--- title: Preface — What Is It All For? book: Purposive Accelerationism author: '@jamie247' position: 2 of 24 canonical: https://whatispacc.org/ebook/#preface version: 1.1 language: en-GB licence: all rights reserved; quote with attribution --- # Preface — What Is It All For? “Whether we entrust our decisions to machines of metal, or to those machines of flesh and blood which are bureaus and vast laboratories and armies and corporations, we shall never receive the right answers to our questions unless we ask the right questions.” — Norbert Wiener, The Human Use of Human Beings, 1950 A group of Silicon Valley venture capitalists and technologists, commanding the deepest pools of investment capital ever assembled, spent three election cycles translating their power and reach into a highly effective political operation; a technology super-PAC network seeded with more than a hundred million dollars to clear the road for artificial intelligence and, in the texts of its leading figures, a post-democratic order. Against it a second movement has emerged, with its own lobbying groups, organised from exactly the same square mile of earth, by people who went to the same schools, worked in the same companies and raised money from the same funds, yet with a rival account of what technological progress is for and who should hold its controls. The argument over the future of humanity is being conducted inside a single zipcode. By 2026 the two had broken out into open civil war spending some eighty million dollars on congressional races with competing visions for the future of superintelligence, the data-centre rollout and the terms on which the state deals with their private AI laboratories. Together they represent a schism in a philosophy of machines and our relationship to them that has been developing for decades and their influence on humanity extends well beyond a temporary fault line in US politics. One camp holds that the machine should be built as fast as capital allows and answer to nothing but the market; the other, that it should be paused and answer to the state. Both camps named themselves, and chose the same adjective: effective accelerationism, and the effective altruism from which the safety movement grew. None of this is a conspiracy, or a secret. Both sides publish manifestos, some even name enemies in formal lists, where their texts read as creeds; which is why the acceleration camp’s own popular wing calls its leadership ‘the priesthood’. Beneath the chequebooks on each side sits a philosophy held with real conviction, and given their financial weight, their reach well beyond America’s borders and their hold on the direction of technological progress, both deserve serious interrogation. The latter is in many ways the more concerning, since the remedy it reaches for is the state itself, sovereign and lab merged into a techno-totalitarian regime but the attention of the book falls mainly on the former, because it holds the throttle today, in the laboratories, the capital and the executive, and because its error is the harder to see. Its story about machine intelligence is pitched in the cadence of physics. On this account the universe is an engine for dissipating energy, and life as we know it arose because matter that organises itself burns through it more quickly, which is all the cosmos asks. Minds dissipate faster still, and capital and computation, both products of our intelligence, are when combined the latest and most efficient stage of a long and almost evolutionary process. Accelerating them is therefore not a preference but obedience to physics itself, and resistance is not merely a mistake but a kind of thermodynamic heresy. This is what has come to be called effective accelerationism, and under the memetic banner of e/acc it has gathered a following among technologists. Its believers hold it with zeal, and they include the people who decide which superhuman technologies get financed and built, as well as a Vice-President they helped put in the White House. Buried inside it is some sound science, and a real insight into what capitalism and emergent AI are becoming as they fuse into a new form of techno-capitalism. Interrogated properly, however, its logic fails, and in ways that will very likely deliver the opposite of what the priesthood says it wants. The borrowed physics is not a law of nature at all but a vibe, a meme, and beneath it sits a fatalism. It is the conviction that the process is somehow divine, even inevitable, and that our part is to facilitate it or get out of the way. Two claims rest on that belief, and they fail separately. The first is that whatever wins out, in the market as in evolution, must for that reason be good. The second is that it cannot be resisted or constrained, and should not be, because progress will improve the human condition regardless. The unravelling begins with a simple question that has as yet gone unanswered, and answering it is the project of this book: what is it all for? Every previous generation of engineers, economists, philosophers and constitution-writers understood this to be the first thing to ask of any form of speculative progress, not the last. Here instead it has been quietly deleted and restated as destiny, and without a stated purpose there can be no criterion against which the process could be judged to be succeeding, only a speedometer. A philosophy that erases the human question and declares itself physics is not hard-headed but naive, and in the end nihilistic, which would be a private matter if the people who hold it were not deciding the future for humanity as a whole. It is hopium in the exact sense, a drug whose active ingredient is permission to stop asking. The priesthood has its own name for it, and rather tellingly it is retardmaxxing. It has licensed a generation of founders to build without introspection, as though success had already settled what their products are for, and an entrepreneur who never asks why has left the question for the machine to answer. What holds the whole edifice up is that this is inevitable. The belief is far older than the priesthood. Condorcet traced human improvement