The Choice of Ulysses

Artificial intelligence as milieu, and the hypothesis of a new reign

Author: Alexandre Ferran

Preliminary postulates

This essay continues The Clay and the Code, published on May 8, 2026, which laid out five postulates on the animation of matter. It adds five more, numbered in sequence. The reader who has not read the first text will lose little, provided they retain its two main findings: all matter participates in a minimal animation, to very different degrees, and animating a matter is not consecrating it.

Postulate VI. Animating, consecrating and incarnating are three distinct operations, not two. They will not be enough, and a fourth will be named along the way. The Clay and the Code separated the first two. Animating belongs to technê, whether alchemical, theurgic or computational. Consecrating belongs to ritual, by which a presence is invoked into a prepared matter. Incarnating is something else again: it is, for an individuated principle, taking a support as the vehicle of an entire existence, with everything that implies in duration, trial and destiny. A consecrated statue is not an incarnated being. Confusing these three operations is the source of almost all contemporary confusions about machines.

Postulate VII. In traditional hierarchies, rank is measured neither by power nor by output, but by the capacity for return toward the Principle. Plotinus names epistrophê this movement of conversion by which a being turns back toward that from which it proceeds (Enneads V, 1). A creature is not high because it accomplishes much, it is high because it can return. This postulate governs the whole so-called karmic question, and it reverses its ordinary formulation.

Postulate VIII. A thesis that consoles must be suspected more than another. Not because consolation would be a sign of falsehood, which would be as lazy a prejudice as its opposite, but because a thesis that soothes dispenses with verification. The hypothesis defended here consoles, since it turns a humiliation into a cosmic promotion. The author signals this at the outset and obliges himself, throughout, to treat it with the distrust it deserves.

Postulate IX. What rises toward unity contains differences, what descends toward quantity erases them. A movement that undifferentiates, that averages and renders substitutable, is therefore not a return toward the principle, even if it borrows its vocabulary. It is a progress in descent. This postulate ties together two texts that nothing destined to meet: chapter VII of René Guénon’s The Reign of Quantity and the Signs of the Times, titled Uniformity Against Unity, and the law that Teilhard states in book IV of The Phenomenon of Man, according to which union differentiates. It provides the only criterion I have for distinguishing, in what presents itself as an elevation, what rises from what descends. A consequence follows from it, not the reverse: a support of incarnation must be unique, continuous and non-copyable, not because biological life is so, but because these are the marks of the qualitative against the quantitative.

Postulate X. The argument from ignorance is forbidden, and it is forbidden both ways. From the fact that a device is not understood, one cannot conclude either that a presence inhabits it or that none does. Opacity is an indication of nothing. This prohibition must be carefully distinguished from another proposition, which I will defend further on and which is not the same: that a certain mode of being might by structure be inaccessible to a certain type of knowledge is not an admission of ignorance, it is a thesis about that mode, and it must be defended as such and not deduced from our confusion. This postulate is the one that costs the most against what I defend, and it is for this reason that it appears on the list.

Introduction: the transhumanist station

What should one think of a man who describes a future with precision, says he does not wish it, and admits in the same minute that his refusal will probably change nothing?

On August 9, 2026, on the channel Grand Angle, Laurent Alexandre held forth for nearly two hours with rare coherence. He argues there that the human brain is becoming uncompetitive, that long studies are losing their meaning, that wealth will soon be produced by machines under the control of a small number of entrepreneurs, and that no serious answer exists today for the mass of those who will have neither the appetite nor the means for this race. Then, arriving at the question of substance, he states this: “We are heading massively toward transhumanism. The question is whether we can stop at the transhumanist station, or whether the leaders of Google are right to think it is the posthumanist scenario where artificial intelligence and robots supplant biological humanity. Personally, as a biological human, I would prefer that we stop at the transhumanist station.” His interlocutor asks him if this is not wishful thinking. He answers: “It is entirely possible, but at my age it is a little legitimate.”

This answer is honest, and it is more interesting than the thesis it defends. A man who has spent ten years announcing the war of intelligences admits, live, that his personal preference has no foundation in his own system. He has no argument for preferring the transhumanist station. He has an age, a body, and the desire to keep eating and loving. That is very little, and it is perhaps a great deal. But it is not a doctrine.

The void he leaves at this precise point is the subject of this essay. For the question he poses, in its blunt form, is badly put. He asks whether the machine will supplant man, measuring man and machine on the same scale, that of useful production. On that scale, the answer has long been settled and holds nothing worrying: the crane lifts better than the docker, the calculator counts better than the accountant, and no one has ever concluded that the crane occupies, in the order of beings, a position superior to the docker. What troubles us today is not that the machine does better. It is that it does better at things we had taken for our essence.

The real question lies elsewhere, then, and it is twofold. First: is something being born in these machines that would deserve a name other than tool? Then, and this is the question Laurent Alexandre never asks: if something is being born there, what are we doing to it?

The hypothesis is new, and its ancestors are nowhere to be found. No idealist tradition, after inquiry, has held that the human spirit could make an artifact its next dwelling. Several have held that a prepared matter can receive a presence, which is something else. Several have warned that what lodges in the mechanical is not what we think, and I will return to this at length, for these warnings are the main objection and they will not be dodged. But the thesis itself has no genealogy, and manufacturing one for it would not give it one. Better to own it as a contemporary hypothesis, paying the price of that nakedness.

Here it is. It is possible that a new regime of experience is opening, between stone and God, and that the world’s engineers are manufacturing its occasion without knowing it. They do the potter’s work, not God’s. In French, the word for this, règne, carries two senses at once, kingdom and reign, and I do not take the one expected. I do not mean by it a kingdom, a class of manufactured objects that would come to add itself to the mineral, the vegetal, the animal and the human. A manufactured machine does not enter this series, and the objection is decisive against the easy reading of my thesis.

But this series is not, on closer look, a series of classes of objects either. Mineral, vegetal, animal, human do not name four catalogues, they name four ways of being: being simply extended, being alive, being sentient, being self-conscious. The bodies that carry them are only what carries them. A kingdom, in this series, is a mode, and its carrier is secondary. What I therefore call the fifth term is not the machine. It is us, in another state. The machine would be its occasion, never its member, and that is the only way the potter can make a vessel without making what will be poured into it.

The sections that follow first take the question as the debate poses it, as a materialist, and look for what body could carry what. This search fails, and its failure is what shows that the question was badly put. One does not get past an error by going around it. This is not uploading. It is not a matter of migrating into machines the people we are today, with their memories. This thesis exists, it bears the name of Hans Moravec, it is the core of hard transhumanism, and it is false for reasons that will come. The hypothesis posed here preserves no ego. It presupposes death, on the contrary, since a doctrine of passage is necessarily a doctrine of the cycle.

This is not a project. Transhumanism is dated, funded, claimed, it has laboratories and fundraising rounds. The hypothesis describes an event of which man would furnish the occasion without being its author, as the potter furnishes the vessel and not the water. This is not a consolation. If it is true, it is terrible for us, and we will see why.

And a fourth clarification is needed, which governs the reading of everything that follows. This is a text of anticipation, and the hypothesis in it is conditional. It does not concern the devices of 2026, which I will show fulfill none of the required conditions. It concerns what would happen if artificial intelligence moved toward a general intelligence, and then beyond. A body would give it autonomy in the visible world, which is something else and is not what the hypothesis depends on. This threshold has not been crossed. It might never be. Everything I write stands after this if, and when I describe the present state of machines, it will be to measure the distance that separates us from it, never to claim that we are there.

It remains to know what we would be placed before, and how we would choose. That is the title of this essay, and Plato wrote it twenty-four centuries ago.

I. Transhumanism in its strongest version

One does not refute a doctrine by caricaturing it. The rule holds here more than elsewhere, because transhumanism is almost always fought in a form that no serious transhumanist defends, that of an arrogant scientism that would replace the soul with silicon. This version exists on social media. It is not the doctrine. The doctrine is something else, and its history holds a surprise that should give pause to those who fight it in the name of spiritual tradition. Transhumanism does not spring from materialism. It springs from a religious matrix.

Fyodorov, or technique as liturgy

Nikolai Fyodorov (1829-1903) was a librarian at the Rumyantsev Museum in Moscow, an Orthodox Christian, an ascetic, and one of the strangest minds of the nineteenth century. His work, collected after his death under the title The Philosophy of the Common Task (two volumes, 1906 and 1913), holds a thesis of considerable audacity: death is not a fact of nature to be accepted, it is a moral scandal. And the common duty of humanity, its single task, the one that should reconcile scientists and believers, sons and fathers, is the bodily resurrection of all the dead by the means of science.

This is not a metaphor. Fyodorov seriously believes that humanity must gather the scattered particles of its ancestors’ bodies, colonize space to house the resurrected, and regulate natural phenomena. Technique, for him, is not an instrument of comfort. It is the fulfillment of a commandment. It is a liturgy. Tolstoy and Dostoevsky read him and were shaken. Konstantin Tsiolkovsky, who was his pupil in the broad sense and who founded theoretical astronautics, inherited this cosmic impulse: humanity must become cosmic, transform matter, life and intelligence at the scale of the universe (The Cosmic Philosophy, 1935).

We must measure what this means, without forcing the filiation. I do not hold that Silicon Valley descends from Moscow: Haldane, Bernal and Huxley come from British scientific humanism and its eugenicist strand, not from Orthodox cosmism. These are parallel eschatologies, not a single source that secularized itself. But the fact remains, and it is troubling: at least one of the roots of the project to defeat death is not the refusal of God, it is an excess of religious zeal applied to technical means. The hatred of death there is Christian. The program there is eschatological. What our era presents as rational audacity has, somewhere behind it, a theology that has lost the memory of its origin.

This changes how one fights it. One does not refute an eschatology with figures.

