically equivalent. The questions raised by the idea of universally calculable matter are interesting on both practical and philosophical terms, but I raise them here to provide conceptual context for other questions.

78

Earth Layer

At the time of this book ’ s publication, no one (certainly not I) can pronounce on the practical validity of quantum computing or industrial-scale atomic-level design, and so for that reason alone, we are careful to separate computation from computers, and not to confuse the mathematical genericity of computation as a process with the actual and comparatively feeble algorithm-crunching machines at work now and in the near future. ¹⁰ Surely the latter is a sad cartoon of the former, and Turing ’ s model of a universal computing machine specifies a break between what mathematics could describe as the computation of natural information and what artificial computing technology could ever program or perform. Even supercomputing grids are just machines particularly efficient at calculation tasks at predictable speeds, but they are not in themselves “ computation, ” just as light bulbs are machines good at conducting electrical currents toward illumination but are not in themselves electricity or light. Many things process information algorithmically and could be said to “ compute ” in a meaningful sense (DNA and RNA, for example) without also demanding that we must see in them the reflection of our computing machines. ¹¹ We might even assume that the “ next machines ” (the ones that come after planetary-scale computation) will look less like today ’ s computers and more like biology itself. ¹² Turing described a capacity to mimic natural computation, providing a measurement of the gap between it and artificial machinic computation, but as we learn to design by the comingling of bits and atoms into strange hybrids, we may in time need to retake that measurement. ¹³ Even if so, we will still resist the conflation of the quasi-universality of theoretical computation with the scope of real or near-future computational infrastructure, not to mention the indeterminate geopolitical effects of that infrastructure on those effects. Unfortunately, both utopian and dystopian alliances are happy to underwrite and expand that conflation to support their own deterministic narratives, but for The Stack, we won ’ t. One hopes that instead of dressing up contemporary techniques as ontological principles, an emphasis on the discovery (versus invention) of computation should make the practical distinction between the formal and the functional more apparent. The geopolitical effects of accidental computational megastructures remain design problems precisely because they are not determined by inflated notions about immanent (and imminent) logics.

Earth Layer

79

Anything else risks misleading conclusions drawn from deceptively conditional extrapolations from the present. For some, logarithmic arcs like Moore ’ s law suggest exponentially accelerating change in the most general sense and lead them to posit a big computer in the night sky, or a quasi-theological master operating system, and to identify computation as some vital force pulsing through all media. ¹⁴ From there, a future horizon filled with formal convergence, magic, and rapture looks inevitable. For some, that is the best possible news and for others the worst, but they can agree on the basic outline of a script that is too simplistic. At the same time, perhaps this genre of extrapolative futurism is just another way of saying that mathematics is universal and algorithmic technologies are medium independent, and so we need to employ them differently than we do now (which is true). Or, perhaps, is it but an elegant humming along with the most superficial and poorly theorized aspects of digital capitalism as it dissolves raw materials into interchangeable goods and services, and so the melody is rewarded for flattering the illusion that these linear speculations represent timeless principles? Industrial capitalism had social Darwinism, and today do we have instead a social Turingism: financialization as social metaphysics? Perhaps it is both (and more besides). Like Freud ’ s figural model of the mind in the rough image of the thermodynamic industrial engine, ¹⁵ the paradigmatic idea that the ongoing reassemblage of the world is, in some literal sense, given to computational processes is more than just ideological metaphor run amok. ¹⁶ It is more than partially true because we have made it true. The long-term technical question is to what extent we can describe and practically manipulate material systems by algorithmic calculation — flipping bits on and off, like the motes present or absent at any given point in the light, and gathering them into the forms we recognize as the world — and to what extent not at all, and more important, when we can and cannot do so, and toward what outcomes they should be directed.