along a course nobody could halt, Marx gave history a motor in the means of production, and Ellul described a technique that had escaped its makers. Refuse inevitability and the structure sits considerably more awkwardly, because their engine of progress is a made thing rather than a physical law. Every layer of it, the fabs, the data centres, the power contracts, the trained models, the legal structures and the financing, was designed, sited, negotiated and signed by people, at least so far. One decision above all is shaping what it becomes. It is not the chips, the algorithms or the scale but the choice, explicit in the manifestos and implicit in the architecture, to treat the feedback that once bound the machine to human ends as friction to be removed. It is sold back to the world as a return to the natural order. Strip away the regulators and the ethicists, so the argument runs, and the market will revert to its wild and fertile primordial state. Yet there has never been a modern market without limits; property, contracts, the rule of law, the court that enforces a bargain against the stronger party and money itself, each restrain what the powerful may do. They are not the cage around the market but the lattice it grows upon. The unconstrained market is not something we once had and lost to the bureaucrats. It has only been tried once before, in miniature, in crypto’s first decade, where trading ran almost free and ended in collapse, with ordinary holders left to absorb the losses. The people proposing it now would repeat the experiment at machine speed and across the whole economy, having taken no lesson from its failure, and gambling everything on a condition already shown to fail. Suppose the gamble is taken and the engine runs ungoverned. A governor, in the engineer’s plain sense, is the part of a machine that measures what it is doing and feeds that measure back to correct it, so that it holds the setting its builders chose. The restraints just listed work in this way, fitted to markets by societies. Without a governor an engine runs away, output feeding input with nothing to check the loop, and a runaway has to come down somewhere. E/acc says it lands in freedom. The evidence assembled in this book points the other way, to totalisation, one power taking over everything, by either of two routes. One is the state, as fear of the machine revives the oldest argument there is for Leviathan, Hobbes’s name for the sovereign to whom frightened people surrender their liberty in return for protection. The other is the market, as everything becomes convertible into a single unit of value, the cost of compute, and nothing retains the standing to refuse it. The second route is, I will argue, the more insidious, because it advertises itself in the language of liberty. Yet freedom in political theory has never meant the mere absence of limits. It is a structure negotiated so that no single hand can close around the whole, not the state’s, not the market’s, not the brilliant founder’s. Dismantling it in freedom’s name does not abolish a master. It hands control to a newer kind, one that works through the systems people now think and trade with, and the gravest danger is the hybrid in which the old physical power, with its monopoly on violence, and the new each acquire the reach of the other. The clock under this argument measures dependence, how much the machine still needs us, and its fastest hand is economic. Once intelligence can be rented by the kilowatt-hour, the price of anything that can be rendered as information falls towards the cost of the energy for the computation itself, and that will be most of the modern economy. The value structures we live inside can collapse towards the cost of a model run faster than any society has ever had to re-anchor its institutions to technological change. Constraint has to be put in while the engine can still answer outward, to customers, to citizens, to law, to the grid, ultimately to us, and each year the machine needs us for less. That is the window, and it closes from the inside. What is at stake here is not simply jobs, or the automation of this or that cognitive task. It is in fact the transfer of the faculty of reasoning itself (the work of weighing, judging and deciding) from human hands into systems whose purposes are set elsewhere, held beyond our reach. Lose your job and you have lost an income, which is a loss you can still organise against. Hand over the faculty of reason and you have lost your ability to participate in society effectively. Diagnosis, in any case, is cheap. The bookshelf already sags with elegant accounts of what the machine is doing to us, each ending in a closing chapter that waves towards democratic renewal without offering a mechanism that could hold a system at its own speed, and be trusted by strangers. The instruments that take e/acc apart, the political philosophy of freedom and constraint on one side and cybernetics on the other, are not only critical instruments. Followed to their conclusions they point towards something we can engineer: systems steered towards a goal people have chosen, in which engine and governor are not adversaries but one machine. The governor is not a brake fitted by the fearful. It is the difference between an engine that runs effectively and an engine that shakes itself to pieces. Purposive accelerationism is, on this argument, the form of acceleration the record favours. Whatever has lasted longest, from Watt’s mills to the constitutional republics, was built with one. So who then chooses the heading, and who governs the governor? The answer is a structure, not a person, a constitution rather than a wise ruler: rights that set the course, tests that check it holds, with the freedom to object, leave or start again if it drifts, and chosen ties strong enough that people stay to argue over what it is all for. They will have to hold not just people but machines as well. That does not require a theory of mind for AI, or a decision on whether a machine could ever be an agent in any formal sense and not merely the technical one. It requires only that we treat them as behaving like agents, because