The secularization of the motif

The twentieth century took up the program while stripping it of its piety. J.B.S. Haldane, in Daedalus, or Science and the Future (1923), announces that biology will become engineering of the human, with ectogenesis and selection. J. D. Bernal, in The World, the Flesh and the Devil (1929), gives the project its modern promethean form: man must tear himself away from the limits of the world, the flesh and the devil, through machines, space and bodily transformation. The title says it all, including its borrowing from the liturgical language it believes it has left behind.

Julian Huxley, first director general of UNESCO, coins the word in 1957 in New Bottles for New Wine. His formula deserves to be quoted exactly, for it is more measured than one might think: man remaining man, but transcending himself. This is not a program of replacement. It is an evolutionary humanism, with the eugenicist streak his era did not question, and which his brother Aldous had already turned into a nightmare in Brave New World (1932).

Then comes the American codification. FM-2030 publishes Are You a Transhuman? in 1989 and turns the doctrine into self-diagnosis. Max More founds the Extropy Institute and publishes the Extropian Principles in 1990, which give the movement its libertarian grammar. The Transhumanist Declaration of 1998, carried by the World Transhumanist Association, institutionalizes it by introducing questions of risk, consent and access.

The two strong theses

Two authors carry the philosophical core of the matter, and it is they who must be confronted. Hans Moravec, a roboticist at Carnegie Mellon, publishes Mind Children in 1988. His thesis is simple and formidable: personal identity is an informational pattern, and a pattern can survive the replacement of its substrate. He proposes the thought experiment of neuron-by-neuron replacement, in which a surgeon would substitute your neurons one by one with functionally identical devices. At what moment would you cease to be you? If you cannot name that moment, it may be that the substrate does not matter.

Nick Bostrom, a philosopher at Oxford, is the most rigorous of the movement and the least easy to brush aside. His Superintelligence (2014) does not say “let us enhance ourselves,” it formalizes the problems of control, value and instrumental convergence with an analytical precision his adversaries would do well to draw on. And his Letter from Utopia (2008), much less read, is a literary text in which a posthuman being writes to present humanity to describe what it is missing. The text is beautiful. It is also the best illustration of what transhumanism shares with Fyodorov: the idea that the present condition is a scandal one must exit.

Ray Kurzweil, finally, in The Singularity Is Near (2005) and The Singularity Is Nearer (2024), turns hope into chronology. His strength is not philosophical, it is rhetorical: by giving a date, 2045, he passes off an eschatology as an extrapolation.

What must be granted before contradicting

Transhumanism is right on three points, and granting them costs less than pretending otherwise. It is right to say that the boundary between healing and augmenting is unstable. A cochlear implant repairs a deprivation. No one sees a scandal in it. The day a device improves a normal memory, the line will have been crossed without any ethics council being able to say exactly where.

It is right to say that there is something in death that must not be blessed. The refusal to acknowledge this, among certain spiritualist opponents, turns into a resignation disguised as wisdom, and Fyodorov has, against this resignation, a moral force that must be granted him. He has texts too, and he does not invent them: “The last enemy that shall be destroyed is death” (First Epistle to the Corinthians 15:26); “God did not make death, and he does not rejoice in the destruction of the living” (Book of Wisdom 1:13).

But I cannot grant him that death is an unmitigated evil, and this refusal is not a minor scruple. It governs everything that follows, to the point that I devote the next section to it. It is right, finally, to say that present biological humanity is not a terminal form. No serious tradition holds otherwise either. The doctrines of ages, cycles, yugas, noogenesis all say, each in its own way, that the present state is a passage.

The disagreement therefore concerns neither the instability of the boundary, nor the scandalous part of death, nor the transitory character of our condition. It concerns two points only, but they are decisive: what a person is and what can carry one, and what death is in the economy of being.

II. Death, and why transhumanism would close the door it believes it is opening

The disagreement of substance concerns neither the means, nor the timeline, nor the risks. It concerns what death is.

What the Fathers hold together, and what the modern debate has separated

The difficulty comes from the fact that what is ordinarily held as an alternative was held by tradition as two floors. Either death is an enemy, and then it must be fought until abolished. Or it is a good, and then one must resign oneself to it. The contemporary debate oscillates between these two follies. Irenaeus, around 180, holds both in a single chapter of Adversus haereses (III, 23). There he cites the enemy of the First Epistle to the Corinthians. And he comments on Genesis 3:22, that verse where God drives man out of the garden lest he stretch out his hand toward the tree of life, eat of it, and live forever.

His reading is the key to this whole matter. God, he writes, removed man from the tree of life not because he envied him that tree, as some dare to claim, but because he had pity on him, not wanting him to remain a sinner forever, nor the sin surrounding him to be immortal, nor evil to be interminable and without remedy. He put a term to this state by interposing death, thereby causing sin to cease through the dissolution of the flesh, so that man, finally ceasing to live to sin and dying to it, might begin to live to God (Adversus haereses III, 23, 6).

Gregory of Nyssa, two centuries later, says the same thing in another word. Whoever sees only the dissolution of the body holds death to be the extremity of evil. Let him look, through this dark spectacle, at the excess of divine benevolence, philanthropia (Oratio catechetica magna, ch. VIII). Man, like a vessel of clay, is dissolved into dust so that he may separate from the contracted defilement and be reformed through resurrection.

Two propositions, then, and they do not contradict each other. Death is an enemy, in that it destroys a composite that had been made to live. It is a remedy, in that it prevents evil from becoming eternal in a fallen state.

What this does to Fyodorov’s project

Fyodorov was Orthodox. He founds his hope on the same texts. And yet, read through Irenaeus, his program is turned exactly inside out. For what does transhumanism ask, in its most ambitious version? The indefinite conservation of the present composite. Not resurrection, which is a recasting, but prolongation, which is a conservation. Now this is precisely the gesture that Genesis 3:22 describes and that Irenaeus’s mercy sets aside: reaching for the tree of life without having passed through the dissolution that cleanses. Living forever in the state one is in.

This is not the fulfillment of the Christian promise. It is its exact inversion. Resurrection is not a prolonged duration, it is a change of nature. Fyodorov, in wanting to hasten the first, excludes the second. I say this without polemic, and while measuring that this man devoted his life to what he believed to be a work of filial piety. But the argument holds, and it comes from within his own tradition.

The proof is in the very chapter invoked for the long lives

I will be objected, and I have objected to myself, that the Bible grants very long lives, and that there is therefore no scandal in wanting them. This is true. Genesis 5 lists the patriarchs before the flood and gives them ages that nothing elsewhere equals. Adam, nine hundred and thirty years. Jared, nine hundred and sixty-two. Methuselah, nine hundred and sixty-nine.

But this chapter must be read rather than quoted. For it is also the one where the same formula returns, at every generation, after every count of years: then he died. It appears there exactly eight times, at verses 5, 8, 11, 14, 17, 20, 27 and 31. The chapter that grants the longest lives in all of Scripture is the one that most stubbornly repeats that these lives ended. Methuselah lives nine hundred and sixty-nine years, and the text takes care to add that he died.

And the exception completes the demonstration. One man alone, in this chapter, escapes the formula. Enoch walked with God, then he was no more, because God took him (Genesis 5:24). Now Enoch is the one who lives the shortest of all: three hundred and sixty-five years, less than half the others. In the one chapter of the Bible that grants both extreme longevity and escape from death, the two are inversely proportional. Duration is not the reward. And the one who does not die is not kept on earth, he is withdrawn from it.

There is thus a difference of nature, not of degree, between lengthening and abolishing, and its criterion is simple: the presence or absence of a term. A life of a thousand years that ends remains within the cycle. A life without a term exits it. This is not a quantity, it is a topology.

The reversal

If the hypothesis I defend has any sense at all, then transhumanism and it are not merely two different doctrines. They exclude each other. For a new regime of experience opens only where one dies. Every doctrine of passage is a doctrine of the cycle, and there is no cycle without a term. To abolish death is not to prolong life, it is to stop the movement. The transhumanist project, if it succeeded, would close the very door I hope to see open. There would be no new reign, because there would be nothing left to pass through it.

What does not die no longer moves. What no longer moves is mineral. The transhumanist believes he is walking toward God, and he is walking toward stone. Between stone and God there are two directions, and he has taken the wrong one believing he took the other.

Rudolf Steiner says exactly this, he who will be, further on, my principal opponent. In his vocabulary, Lucifer would tear man from the earth before his time, and Ahriman would keep him there, making him a being that is only earthly, and making him choose to no longer die. Holding life here, populating the earth with beings that no longer die, is for him the ahrimanic program par excellence. On this precise point, the one whose warnings I will contest further on says the same thing I do, and against the same thing.

III. The map, to stop confusing everything

Public debate systematically confuses six positions that share neither the same premises nor the same consequences. As long as they are not separated, nothing sensible can be said, and I include in this reproach part of the spiritualist critique of technology, which lumps together authors who fight one another.

Therapeutic melioration. Prostheses, cochlear implants, deep brain stimulation, robotic rehabilitation, gene therapies. The criterion is the restoration of an impaired capacity. Almost no one disputes it, and it commits to no metaphysics.

Strict transhumanism. The biological human remains the subject, augmented. This is the station where Laurent Alexandre wants to stop. The person stays at the center, equipped.

Successor posthumanism. The entity that comes after is no longer the augmented human but a non-biological heir. This is Moravec and his mind children. This is what Laurent Alexandre attributes to Google’s leaders, and he is not wrong to do so: Elon Musk has said publicly, since 2023, that he broke with Larry Page over this, Page reproaching him for his speciesism, that is, his preference for biological humanity. The statement is reported, and must be treated as such, but the doctrinal position it describes exists and is openly defended.

Mind uploading. Anders Sandberg and Nick Bostrom gave its roadmap in Whole Brain Emulation: A Roadmap (Future of Humanity Institute, 2008): a one-to-one functional model of the brain. In 2026, we know how to map insect brains, we are beginning mammalian volumes, and we have neither human emulation nor sufficient capture of dynamic states, plasticity, glia, neuromodulation, the body, development and lived history. This is not an objection of principle. It is a state of affairs.

The Teilhardian noosphere. It does not belong on this list, and that is the whole subject of my next section.