As is well known, hyperfast algorithmic asset trades flourish in global equities and futures markets and account for huge percentages of all executed trades; their swerves also form value but do so in ways that can exceed or outpace human oversight. Algorithmic capitalism ’ s own story is often too happy to confuse means and ends, but perhaps counterintuitively, we may conclude nevertheless that the synthetic inhumanity of computational capital is actually the most direct vector out of the anthropocentric humanism that places short-term human needs at the center of public philosophy and engineering. The limits of machinic calculation are not the same as the limits of deterministic rationality, and the social effects of computational systems are certainly given to creative accidents. ¹⁷ Reactionary analog aesthetics and patriotisms, Emersonian withdrawal, and deconstrucivist political theology buy us less time and far less wiggle room than they promise, even less actually than the unfortunate notion that planetary-scale computation could emerge and mature without fundamental constitutive violence against traditional (that is, “ modern ” ) concepts of individual, society, and sovereignty. Because they simulate logic but are not themselves necessarily logical, computers make the world in ways that do not ultimately require our thinking to function (such as the interactions between high-speed trading algorithms that even their programmers cannot entirely predict and comprehend). The forms of inhuman intelligence that they manifest will never pass the Turing test, nor should we bother asking this of them. It is an absurd and primitive request. ¹⁸ It is inevitable that synthetic algorithmic intelligences can and will create things that we have not thought of in advance or ever intended to make, but as suggested, because they do not need our thinking or intention as their alibi, it is their inhumanity that may make them most creative. ¹⁹ Like Deleuze on the beach making sand piles, humans wrangle computation with our algorithm boxes, and in doing so, we make things by accident, sometimes little things like signal noise on the wire and sometimes big things like megastructures.

80

Earth Layer

1. Digestion

In the dynamic between natural computation expressing itself through artificial computing machines and those machines in turn remaking the world, each bends and countersimulates itself awkwardly and incompletely in the other. The Earth layer is shaped by this irregular and perhaps unsustainable interplay between the one and the other — sensing, drawing, enumerating, consuming, effecting. While it is a mistake to imagine computation as something that just sprang into existence with the rise of computing machines, or as the superimposition of some synthetic layer on top of everything organic and analogous and ultimately separate from natural algorithmic processes, it is also wrong to imagine computation as existing on a dry virtual plane sealed off from wet economies of energy, water, arbitrary valuation, remote capture, and geographic chance. So irrespective of the mathematical limits of algorithmic reason, The Stack is interested instead in the limited and sufficient compositional capacities of a megastructure already under construction, the thresholds of which are geological, sociological, economic, chemical, and geopolitical as much as they are calculative. This chapter draws on those limits and on the risks that come with positioning the Earth as a layer in a synthetic machine, and for this Earth is the Earth— a physical planet — not a metaphor for “ nature. ” There is no planetary-scale computation, now a vast network of many billions of little Turing machines, that does not intake and absorb the Earth ’ s chemistry in order to function. The Stack is a hungry machine, and while its curated population of algorithms may be all but massless, their processing of Earthly material is a physical event, and therefore the range of possible translations between information and mechanical appetites has another limit that is not mathematical but defined by the real finitude of substances that can force communication between both sides of this encounter. ²⁰ Furthermore, like any megamachine the Earth layer is as socially constrained as it is technologically configured, and so there are political economies of Turing machines that are only accessible through misaligned and uneven hierarchies of geography, energy, and programmability. ²¹

Earth Layer

81

This is made clear by unpacking and sifting through the hardware on which The Stack depends. Silicon is far from the only important substance required in its manufacturing and maintenance, and the economics of its assembly are far from crystalline. The Stack ’ s need for more exotic elements is intense, and even relatively mundane consumer electronics and Cloud tethers (aka “ phones ” ) contain dozens of different minerals and metals sourced from every continent. Some crucial metals are drawn largely from rich and vulnerable mines in central Africa. In the east of the Democratic Republic of Congo (DRC), for example, big chunks and little pebbles of tantalum (coltan), cassiterite (tin), wolframite (tungsten), and gold are pulled by hand from cold, sludgy mountain rivers, often by children, and eventually they make their way into the device component supply chain. ²² In 2009 a few mines here produced 13 percent of the world ’ s mined coltan, an inert metal used in ubiquitous tiny capacitors, especially for cell phones. ²³ From this same land, the Belgians took ivory, the Americans cobalt, and now billions of Earthlings everywhere carry little bits of Africa around with them in their pockets. The financial rewards of mining and trading in electronics have contributed to devastating effects in the region, including overlapping civil wars in the DRC and next door in Rwanda (from 1998 to 2003, upward of 5 million people died in the Congolese civil war, making it by one measure the deadliest conflict since World War II). ²⁴ Extraction and export of minerals, both legal and illegal, have been controlled and taxed by competing militias and organized crime; away from the relative stability of the cities, these groups continue to terrorize local populations and use the proceeds of this export trade to finance ongoing wars over local territorial positions. The smoldering conflict is a war partially financed with the manufacturing capital of smart phones and laptops; inevitably, the smooth skin of the device demands gore to feed its gloss. Deforestation in the pursuit of new sources of coltan in remote areas populated by gorillas has also led to an increase in the trade and consumption of bush meat, a quasi-cannibal economy that may also allegorize widespread war crimes in nearby villages. Exported minerals are sent to smelting companies, mostly in China, India, Thailand, and Malaysia, where they are mixed with metals sourced the world over (Australia and Brazil are the other major sources of coltan). More recently, the “ conflict minerals ” cause has taken its place among other Konyisms, but potentially effective legislation has been passed, including in the United States, that requires electronics manufactures to better police their own supply chains. ²⁵