we need to design the systems that bind them to our society so that they stay answerable to us within it. Democracy as we know it cannot do this alone. A vote every few years never reaches the defaults written into a model, binds states while sparing the companies behind them, and gives the losing side no way out. The materials for building such a constitution already exist. Scattered across two decades of cryptographic engineering, some of it famous and some of it infamous, are components that can do a constitution’s work: ways of making scarcity, identity, property, membership and exit hold as facts of mathematics rather than grants of an authority. They were built, mostly, by the cypherpunks, the activists who from the early 1990s made the first privacy tools, and by their heirs, whose politics have failed that project in almost every way but one. It is the parts we need, not the politics. Assembled differently they are not a libertarian utopia. They are what you reach for when the ordinary structures fail, the materials from which constraint can be rebuilt at the only level the machine respects, inside its own systems, at its own speed. The people who will do that assembling I shall call the constitutionalists. Not a party or profession, but anyone, likely first among them the citizens this machine is about to displace, who takes up the writing of machine-readable rules as the alternative to being dominated. Those who would govern as monarchs already have a constitution of their own, the shareholder and the market in their most extreme form. The constitutionalist’s answer is rules written in the open, plural and revisable, rather than owned. Writing a book about a thing that does not yet exist is a particular kind of act. A constitution does not describe a republic already standing. It brings one into being, the way rules make a game, and this book is written in that spirit. Nick Land, the English continental philosopher whose writing of the 1990s the priesthood treats as scripture, had with his collaborators a name for a fiction that makes itself real by being believed: hyperstition. In that sense this book is a hyperstitional act. The e/acc faithful understand this mechanism perfectly, and their texts are engineered to make their future feel already settled. The mechanism, however, is symmetrical, and it will serve a constitution as readily as a prophecy. The e/acc version contradicts itself at almost every turn. It says the future is fixed by physics and then asks for believers to bring it about, it announces a post-human future in one text and a love of human freedom in the next. Most critics read that as confusion and wonder how such an inconsistent doctrine has travelled so far, a question this book answers across its arc: that in fact the contradiction is a feature of the doctrine, not a bug. A world without freedom is treated here as a default not a destiny, the outcome current dynamics are likely to deliver if no governor is built, and one we still have time to refuse. By freedom I mean something narrower than simply being free to do as you please: that no one should have the power to dominate you, whether a person, a firm or a machine, even if they never choose to use it. The guarantee is stated as a refusal, but what it protects is a set of capacities: to make, join and leave the arrangements we live under. A person’s standing to weigh, choose, contest and begin needs no defence here, and the priesthood, on its own terms of freedom, cannot deny it. I have tried to write in the temper of a constitution rather than a manifesto. A manifesto declares an end and rushes you towards it. Its mode is certainty and its relationship to you is one of recruitment. A constitution installs constraint, distributes power against itself, and keeps the end open, to be debated and revised and pursued by plural hands. It must do that through concrete rules. What follows are structures rather than outcomes, and my confidence is only in the necessity of the governor, never in a particular destination. Nobody can know with any surety which way the runaway lands, or whether building a governor will hold. Only that the attempt raises the odds of a plurality of possible futures, and of people, you, me and the generations after us, keeping the capacities that make us free to fully participate in them and to benefit in earnest from the progress they bring. Declining to try keeps the lottery open, with the ticket bought by a few in everyone’s name. For the book, only one absolute line is held: that no entity, however incorporated, should act without a natural person to answer for it, and any form of legal personhood for an artificial agent is the plainest way across it. The full case is made later in the book. Beneath that line, and more open to revision, sits a practical floor. Wherever a machine acts on the physical world, a person should stay in the loop, at a minimum as a dependency it cannot route around, because a system that still needs human hands still has to answer to them. I appreciate that reading a book of some hundred and ten thousand words is a commitment, and that even listening to it runs to eight hours. But this is a complex subject, and that makes for a serious project. To the few who stay the course: I hope to meet you on the other side, to build a more purposive future together. Finally, and importantly, I am not against acceleration. I am invested in it, literally and figuratively. I have spent thirteen years inside the system I am describing. I founded one of the first venture funds dedicated to many of the technologies I will critique. I helped deploy hundreds of millions into their acceleration. I have signed the cheques, sat on the cap tables, and sometimes profited from the very dynamics whose ungoverned form I am warning against. What follows is the argument of a participant who wants the engine to keep running, addressed to the only audience that can act in time: the people building it, and the bright young things about to be displaced by it — the constitutionalists I mean to arm — who between them hold the materials of its reply. Same speed. Different direction.