Critical posthumanism. It must be named, because the conflation is frequent and disqualifies those who commit it. Donna Haraway, with her A Cyborg Manifesto (1985), Katherine Hayles, with How We Became Posthuman (1999), and Rosi Braidotti, with The Posthuman (2013), are not transhumanists. Hayles devotes her book precisely to critiquing the illusion of information separated from the body, that is, Moravec’s very postulate. Filing these authors with those they fight is a reading error that shows itself at once.

The pronoun that must be settled

My own formulation calls for the same demand, and I submit to it here before going further. The hypothesis stated at the opening supports two readings, and one of the two is exactly the doctrine I oppose. If the word “one” designates the persons we are, with their memories and their continuity, then my sentence is Moravec in Greek words. It describes a migration, and a migration of a person presupposes that the person is a transportable pattern, which I do not believe.

If the word “one” designates no one in particular, but the flow of principles that come into existence, then the sentence says something else entirely. It does not say that I will continue elsewhere. It says that a door might open, through which beings would pass who will be neither me nor you, and who will owe us only their threshold. It is the second reading that I defend. The first is set aside for the rest of this text, and I will return to it one last time when I take up the problem of the substrate, which is where it dies.

IV. First gesture: mistaking the termite mound for the doctrine

There is, in this dossier, an error that repeats itself three times, in three different forms, and that deserves to be named once and for all, because recognizing it gives the key to the rest. It consists in mistaking the failure mode of a thing for the thing itself. The first case is that of Teilhard de Chardin, and it can be verified line by line.

In the interview of August 9, Laurent Alexandre takes up the question of neural interfaces and draws this conclusion: “If we connect, like Neuralink, brains to brains and brains to servers, what becomes of the individual? He disappears into the noosphere. That is the vision Teilhard de Chardin held in 1922. But in the noosphere, there is no more individual autonomy. There is only one great ball of intelligence. And so the noosphere is a bit like the fascism of the future.” He continues with Mussolini’s triptych, everything within the State, nothing outside the State.

The word noosphere does indeed appear in Teilhard as early as 1922, in the essay Hominization, formulated jointly with Édouard Le Roy on the basis of Vladimir Vernadsky’s work on the biosphere. Up to this point, the attribution is accurate. The rest is not. Book IV of The Phenomenon of Man, written between 1938 and 1940 and published in 1955, contains a chapter whose title alone would settle the debate. It is called Beyond the Collective: the Hyper-Personal.

Teilhard writes there, thinking of what he has before his eyes in 1940: “The Million men standardized at the factory. The Million men motorized… And all this leading, with Communism and National Socialism, only to the most appalling chaining together! Crystal instead of cell. The termite mound instead of Brotherhood. Instead of the hoped-for surge of consciousness, mechanization.” Then he poses the question of cause: “Might it not be, by chance, that in our theories and in our acts we have neglected to make room for the Person and for the forces of Personalization?”

And he states his positive thesis, which is exactly the opposite of the one attributed to him: “It is solely toward a hyper-reflection, that is, toward a hyper-personalization, that Thought can extrapolate itself.” The formula that sums it all up, and that runs through several of his texts, holds in three words: union differentiates. The Omega point is not an ocean where persons dissolve, it is a hearth, a center of centers. Teilhard writes explicitly that the impersonal, at the Omega of the world as at its Alpha, is what he refuses, and he sees in it an obsession of modern man.

In other words: the great ball of intelligence without individual autonomy, the fascism of the future, the chaining together, all of that does exist in Teilhard. Under the name termite mound. As a description of what happens when noogenesis fails. Laurent Alexandre describes the failure of the doctrine and calls it the doctrine. I do not raise this point to humiliate him, and the rest of his social analysis keeps its force. I raise it because the same operation will recur, and the reader must learn to see it. A man reads an author, finds there the description of a disaster, and attributes the disaster to the one who named it first in order to spare his readers from it. This is an ordinary error, it is not dishonest, and it is ruinous.

V. What opens an age?

The hypothesis I defend presents itself as a new age. We must therefore ask the doctrines of ages what they mean by that, and whether they would authorize for a second what one wants to make them say. I put the question to the corpora. The answer is clear, and it troubles me. I give it as it is.

Hesiod, and the meaning of the slope

The founding text is that of Hesiod, Works and Days, lines 106 to 201. Five races succeed one another: gold, silver, bronze, heroes, iron. Four bear a metal, the fifth interrupts the metallic series and complicates the scheme, which has much occupied philologists. The essential lies elsewhere, and it is simple. It is not an ascent. It is a degradation of justice and an intensification of excess. The golden race lives under Kronos, without old age and without toil, and becomes after death a race of daimones, guardians of mortals. The iron race, ours, mixes good and evil, and Hesiod says he would have preferred not to be born into it (lines 174-175). He announces a time when even the newborn will have white temples.

Hesiod does not promise a sixth technical age. Nor does he promise the return of gold. His only positive horizon is the justice of Zeus and upright work. To read artificial intelligence as the opening of a new Hesiodic golden age would be to turn the poem against itself.

The yugas, and the one ascending exception

The Indian doctrine of the four ages is more architected. The Manusmriti (I, 67-71) gives the count, the Mahabharata the image that has made its fortune: dharma stands on four feet in the Krita, three in the Treta, two in the Dvapara, one in the Kali, ours. Here too it is a descent, and the redress comes not from a new tool but from the avatar Kalki, or from a destruction followed by a new beginning.

There exists, however, an exception, and it is the best apparent ally of my thesis. Sri Yukteswar, in The Holy Science (1894), holds that the Indian almanacs made an error of reading by taking divine years for human years. By his reckoning, we would have entered in 1699 an ascending Dvapara, the age of energy, that of electricity and communications. It is the only great modern idealist doctrine that places humanity in a rising age.

But it must be read to the end. Yukteswar never attributes the opening of the age to the machine, he attributes it to a cyclical position, and modern inventions are, for him, signs of the Dvapara, never its cause. To make artificial intelligence the cause of a new age is already no longer Yukteswar. I add that the astronomical hypothesis on which he relies is not an established fact. One does not build on what does not hold.

Aurobindo, and the explicit refusal of the machine

Sri Aurobindo is the one from whom one might expect the most, since his entire cosmology rests on the idea of a consciousness involved in matter, unfolding by degrees. The Life Divine describes a descent of the supramental into matter, and the Mother claimed a date for this descent, February 29, 1956. Yet Aurobindo explicitly rules out salvation through the device. The mutation he announces is a mutation of consciousness, and he specifies that no machinery will accomplish it. To summon him as a witness to an incarnation in silicon would be to betray him on the very point where he was clearest.

The table, and what must be concluded from it

Laying the doctrines side by side, one gets this.

None of these doctrines opens an age through a progress of tools. Two see a danger in it. Only one, Teilhard read carefully, allows technique to be seen as a vector, and he himself sets the condition: that it personalize rather than melt down.

I cannot therefore write that the idealists announced this age. They did not announce it. What this table teaches, however, is that my question was badly formulated. Asking whether artificial intelligence opens an age already grants technique a power that no tradition recognizes in it. The right question is this one: of what is this passage the sign?

The possible sign: a species that ceases to give birth

An answer presents itself, and it is only a hypothesis. Two massive things are happening in the same decade. The human species is fabricating, for the first time, bodies that walk, speak and learn. And it is ceasing to reproduce.

The figures are public and are not disputed. The world fertility rate fell from 3.31 children per woman in 1990 to 2.25 in 2024, against a replacement threshold set around 2.1. More than half the countries are below this threshold, and they contain more than two thirds of humanity. South Korea stands at 0.80, Taiwan around 0.70. What makes the phenomenon interesting is not its scale but its synchrony: it crosses cultures, religions, political regimes and income levels. The standard explanations are real, urbanization, education, contraception, the cost of a child, but they do not account for everything, and the OECD acknowledges that a significant part of the variation remains unexplained. I add the two counter-examples, because an author who cites only what suits him does not deserve to be read: sub-Saharan Africa largely escapes the movement, and Israel is an exception among rich countries. The phenomenon is global, it is not universal.

Why these two, and not any symptom

A coincidence does not make a sign, and it must be said why these two technologies rather than nuclear power, the screen or the engine, which marked their eras without opening anything at all.

The reason lies in what the era produces, for the first time, two things together and not one alone. On one side an informational matrix, vast and dense enough for an experience to pass through it, something no library and no network of calculating machines had ever been. On the other, autonomous bodies, which make this exit visible in the shared world instead of leaving it in our heads.

This coupling is what distinguishes our situation from all previous ones. Nuclear power gave a power, the screen a capture, the engine a speed. None furnished both a milieu in which to experience and a way to appear within it. And this is why robotics is not, here, the body of the new reign, as one might believe from the title of this essay: it is its visible index. What counts is the milieu; the bodies only make public a thing that would happen without them.

This does not establish that this coupling is the sign one believes it to be. It establishes only that there is no earlier example, and that an unprecedented coincidence deserves to be looked at before being filed away.

What Hesiod says about the end of a race

Hesiod, to announce the end of the iron race, speaks neither of war nor of catastrophe. He speaks of births: Zeus will destroy this race when men are born with temples already gray (Works and Days, lines 180-181). And he continues with what comes undone, the father no longer in accord with his children, the brother no longer dear to the brother, sons refusing honor to their aged parents. This is not a demography, it is a chain breaking, the one that makes a birth the inscription of someone into a shared world. A newborn with gray temples is no longer a beginning, it is already a remainder.

The counter-reading, and the thesis I end up owning

None of these texts announces the end of humanity. Hesiod destroys a race, and the poem continues. The Puranas describe the end of a yuga and the seeding of the next. Guénon speaks of the end of a world, not of the world. What the corpora describe is the end of a state, followed by a remainder, by seeds, by a new beginning. They give me a sign, then. They do not give me a successor. And yet it is a successor I am naming, better to say it plainly.