82

Earth Layer

The most heinous circumstances are the most allegorically rich, but even absent the anarchic brutality of these wars and the Conradian odor of campaigns against them, the lesson is more global: there is no Stack without a vast immolation and involution of the Earth ’ s mineral cavities. The Stack terraforms the host planet by drinking and vomiting its elemental juices and spitting up mobile phones. After its short career as a little computing brick within a larger megamachine, its fate at the dying end of the electronics component life cycle is just as sad. What is called “ electronic waste ” inverts the process that pulls entropic reserves of metal and oil from the ground and given form, and instead partially disassembles them and reburies them, sometimes a continent away and sometimes right next door. Minerals originally sourced from the Congo might make their way to California via China, before being pulled by hand from a dead phone and burned or buried in Agbogbloshie, Ghana, or Lagos, Nigeria, two of the most active repositories, a short distance from their source. ²⁶

1. Geo-graphy and Geoaesthetics

As a transcontinental effect, this digestive cycle can also be seen as a sort of distributed composition, a discontiguous plastic-metallic mega-assemblage remade every day with little molecular bits floating in the light, across the ocean and our lives. To call these flows “ compositions ” is not to excuse their due ethical weight, but instead to remind ourselves that they are recomposable, and as such, that any sense of inevitability about how they are now arranged today is shortsighted. How unfamiliar could its flux and churn be from what it is now? At the radical end of that contingency, what is the ultimate recomposability of such materials? The answer may depend on how well we can collaborate with synthetic algorithmic intelligence to model the world differently — in other words, thinking takes place in relation to territory, and concepts not only represent the world but also make it according to their situation: diagramming, deforming, drawing and redrawing, and segmenting of the Earth back on itself. We internalize ideas from ground and sky according to whatever perspectives are available (Copernicus, Solaris, Hubble) until like the upside-down spacewalking astronauts, we are dislodged from this reference and given back to the void.

The Stack is not only on the Earth and built out of the Earth; as a composition, it is also a framing of the Earth, and so its geodesign works through its specific sorts of line-making and putting segments of the world in motion. Elizabeth Grosz develops a philosophical trope of demarcation as an elemental principle of animal world-making. She writes, “ The earth can be infinitely divided, territorialized, framed. … Framing is how chaos becomes territory. Framing is the means by which objects are delimited, qualities unleashed and art is made possible. ” ²⁷ Making enclosure by drawing a segment of the world into a presentation is elementary place-making; it is the gesture of geography. Schmitt would not disagree up to this point. The frame, however, is a peculiar sort of introduction of difference whereby the surface of things appears to fold in on itself. It captures and exhibits its subject curled back on itself by a delineation of figure and ground. Grosz links the act of framing, however, not to the subtractive competition of natural selection but to the multiplicative energy of sexual selection and its economies of display, expenditure, and abundance. In her Darwinian parable, the animal draws territory with its paw, its wing, or song refrain not only to fend off predators but also to stage itself in the view of a mate. She quotes from Deleuze and F é lix Guattari ’ s last collaboration: “ Every morning the Scenopoetes dentirostris, a bird of the Australian rain forests, cuts leaves, makes them fall to the ground, and turns them over so that the paler internal side contrasts with the earth. In this way it constructs a stage for itself like a ready-made; and directly above, on a creeper or branch, while fluffing its feathers beneath its beak to reveal their yellow roots, it sings a complex song made up from its own notes and, at intervals, those of other birds that it imitates; it is a complete artist. ” Beyond what is needed for survival in the moment, this act of self-framing sets the world in motion with a composition that motivates communion. The composability of the Earth is “ not linked to some intrinsic relation to one’s own body but exactly the opposite: it is linked to those processes of distancing and the production of a plane of composition that abstracts sensation from the body. ” Grosz writes that the frame “ is the particular contribution of architecture to the taming of the virtual, the territorialization of the uncontrollable forces of the Earth. It is the frame that ... liberates the qualities of objects or event that come to constitute the substance, the matter. ”

Earth Layer

83