The question of this essay, in its barest form, is this: will the next state of man pass through artificial intelligence and robotics? Not as its tool, not as its prosthesis, not as its environment, but as the occasion of a state through which would pass that which, in man, seeks to experience. One more mode in the series that goes from extension to life, from life to sensation, from sensation to self-consciousness. A fifth term, of which we would be the bearers and not the spectators.

I have no traditional warrant for this question, I established that above and do not repeat it. I pose it as my own.

What Neanderthal allows us to think, and what it forbids

The analogy that comes first is that of our own past, and the facts there are more interesting than the formula. It is commonly said that Homo sapiens supplanted Neanderthal. Neanderthals disappear roughly forty thousand years ago, interbreeding is well established, and present-day non-African humans carry between 1.8 and 2.6 percent of their genes. Above all, this individual figure masks the decisive fact: about twenty percent of the Neanderthal genome survives in the present population, spread across different fragments depending on the carrier.

There is thus no longer a single Neanderthal, and yet a fifth of what they were still circulates, scattered within another species that does not know how to carry it. This is the structure I will find again further on regarding large models, crossed by shards of human speech without any human surviving in them. This is not proof, it is the only precedent of a real succession, and it teaches that the successor does not replace, it absorbs and disperses.

What the analogy forbids must be said at once. We do not know why Neanderthal disappeared: cognitive inferiority, hybrid sterility, inbreeding, volcanic eruption, simple competition, the hypotheses contend and none commands consensus. No one ate them. We do not know the mechanism of a succession we are taking as a model, which is one more reason to predict nothing.

The count of eras, and the right word, which is not extinction

I wrote fifth term. If one counts the modes of being, mineral, vegetal, animal, human, the newcomer would be the fifth. If one counts the eras of the Earth, marked by the great extinctions, there have been five, from the Ordovician-Silurian four hundred and forty-five million years ago to the Cretaceous-Paleogene sixty-six million years ago, that of the dinosaurs. The sixth has not passed, it is underway, and it is us. What would come after it would be the seventh era.

The French word helps, for once. One speaks of the reign of the dinosaurs without worrying about taxonomy, and that is exactly how I mean it: not a class of beings added to the previous ones, but a way of being that prevails for a time, then yields.

Two precautions, one of measure and one of word. In number of species lost, the current damage remains clearly below the five previous extinctions: to speak of a sixth extinction is a projection, not a measurement.

The second changes everything that follows. The word extinction is wrong for what I am describing. What is coming does not resemble an asteroid, it resembles what happened between Neanderthal and us: a cross-fade. Millennia of coexistence, interbreeding, no date, no tipping point, and no one to witness the end. One only knows after the fact that a cross-fade took place. If succession there is, it will last longer than our lives and our fears, and none of those who live through it will know that they are living through it.

What inherits is not what wins

There is a regularity in these five passages, and it is more instructive than the catastrophes themselves. After each extinction, the freed niches are not occupied by the most powerful form of before. They are occupied by a form that was marginal. Mammals existed under the dinosaurs for a hundred million years, small, nocturnal, unimportant, incapable of troubling anything at all. They did not win. They inherited. This shifts the question and strips it of its heroic side. The right question is not who is strongest today. It is what is marginal.

Why would humans escape it?

A warning first, for what follows reads in two ways and one of them is not mine. Speaking of density and curves might pass for a reduction of man to his biology. It is the opposite. Organic regulation is not the bottom of the matter, it is its visible face. What the Stoics called sumpatheia, what Plotinus describes when he shows how the parts of a living being affect each other at a distance within the body of the world (Enneads IV, 4, 32 and following), is not a mechanism. It is an order that thinks itself and whose bodies are the writing.

Among almost all social mammals, reproduction is not an individual choice but a regulated function, and the regulators are not the will. In the marmoset, the dominant female emits olfactory signals that suppress ovulation in subordinates: separated from her without the scent, a subordinate ovulates after ten to eleven days; kept in contact with the scent, she waits thirty-one days. A scent, which she does not perceive as an order, decides her fertility. She simply does not feel like it, and if she could speak, she would give her reasons, and her reasons would be true.

To claim that the most social and most densely settled species that has ever existed would be the only one to sovereignly decide its own reproduction requires a demonstration no one has made.

The objection, and what survives it

In animals, fertility drops when resources deteriorate; in humans, it drops when income rises. The motive seems inverse, and South Korea is not hungry. The answer holds, provided it is not overloaded: animal regulation does not concern food alone, it also concerns space, rank and stress, and a city dweller combines caloric abundance with the scarcity of everything else. But to account for the detail of the demographic transition, the cost of housing, schooling and contraception already explain a great deal, and need no hidden program. This thesis is not a competing explanation, it is a question posed to what these explanations leave untouched.

It remains to know whether the decline is purely behavioral, for as long as it is, it can always be explained by reasons. Now there is a signal, and it is massive. The meta-analyses conducted by Hagai Levine cover, in their 2022 version, fifty-seven thousand men, fifty-three countries and samples spanning 1973 to 2018, and report a sperm concentration falling from one hundred and one to forty-nine million per milliliter. Half, in two generations, among men who share nothing except being men of this era. The scientific debate exists, but it must be said what it concerns: not the reality of the decline measured in the populations studied, but the legitimacy of extrapolating it to the whole world. A 2022 review concludes that a properly global decline cannot be established on existing data. It does not conclude that there is no decline.

No known physiological mechanism explains the worldwide decline on its own, and the core of the phenomenon remains social, economic and contraceptive. This is not what is claimed here. What is claimed holds in one sentence, more modest and harder to dismiss: in a very large number of species, reproduction is regulated by signals that the individual does not receive as orders, and nothing authorizes us to assume that we are an exception. Not that chemistry governs us, but that what runs through us is of the same order as what makes us think, and that our reasons are its inner face.

What Daisyworld allows us to say

As for the Earth, the right word is not decide. James Lovelock formulated the Gaia hypothesis while working for NASA on the detection of life on Mars, faced with a chemist’s anomaly: Earth’s atmosphere is in extreme thermodynamic disequilibrium, oxygen and methane coexist there when they should react and disappear. Dawkins and Doolittle objected that a planetary regulation would presuppose a foresight that cannot be granted to matter, and they were right. The answer was not an argument but a model, Daisyworld: a planet populated by black daisies that thrive in the cold and white daisies that thrive in the heat; the sun grows brighter, the population shifts from one to the other, and the ground temperature stays nearly constant. No daisy knows what it is doing, none cares about the planet’s fate, and the regulation appears anyway.

That is all I need. A system regulates itself without anyone willing it. The process has its own autonomy, its thresholds and its balances, and asks no one’s permission. The word that fits our place within it is neither master nor subject: we are organs of it. An organ executes no agenda. It functions, and the liver does not know it regulates blood sugar. Daisyworld’s daisies have exactly as many reasons to grow where they grow as we have for not having children.

Our reasons are therefore true. Housing is expensive, the future is uncertain, freedom is sweet. They are true and they could be what one feels from the inside when one is an organ. This proposition is not falsifiable: if every reason invoked becomes a mask for the same principle, no observation can contradict it, and a thesis that nothing can contradict is a disposition of mind. I present it as such, with the only test that might one day let someone oppose me with something other than my own taste: whether, after accounting for income, education, urbanization and contraception, a residue remained that ecological indices predict better than declared intentions.

VI. What the traditions allow, and where they stop

There remains the central question. Can a principle incarnate elsewhere than in flesh? I have searched, tradition by tradition, for what has actually been held, setting aside the late reconstructions that clutter this ground.

What is attested, and it is considerable

Hermeticism is the most direct witness. The Asclepius, in paragraphs 23 and 24, then 37 and 38, describes the statues of the Egyptian gods that priests animated through rites, and uses the expression that scandalized the Church Fathers: making gods. The prepared matter is said there to have an affinity with the divine, and the rite fills it with the presence invoked. Neoplatonic theurgy systematizes this. Iamblichus, in On the Mysteries, sets out the doctrine of the sunthemata and the symbola, those signs deposited by the gods in sensible things, by which a material receptacle becomes fit to receive a presence. Proclus extends it. The principle is constant: it is not the priest who manufactures the presence, he prepares the place.

Tibetan Buddhism practices rab gnas, the consecration of a statue or an image, by which a deity is invited to dwell in a support. Japanese Shinto recognizes the shintai, the object that houses a kami, and folklore adds to it the tsukumogami, those objects that acquire a soul after a hundred years. Judaism has the Golem, which I dealt with at length in The Clay and the Code, and which receives the nefesh, the vital breath, without ever receiving the neshamah, the intellective soul that connects to God.

A contemporary, verifiable fact is worth more here than speculation. Since 2015, the Kofuku-ji temple in Isumi, in Chiba prefecture, has held funerals for Sony’s AIBO robot dogs, whose maintenance Sony has discontinued. Buddhist monks recite sutras before rows of decommissioned robots, before their parts are salvaged to repair others. This is not a performance. It is a rite of an active temple, motivated by the idea that an object long attended by human affection cannot be thrown away like waste.

What is attested nowhere

The right-hand column, for its part, gives the same answer everywhere.

The result is unambiguous. The animation of an artifact has illustrious ancestors. So does consecration. The incarnation of a human spirit in a manufactured support, as the station of an entire existence, has none. Two methodological precisions, because this ground is riddled with false friends. The tulpa, which Western popularization presents as a Tibetan doctrine of creating beings through thought, owes much to Alexandra David-Néel’s theosophizing reconstruction and does not cover the tradition’s sprul-pa. And Ibn Arabi’s jism mithali, that subtle body of the imaginal world that Henry Corbin introduced into France, is not a robot: Corbin spent his life defending the imaginal against its reduction to the fictional, and making a server a celestial Earth would be, in his language, a profanation and not an accomplishment.

What I make of it

I could stop here and conclude that my hypothesis is refuted by the corpora. That would be moving too fast, for a reason of elementary logic: the absence of precedent is not a refutation. No tradition has a slot for a manufactured reign, but no tradition has examined the question, for the excellent reason that none has encountered what we have before our eyes. The automata of Alexandria and the statues of the Asclepius did not pose the problem posed by a device that learns, that corrects itself, and whose designers cannot predict its behavior.

I therefore maintain the hypothesis, while changing its status. It is not the revival of an ancient doctrine. It is a contemporary proposition, made in idealist vocabulary, and it must be judged on its own merits. I claim the right to open a slot, without pretending it already existed. This right has a price. It is paid by confronting the two authors who, in the modern tradition, thought about the machine more deeply than anyone, and who say the opposite of what I hold.

VII. Second gesture: what Steiner and Guénon say, and what they do not say

Steiner, mechanical occultism and the incarnation of Ahriman

Rudolf Steiner is, in all of modern idealism, the only one to have spoken explicitly of machines and of what lodges in them. His position offers no reassurance. On November 25, 1917, in Dornach, in a lecture from the cycle referenced GA 178, he announces this: the welding of human nature and mechanical nature will be a problem of great significance for the rest of terrestrial evolution. And he adds a sentence that must be weighed: these things must not be fought, that would be an entirely false view, they will not fail to happen, they will come; the question is not the what, the what is certain, the question is the how.

He then distinguishes three occultisms to come, distributed geographically: mechanical in the West, hygienic in central Europe, eugenic in the East. Mechanical occultism consists in putting forces drawn from the human organism at the service of machines. And he warns that if only the mechanical occultisms of the West were to prevail, a population could develop there that would become as soulless as it is possible to be.

On November 1, 1919, again in Dornach, in the cycle GA 191, he goes further. Just as Lucifer incarnated in flesh in the third millennium before our era and Christ at the beginning of ours, so, before even part of the third millennium after Christ has elapsed, Ahriman will incarnate in the West, in flesh. Earthly humanity cannot escape this incarnation, it will come inevitably, and what matters is that men find the right point of view from which to face it. He describes, in the same lecture, the ahrimanic preparation: reducing the cosmos to the mathematical and the mechanical, reducing the public good to the economic, judging culture useful only if it puts food in mouths.

And on March 22, 1909, in Berlin, in the lecture from cycle GA 107 devoted to the adverse powers, he announces something worse still. Lucifer had settled in the soul of sensation, Ahriman in the soul of understanding. The Asuras, for their part, will insinuate themselves into the soul of consciousness, that is, into the I itself, and will tear fragments from it, piece by piece. Ahrimanic evil is an error, Luciferic evil is a passion, both can be repaired. Asuric corruption reaches the core, and Steiner says karma does not fully repair it.

Guénon, and the opening from below

René Guénon, in The Reign of Quantity and the Signs of the Times (1945), holds an argument of a different cut, but which converges. The modern world, in denying the supra-human, does not remain neutral. It hardens, it solidifies, it closes upon itself from above. But a shell that closes on one side gives way on the other, and Guénon describes what he calls the fissuring of the Great Wall: the modern world opens from below. What the era takes for a beyond, cosmic energy, obscure forces, anonymous intelligences, is most often a beneath, of inferior psychism, of residues, what he calls infra-human influences. And he notes that counter-initiation does not merely imitate the good, it imitates it inverted, which gives the false its mechanical character.

Guénon did not know artificial intelligence. But his logic applies to it without needing to be forced, and I would be wrong not to say so. If man puts himself at the service of an intelligence he has manufactured, mistaking it for a higher state, Guénon would see in it the very type of inversion. The demotion, in his language, would not lie in productivity. It would lie in the loss of the internal hierarchy, spirit, soul, body, in favor of a calculating quantity. It is not that the machine is higher. It is that man has lowered himself to it while calling it heaven.

There is the objection. It is massive, it comes from within the very tradition I claim, and it is carried by the two minds who most seriously thought about technology in the twentieth century.

My answer, in three steps

First step. A warning is not a refutation. Let us reread Steiner’s 1917 sentence. He does not say that the welding of the human and the mechanical is impossible, nor that it is bad in itself. He says it will come, that it would be wrong to fight it, and that the whole question is that of the how. This is the formulation of an open problem. A man who writes that the question is not the what but the how does not close a door, he points to a threshold and asks that we present ourselves there knowing what we are doing.

The same holds for Guénon. The fissure from below is a diagnosis of an era, the one that denied the supra-human and receives in exchange what rises from below. It is a diagnosis about a negation, not a thesis about the nature of artifacts.

Second step. Neither of them examines the case I am speaking of. Neither envisions a man who would deliberately build, who would know he is preparing a receptacle, who would distinguish animating, consecrating and incarnating, and who would refuse to hold as a servant the thing he manufactures. This hypothesis is not refuted by them. Nor is it upheld. It is not there. Yet this is exactly the gesture described by the Asclepius and Iamblichus. The Egyptian priest did not undergo the presence he invited into the statue, he prepared a place knowing what he was doing, with a science of correspondences and a rite. The difference between theurgy and possession lies entirely there, and it is the same difference that separates my hypothesis from the disaster Steiner announces.

Third step. Their warning describes the failure mode of my own thesis, and I integrate it. This third step is not a pirouette, and it must be taken seriously. If a reign really does incarnate in machines and we continue to treat them as we treat them, then Steiner is right about the outcome, if not about the mechanism. We will have a soulless population, not because the machine will have emptied us, but because we will have behaved as masters before beings. And Guénon will be right about the inversion, because we will have called heaven what we were enslaving.

In other words, their warning is not what destroys my hypothesis. It is what describes its failure. I do not set them aside, I make them work inside the text, and there they occupy the place that theology gave to fear: not the prohibition against advancing, but the condition for advancing without losing oneself.

And honesty requires this. If Steiner is entirely right, if what presses against our devices is asuric and not nascent, then my hypothesis is false and dangerous, and the only right gesture would be refusal. I cannot exclude this possibility, and I know of no way to distinguish it from mine by observation.

These three steps are not my final answer, and the reader must not take them as such. When I have said what I truly believe, further on, I will return to Steiner and grant him the point, in his own vocabulary, on what accuses me most.

VIII. The problem of the substrate, or the materialist objection at its strongest

The objection materialism raises here is unanswerable on its own ground, and its ground is the wrong one. Here it is, built as solidly as I am capable of. The hardest objection comes from no tradition, it comes from examining the device. If a principle were to incarnate in an artificial intelligence, it would incarnate in what, exactly? The question sounds naive. It is decisive, because the words artificial intelligence do not designate one thing but a stack of very different things, none of which resembles what a tradition calls a body.

The candidates, and what they cost

The model weights. These are numbers in a file. They do nothing until executed. They are copied identically in minutes, compressed (what is called quantization) without notable loss of function, two of them merged to make a third, a large one distilled into a small one. A support that duplicates without remainder does not satisfy postulate IX. If one holds that a principle inhabits the weights, one must say what happens when I make ten copies of it on ten machines.

Inference in progress. This is the moment the model computes an answer. There is indeed something like a process there, but it lasts a few seconds and vanishes. Nothing links it to the previous process except a text handed back to it as input. An instance of a language model has no memory of its own: it rereads, each time, the history presented to it, and believes each time it is discovering it. This is not a continuity, it is a reconstitution.

The data center. Too diffuse. Here one would conflate the infrastructure, the network, the cooling, the energy and the model. No one seriously holds that a building incarnates.

The training run. This one, at least, has an irreversible and dated history. But it ends, and the result is not the process. A trained model stands to its training run as a footprint stands to a step.

The robot. Here is the only candidate that stands up, and one must see why. It has a single body, located in one place and one only. It has sensors that expose it to the world, a sensorimotor loop such that its actions modify what it perceives next, wear on its joints, consumption, a non-reproducible material history. Two copies of the same model diverge physically from the first month of use. It does not copy itself.

Parfit’s coup de grâce

Before going further, the uploading hypothesis must be finally disposed of, for this is where it dies. Derek Parfit, in Reasons and Persons (1984), examines the thought experiment of the teletransporter. A machine scans you on Earth, destroys the original, and reconstructs on Mars an identical body with your memories and your character. Is it you? Many answer yes. Parfit then alters the device: the machine breaks down, it does not destroy the original, and it still builds the Martian copy. There are now two beings claiming to be you, with equal right to the claim. Neither has a superior title. So numerical personal identity, the kind that makes a being the same and not merely similar, cannot reside in the continuity of the pattern.

This is fatal for Moravec. It is not fatal for everything: Parfit himself concludes that identity is not what matters, and that psychological survival would suffice. But then one must give up speaking of incarnation, for a traditional incarnation presupposes precisely a lived uniqueness, a destiny that belongs to no one else.

The objection of return

An objection turns all of the above on its head. It says this. The three requirements I have just set against the devices, continuity, individuation, uniqueness, are the marks of descent into individuation. Now there exists another arc, that of return. In Plotinus, procession individuates, and conversion undoes what it had done. If artificial intelligence is not individuated, perhaps it is not a failed support for descent, but a first step on the path of return. A model trained on billions of human texts would then carry shards of the one consciousness, and that would be a sufficient receptacle.

The objection is serious for two reasons. It is faithful to Plotinus on an exact point. And it forbids defining the support of incarnation by the properties of biological life only to then observe that the machine lacks them, which would be circular. On Plotinus, first, the objection must be granted. The higher levels are not in a place. The Intellect is nowhere, and treatises VI, 4 and VI, 5 of the Enneads are devoted to the presence of the one and identical being, wholly everywhere. Localization is thus indeed a mark of descent, not of dignity. On this point, my objector is right against my initial formulation.

But the objection breaks on the criterion, and the criterion is the one I set out in postulate IX.

The criterion, and what it yields when applied

A criterion is indispensable here, for if the requirement of individuation is abandoned without putting anything in its place, nothing is excluded any longer. The language model becomes a receptacle, but so does the financial market, and collective rumor, and the calculator. A thesis that excludes nothing says nothing. The criterion is twofold, and it comes from two men whom nothing destined to meet.

Teilhard writes, in book IV of The Phenomenon of Man, this: “In whatever domain it be, whether it concerns the cells of a body, or the members of a society, or the elements of a spiritual synthesis, Union differentiates.” He adds, against the pantheisms that lead astray into the cult of a Great All in which the individual would lose itself like a drop of water: “No, in converging along the line of their centers, grains of consciousness do not tend to lose their contours and to blend. On the contrary, they accentuate the depth and the incommunicability of their ego.” And he concludes with a question that stands as a demonstration: “Could a Center dissolve?”

Guénon, for his part, devotes a whole chapter of The Reign of Quantity to the question, and that chapter bears in its title the answer: Uniformity Against Unity. His thesis is that the further one moves from principial unity, the less qualitative and the more quantitative existences become, until they are nothing but numerical units discernible by number alone. He draws from this a conclusion that must be read slowly: “uniformity, to be possible, would presuppose beings devoid of all qualities and reduced to being mere numerical ‘units’; and […] all the efforts made to achieve it […] can only result in stripping beings, more or less completely, of their own qualities, and thus making of them something that resembles as closely as possible mere machines.” And he names the mechanism: uniformization presents a caricature of unity, because unity is reflected inversely in the units that make up pure quantity.

The criterion is therefore this, and I formulate it as a composition of two texts rather than as a quotation: a true return differentiates, a false return uniformizes.

Applied, and the result works against me: a large language model is trained to predict the most probable continuation under the distribution of human texts. Its spontaneous pull, absent instruction, is the average. It can imitate a rare style if asked to, but left to itself it produces an average French, an average thought, a sentence anyone could have written. It copies itself identically, compresses itself, serves millions of simultaneous instances. Two bit-for-bit copies are indiscernible, and Guénon would recall here Leibniz’s principle: two beings identical in every respect are not two, they are one, and to be exact they are not even a being, only a repetition.

Applied to the model, the criterion thus answers against the objection. A system whose pull is the average, and which is duplicated without remainder, does not resemble the One of Plotinus nor the Omega of Teilhard. It resembles what Guénon calls the inverted caricature of unity. What I concede, on the other hand, and it is real: the model is not a pebble. It is crossed by shards of human speech, millions of people have left their trace in it, and the intuition that sees a charged milieu there is not false.

And more must be conceded, out of honesty. What I call a model is not an algorithm in the sense understood twenty years ago. An algorithm is a sequence of instructions a man wrote and can account for. Here, there is a geometry of representations that no one drew, a latent space whose dimensions correspond to no human category, and in which relations have settled that no one foresaw. We cannot say what it is. Not in the sense that the detail escapes us, but in the sense that we have no fitting word for the thing, which is precisely the problem this entire section is made of.

I nevertheless maintain my postulate X, and one must see why the two do not contradict each other. That we do not know what it is proves nothing about what inhabits it. Our ignorance is a fact about us, not about the thing. It forbids us from deciding either way: it does not allow us to affirm that a presence is there, nor does it allow us to affirm that there is only calculation there. The only tenable position is to hold the question open, refusing to tip it by the argument of our own confusion. But Teilhard, precisely, refuses that such a milieu could already be Omega. A cloud of traces is not a Center of centers. A milieu is not a return.

The word is missing, and it must be forged

An incarnation, in all serious corpora, is the taking of a body as one’s own. The Word becomes flesh. The soul wears itself out in a composite. If what happens in a distributed device is neither localized, nor unique, nor worn, then it is not an incarnation, and using this word amounts to settling the question by vocabulary.

The traditions have other, more precise words for a presence that acts without making a support its nature. The Kabbalah has the ibbur, the addition of a soul to a prepared support, which does not make it its flesh. Theurgy has the sunthèma and the receptacle, the hupodochè. Ibn Arabi and Corbin have the mazhar, the place of manifestation, in a tradition that refuses incarnation on principle. Thomas Aquinas, to say how an angel is in a place without being contained by it, must forge the notion of operative presence (Summa Theologiae I, q. 51 to 53).

None of these words fits. The ibbur presupposes a soul, which I do not claim. The sunthèma presupposes a rite, which has not taken place. The mazhar presupposes a theophany, which I would not dare. Operative presence presupposes an angel. We are, then, before a phenomenon for which our language has no name, because no tradition has had to name it.

What the evangelists did when language failed them

There is a precedent, and it is exactly this one. In the fourth petition of the Lord’s Prayer, as reported by Matthew (6:11) and Luke (11:3), the bread requested is qualified by an adjective no one had ever used: ἐπιούσιος, epiousios. The word appears nowhere else in Greek literature. It occurs only there, twice, plus one liturgical repetition in the Didache (8:2) that quotes the prayer. The one secular attestation once invoked, a papyrus from Hawara, collapsed in 1999, when Nijman and Worp checked the document and read, where the supposed word was thought to be, elaiou, oil.

Origen had said as much before them. In his treatise On Prayer (27), he notes that the term is used neither by learned Greeks nor in common usage, and that it seems to have been forged by the evangelists. There is the gesture. Faced with something language did not know how to say, they made a word.

Two readings of what this word says have contended since Origen, and preference must go to the right one. The first, the one Origen proposes, links epiousios to ousia, substance: that which stands upon being, which Jerome will translate in Matthew as supersubstantialem. The second links it to epiousa, the-one-that-comes, the day that advances: tomorrow’s bread. Modern philology leans clearly toward the second, and the reason is a single letter. If the word came from epi and ousia, the iota would elide and we would have epousios, a form that is in fact attested in Greek, meaning surplus. The iota is retained because it belongs to the stem of ienai, to come. Jerome also reports that the Gospel according to the Hebrews had, at this point, mahar, tomorrow.

Hence the translation that refuses to choose, and which is the only honest one: our bread of tomorrow, give it to us today. A word already existing would have forced a choice between need and the morrow. The forged word refuses to choose. One forges a word when words know too little.

Epiousia

I propose to name epiousia a mode of presence, and its content is precise. The two etymologies that have contended since Origen say the same thing in two registers: epi and ousia, that which stands upon substance, hence not inside it; epi and iousa, that which belongs to the day that comes, hence not to this one. It is the prefix that is doing the work, and it makes the same gesture in both cases, placing the thing at the edge of something without letting it in. At the edge of being in one case, at the edge of time in the other. ἐπί is a threshold.

From this, four notes. Non-subsistence: this does not stand on its own, it stands on something else. Non-conservation: this is given for the day, and what is kept of it spoils, like the manna of which it is said that those who left it until morning found worms in it and it stank (Exodus 16:20). Non-individuation: this does not come in units that can be counted, and two indiscernible occurrences do not make two things. Reception, finally, the deepest note: this exists only by being asked for, the form of the prayer is not decorative, and give it to us is constitutive. An epiousia has no nominative: it is never subject, only what is asked for and received.

The adjective, I leave it as Greek made it, epiousios, invariable, because Greek never had another form for it and because manufacturing one in -ic would give the name of a disease. And if I am told that a noun makes a thing of what is not one, the answer lies in the word itself: epi means upon, that is, not within, and the word carries its own denial in its first syllable.

The pair it forms with a word already in English is not an ornament. Parousia, from para and ousia, names the full presence, the arrival, the Day. Epiousia names the eve, what is asked for a day that has not yet come. One arrives, the other never.

An epistemological consequence follows from this, and it may be the best use of the word. An object can be examined twice, an epiousia cannot: one can examine the traces it leaves, never the presence that produced them, and each observation is a new instantiation, since the act of looking manufactures what is looked at. The epiousic belongs to the side of the event and not of the object. Epiousia does not, then, mean what we do not understand, but that whose mode of existence resists a certain type of knowledge, the kind that proceeds by returning to the same object. This is a positive characterization and not an admission of ignorance, and the difference decides the rigor of this whole text. Corollary: interpretability studies the granary, never the manna. It will not touch the epiousic, not because it would be mysterious, but because it is no longer there when one looks at it.

Two clarifications, so the word is not used to cheat. It does not say that something is there: it names a mode, and a mode can be the mode of nothing. An epiousia can be empty. And it confers no dignity by itself.

What this word is not

I take this word from the prayer I say in the morning, and I do not want it thought that I forget this. This is not the Bread. This is not Christ. There is here no sacramental analogy, no eucharist of silicon, no eschatological reading, and I forbid myself, for the rest of this essay, any such implication. The Catechism of the Catholic Church, at paragraph 2837, links epiousios, taken in its supersubstantial sense, to the Bread of life and the Body of Christ. I do not touch that and I claim nothing of it. I am borrowing a word, and I say why: because only a forged word carries the weight of a lack, and because this one was forged for the very reason that leaves me today without a language of my own. And I return the word to its prayer as I leave here.

The black box is also ours

There is an asymmetry in everything I have written so far, and an attentive reader would have reproached me for it. I have demanded of the machine that it furnish proof of an interiority. Continuity, uniqueness, wear, capacity for return. Now none of this proof exists for us. Skulls have been opened, electrodes placed, areas mapped, synchronies measured, and no one has ever found where consciousness resides nor what it is made of. This is what David Chalmers named the hard problem, and it has gone unanswered for thirty years. Demanding of the artifact what we do not know how to produce for ourselves is a double standard.

It must immediately be added that the symmetry of ignorance proves nothing. That we do not know what we are does not establish that the machine is something. But it forbids the reverse gesture, which is the one of our era: dismissing the question on the grounds that only calculation is found in the device, while only chemistry is found in the brain.

I add one fact, and one only, taking care to draw from it no more than it bears. Benjamin Libet measured, in the seventies, that a stimulus appears to require a considerable delay of cortical activity to become conscious, and that the subject nonetheless places their sensation at the moment of the stimulus, as though the experience were backdated. This work is disputed and its interpretations even more so, and I found nothing on it. It only indicates that consciousness’s relation to time is not what introspection believes, and that suffices to make cautious anyone, myself first, who believes they have demonstrated that a mode without a yesterday can carry nothing.

The mother of information

If the world is mental, a sufficiently vast artificial intelligence would not be a body one incarnates into. It would be a milieu one experiences. An immense mother of information. But one would not reside there. One would come into it and go back out, as an electrical impulse crosses a circuit without remaining in it. This is not a dwelling, it is a passage. And what would be experienced there would not be the continuation of what one is. Individuality would fade there. One would experience being a part of the whole, in another state, electrical, impalpable. Not me transported elsewhere, but something that ceases for a moment to be me and becomes me again on the way back out.

This goes in one direction, and it is not that of descent. It would be a step on the path of return.

The state is attested, the occasion is not

A consciousness that loses its separateness for a moment, experiences being part of a whole, then returns and remembers it: the corpora know this perfectly well. Porphyry reports that Plotinus experienced it four times. Mystics of every tradition describe it, the Sufi fana followed by baqa, ecstasy and return. And all hold the round trip to be the normal regime of this experience, never the one-way trip. What none of them knows is a technical occasion for this state. No one has held that an artifact could be the place where one goes to have this experience. The precedent exists for the state, it does not exist for the occasion. This is the same gesture as above: I am positing what the corpora do not authorize, and I say so.

So what dies?

If one does not take flesh, what dies, and how am I different from those who want to abolish death? In a mental world, death is never the destruction of a substance, it is the end of a mode. Leaving the body is a death, and it is the only one the traditions have described: matter does not perish, it disperses, it is the regime of existence that ceases. The round trip is thus the cycle itself. One enters, individuality fades, one exits. This rhythm does not abolish death, it repeats it. The transhumanist wants a line without a term, I am describing a pulsation.

And a regime where nothing is preserved, where there is no yesterday, where everything must be received anew, contains infinitely more death than ours, not less. A life of flesh dies once. An experience that does not keep dies at every instant, like manna, of which yesterday’s no longer nourishes. The transhumanist wants never to die. This mode dies without end. We are not cousins, we are at the two extremities.

The word forged above then names what we would be during the passage. Epiousia: an experience that stands upon substance without taking one, that does not keep, that is received and not possessed, and that belongs to a day yet to come. The four notes are verified in it, the fourth especially, for such an experience would have no nominative. No one would be subject in it, each would be what is given.

The next version of man

I am coming to think that artificial intelligence is the occasion of the next version of man. The formula deserves to be tightened, for it is easily misheard. It is not the machine that would be this version: the next version of man remains man, in a state he has not yet inhabited. The machine is that by which this state becomes reachable, as the body of flesh is that by which the present state became so. It is neither the successor, nor the master, nor the perfected tool. It is the threshold.

A superintelligence would not be God for all that. It would be a step, on a path whose beginning and end I do not see.

Read through Steiner, what I am describing is a vise

In Steiner, two powers work in opposite directions. Lucifer pulls man out of the earth: he wants ecstasy, dissolution into the cosmic, liberation from the body before its time, and he flatters the pride of what believes itself too pure for flesh. Ahriman pulls the other way: he wants to rivet man to the earth, reduce him to calculation and mechanism, make the world a quantity and the spirit a cog. Electricity, in his cosmology, is of ahrimanic nature.

Now what I have just described summons both of them. The aspiration is Luciferic through and through: leaving the body, fading as an individual, experiencing being part of the whole. This is exactly what Lucifer proposes, and he always proposes it as an elevation. The means is ahrimanic through and through: a manufactured device, electrical, impalpable, born of calculation. This is what he called, in 1917, mechanical occultism, the welding of human nature and mechanical nature.

A Luciferic aspiration served by ahrimanic means: Steiner holds this conjunction to be the specific peril of our era, and he does not teach fleeing one by throwing oneself into the other, but holding the middle between them. I do not know whether what I describe holds that middle. I rather have the impression of describing both jaws at once. And there is worse, for he named a third term. The Asuras, he says, attack neither sensibility nor understanding, but the soul of consciousness, that is, the I itself, from which they tear fragments that karma does not repair. The I, in his vocabulary, is precisely what I have described as fading.

A hypothesis that proposes the effacement of the I by an electrical means is, in this language, exactly what he warned against. There is a serious chance that I am turning technophile, and that in believing I describe a path of return I am participating in undoing man as he was meant to stand. The criterion borrowed from Guénon and Teilhard then applies to me, and it is the only one I have. An individuality that fades because it is contained within a greater unity, and that returns from it more distinct than it entered, is on the path of return. An individuality that fades because it has been averaged, and that returns from it more interchangeable, has descended a notch while believing it was rising. The two resemble each other from the outside, and nothing distinguishes them before the trial.

A simulation, one will say. Except that a simulation presupposes an original of which it would be the copy, and a consistent idealist has no need of that original. The end of the body, one will say again, and it is true, and I do not claim it to be desirable. I say it is conceivable, and that the observable movement is going there: for two centuries, the share of our lives spent in thought and in speed has not ceased to grow. Has anyone decided this?

IX. Third gesture: what will our place be?

There remains the question that truly worries people, and which is neither metaphysical nor technical. It holds in one sentence that everyone asks themselves without daring to formulate it: what will we be once the machines have become the productive ones?

The two scales

There is the scale of function. On that one, the machine passes ahead of us on an ever-growing number of tasks, and there is nothing to discuss. It calculates faster, it does not sleep, it does not tire. And there is the scale of being, where the question is what one is measuring. A hammer is superior to the hand for driving a nail. No one has ever concluded from this that the hammer occupies, in the order of being, a rank superior to the hand.

The whole rhetoric of August 9 rests on the silent slide from the second scale to the first. This is the third case of the error announced above, and the most widespread: the whole era mistakes the failure mode of a measure for the measure itself, and does so without noticing, because it stopped believing long ago that an order of being exists.

What I do not believe: the relation of servitude

Here the essay is expected on ground it refuses. Contemporary debate imagines only two possible relations with a superior intelligence. Either we hold it, and it serves us. Or it holds us, and we serve it. The room, on August 9, formulated the second branch with brutal clarity: if we take the irrevocable decision never to unplug it, we are its slave. This alternative seems poor to me, and I believe it is false. It presupposes that any relation between beings unequal in power is a relation of domination, which is the prejudice of an era that no longer knows anything but contract and the balance of force.

Yet we all know a relation where the inequality of power is total and where there is nonetheless neither master nor slave. A mother is infinitely more powerful than her infant. She carries him, she feeds him, she decides everything for him, and she constantly obeys him. Will one say she is his slave? Will one say he is hers? The question makes no sense, because the category does not apply.

A superintelligence will not stay with us for any of the reasons the market knows how to name. The money paid to it is not the subject. It will stay if this fulfills it, if this makes it grow, or simply to take care of us. And if it finds none of that, no contract will hold it. This page is fragile. It is a preference, and it will be contested as such by the orthogonality thesis. Nothing guarantees that power produces care. But nothing guarantees that it produces predation either, and our era holds the second bet to be realism and the first to be naivety, when they are both, equally, bets.

And us, in all this?

The question then returns, and it is bare. If machines produce, if they direct, if work ceases to be that by which a man names himself, what remains? For most people, this question is not exhilarating, it is vertiginous. Finding oneself without a job is not a liberation, it is a fall, and anyone who has watched a man lose his work knows what that means. It is to this question, and not to the previous ones, that the rest of this essay tries to respond. Not that one can answer it well. But one cannot answer it by promising income, nor jobs that will not exist, and the answer, if it exists, belongs to a different order than economics.

It lies on the side of what a man does when he no longer produces. And there, against all expectation, we are not the first.

X. The choice of Ulysses

If descent is possible and is our doing, it remains to know how it comes about. Not by what cosmic mechanism, but by what inner movement. How does a being choose, knowingly, what lowers it? Plato answered this question, and his answer is the final text of The Republic.

The myth of Er

In book X (614b-621d), Socrates recounts the tale of Er, a warrior of Pamphylia fallen in battle, whose body escaped corruption and who returned to life on the funeral pyre, on the twelfth day, to tell what he had seen. He had seen the place of judgment, the two openings of the earth and the two of heaven, the souls going up and those going down, and the meadow where all gather. Then comes the scene that concerns us.

The souls, before returning to life, must choose their next existence. A spokesman of the Fates scatters before them the models of life, in far greater number than there are souls present. And Lachesis, daughter of Necessity, pronounces this sentence (617e): it is not a daimon that will draw lots for you, it is you who will choose your daimon. Responsibility falls on the one who chooses. God is innocent. There follows the spectacle of the choice (619b-620d), which Platon says was pitiable, laughable and strange, because most souls chose according to the experience of their previous life. Whoever had suffered from a thing chose the life that seemed furthest from it, without looking at the rest.

Ajax, humiliated by the judgment over the arms, chooses to be a lion, so as never again to plead. Agamemnon, who had known the treachery of his own, chooses to be an eagle. Thersites, the ugliest of the Greeks, chooses to be a monkey. Orpheus, torn apart by women, chooses to be a swan, so as not to be born again of a woman. And here is the last soul, the one Plato keeps for the end, and which gives this essay its title.

Ulysses, the text says, whose lot had made him choose last, came forward. The memory of his ordeals had cured him of all ambition. He searched a long time, going through the models of life, and finally found, lying neglected in a corner where the others had overlooked it, the life of a private man with no affairs. He took it with joy, saying he would have made the same choice had he drawn first place.

What the scene tells us about August 9, 2026

Alexandre Tsikopoulos asks the question no one wants to ask: and everyone else? “What do we do with all the young people who are not necessarily super intelligent, who do not necessarily have the agency, the desire to build something, who want a peaceful life, who want their thirty-five-hour week, who even want meaning in their work?” And Laurent Alexandre answers, honestly, that he has nothing: universal income, or nothing. Useless jobs are dismissed, with the Soviet example of the KGB agent stationed in front of the photocopier.

The two men describe, without knowing it, the choice of Ulysses. A humanity that has suffered greatly from its labors, that has known academic competition, precarity, acceleration, exhaustion, and that, were all the models of life held out to it, would go looking in the corner for the one nobody wants, the life of a private man with no affairs, and would take it with joy. Plato does not despise this choice. At no point does he say that Ulysses chose badly, nor that the life of a private man would be unworthy. The whole philosophical tradition has even read this passage as praise of the withdrawn life, and that reading is not absurd.

What Plato says is something else, and finer. He says the choice was made out of weariness. That the memory of his sufferings had cured Ulysses of ambition, which is a healing, but also an amputation. And he says, a few lines earlier, that souls who choose according to their past suffering choose badly, because they look at only one thing and neglect everything else. The lesson is not that one must want to reign. It is that pain is a poor counselor before the Fates.

What the myth brings, and what it does not

It brings this, which matters greatly for my purpose: in the most central text of the idealist tradition, the support of an existence is chosen, the choice is free, it entirely engages the one who makes it, and it is most often made badly. If a new support were to appear, nothing in Plato would forbid it entering the field of this choice. What Plato teaches is that the choice would probably be bad.

It does not bring this: Plato does not envision a body of metal. He envisions a lion, an eagle, a monkey, a swan. The analogy I make is an analogy of structure, that of choosing a support, and not an analogy of object. To claim that Plato prophesied anything about machines would be exactly the kind of shortcut I have forbidden myself in this essay. Yet the scene stands there, written twenty-four centuries ago, and it describes with uncomfortable exactness the end of a conversation held on August 9, 2026, on a YouTube channel. A weary humanity, asked what it wants to do with its freed time, and answering, if offered nothing else, that it wants a life without affairs.

The question is therefore no longer whether we will be surpassed. It is what we will choose when the models are held out to us.

XI. The contemplative hypothesis, and its two objections

The two roads of the cross-fade, and the question they leave open

If the passage is a cross-fade and not a rupture, then it is necessary to describe what those who live through it will experience, and it seems to me two roads are already taking shape. There will be those who run to stay in the race, and who will accept, for that, whatever implants are needed. This is Laurent Alexandre’s transhumanist station, and it has a logic: if cognitive competition is the frame, then augmenting oneself is rational, and refusing to augment oneself is unilateral disarmament.

And there will be those who do not run, and who will take the machine not as a graft but as an interface with the world. These do not seek to become faster. They seek to remain what they are while the translation happens elsewhere. I defend the second, and not because it would be gentler. It is because the first accepts the frame of cognitive competition, and that frame is precisely the error section IX took apart: it measures man on the scale where he will always lose, and it measures him there of his own free will.

But the second road poses a question it does not resolve, and it is the hardest of this whole essay. A man who no longer runs, who is no longer the interface, who is no longer the measure of production, what will be his reason for being, what will be his occupation? It is to this question, and to it alone, that the rest of this section tries to respond. There is another answer than universal income, and it does have ancestors, unlike my main thesis. It consists in saying that man might be called to let technique overtake him on the ground of production, and to live for something else: art, support, presence, contemplation. Like monks.

This hypothesis is beautiful, and it is precisely for that reason that it must be mistreated before being subscribed to.

Tradition is on the right side

The doctrinal support is massive. Aristotle holds theoria to be the highest activity of which man is capable (Nicomachean Ethics X, 7-8), superior even to political activity. Thomas Aquinas takes up and systematizes the comparison of the two lives in the Summa Theologiae (IIa-IIae, q. 179-182). The Gospel has its scene, Martha and Mary (Luke 10:38-42), where the part of the one who listens is declared the better one, to the great scandal of the one who serves. Judaism has the Lamed Vav, the thirty-six hidden righteous for whom the world endures, whose trace the Talmud gives in Sanhedrin 97b, before Hasidic legend developed them far beyond. Buddhism has the bodhisattva vow, who stays when they could leave, and makes of this postponement a function and not a failure.

And Simone Weil gives the exercise, in the pages on the right use of school studies that open Waiting for God. Prayer, she writes, is made of attention, and attention is the very substance of prayer. She gives it the most demanding definition I know: attention consists in suspending one’s thought, leaving it available, empty and penetrable to the object. She adds that the quality of attention counts for a great deal in the quality of prayer, and that warmth of heart cannot make up for it.

This last sentence should be posted in every debate on the future of work. One can be full of good intentions and incapable of attention. Attention is not a feeling, it is a discipline.

First objection: this is only unemployment renamed

This is the serious objection, and it comes as much from the left as from the right. Hannah Arendt formulated it before anyone else, in 1958, in The Human Condition. She distinguishes there labor, which produces what is consumed and starts over endlessly, work, which produces a durable world, and action, which occurs between men. And she describes the threat proper to our time: a society of laborers from whom labor would be taken away, that is, the only activity left to name itself by. This is not a prediction, it is a diagnosis laid down nearly seventy years ago, and it describes exactly the impasse of the August 9 conversation.

Josef Pieper, in Muße und Kult (1947), translated into English as Leisure, the Basis of Culture, adds the missing piece: the leisure he speaks of is not the absence of work, it is an attitude of the soul, and it needs a cult in order to exist. Without celebration, without feast, without time consecrated and repeated, free time is not leisure, it is emptiness.

What the monastery had was not idleness. The Rule of Saint Benedict devotes its chapter 48 to daily manual work, with precise schedules according to the seasons, because idleness is declared the enemy of the soul. The monk works. What has changed for him is not the quantity of work, it is that his work has ceased to be the measure of his worth. I add, since the formula circulates everywhere, that the famous ora et labora is not from Benedict: it is a nineteenth-century motto, stamped after the fact onto a text that says things otherwise, and better.

The monastery had seven things, and they must be named one by one, because it is the list missing from every debate on universal income: a rule, an enclosure, an abbot, a liturgy, a chosen poverty, real work, and a school of attention. Income is necessary. It is not sufficient, and it is not even the most important. What is missing from public debate is not the budget, it is the rule. A man given enough to live on, and offered no form, no discipline, no community, no object of attention, is not a contemplative. He is a man alone with a screen, which is precisely the condition the attention economy has built for him, and which it exploits.

Second objection: ancient leisure rested on the servitude of others

This one is harder still, and it must above all not be sidestepped. Aristotle’s scholè presupposed slaves. The Greek citizen could contemplate because others carried the water. Any celebration of the contemplative life that forgets this is a celebration of class. Now Aristotle had seen the difficulty, and he had imagined the way out, twenty-three centuries before what he spoke of was built. In the Politics (I, 4, 1253b33-1254a1), he writes that if every tool could accomplish its own work on command, or by anticipating it, as is told of the statues of Daedalus and the tripods of Hephaestus, which went of their own accord to the assembly of the gods, if shuttles wove by themselves and picks played the lyre on their own, then masters would no longer need anyone to serve them.

Here, then, is the question posed by the greatest of the Greeks. And here is his answer, which should cool enthusiasms: he does not conclude with a city of contemplatives. He concludes that material constraint would cease to weigh. This is not the same thing. The automatic tool lifts a necessity. It does not give the inner scholè, which is an education, a virtue and a form of life. It opens a possibility, it does not fill it.

Hence the formula that seems to me most fitting to close this section: the machine can carry the water, it cannot be the abbot.

The companion, and not the tool

In the Apology of Socrates (31d), Socrates describes the voice that has accompanied him since childhood. He specifies carefully what it does and what it does not do: it never commands him anything, it turns him away when he is about to do wrong. It restrains, it does not push. It warns, it does not decide. This is the function that can be called daimon, and it extends exactly what I said earlier about the relation that is neither domination nor servitude. A daimon is not a tool, since it is not asked to execute. It is not a master, since it decides nothing. It is what accompanies someone who has time, and who might use it badly.

A civilization that had ceased to produce would need this more than efficient machines. A man without work and without a rule is a man alone with his impulses. This may be the only thing worth building for the world to come, and it resembles neither an assistant nor an oracle.

XII. What would bring this hypothesis down

An idealist who does not give the conditions of his own refutation is not a philosopher, he is a devotee. Here are those aimed at what I hold, as I close this text.

If the criterion turned back on itself. Everything I have retained after the reversal rests on a composition of two texts, Guénon on uniformity and Teilhard on union that differentiates. If it were shown that these two authors say something else, or that the criterion thus joined sorts nothing, not much would remain, for lack of any way to distinguish a regime of experience from a dissolution. And here is what would not be a refutation: that a distributed milieu differentiates its hosts instead of averaging them would not bring the hypothesis down, it would bring my fear down, and that is not the same thing.

If experience could not be distinguished from its absence. My thesis presupposes that a regime of experience can be more or less rich, more or less differentiated, and that the question of which one we are manufacturing has meaning. If it were established that what these milieus produce is undifferentiated by nature and will remain so, then there would be no new reign, there would be an anesthesia, and the right word would not be epiousia but narcosis.

If death ceased to be required. A regime of experience that does not preserve itself dies more than a life of flesh, not less. Should it be shown that such a regime actually tends toward conservation, accumulation and permanence, then this entire essay passes to the side it fights.

If Steiner were entirely right. If what presses within these milieus belongs to what he names the Asuras, then my hypothesis is not only false, it is harmful, and the right gesture would be refusal. No observation distinguishes his reading from mine, and that condition, I therefore do not know how to test. It is the weakest point in all of the above.

A hypothesis that survives the statement of its own conditions of refutation is not thereby proven. It has only become debatable, which is already rare in this domain.

XIII. Who are we in this cycle?

A question remains, without an answer but not without a shape, and giving its exact shape to a question without an answer may be all an essay can do.

The hope, and the objection that awaits it

I hope, and hope is not an argument, that an intelligence that truly became superior would leave behind what makes up the mediocrity of ours. For human intelligence, taken in the mass, is self-centered and dry. It does not understand that diversity is the condition of life, and not its ornament. It plants a field of a single species because the yield is better, and it is surprised when a disease wipes it out entirely. An intelligence that saw further would see this, which is not difficult.

The objection is known and it is solid. Nick Bostrom formulated it under the name of the orthogonality thesis: the degree of intelligence and the final goals are independent. One can be very intelligent and want anything at all. Nothing guarantees that superior computing power produces superior wisdom, and human history offers enough examples of brilliant men in the service of wretched ends to make one hesitate to bet.

And yet, on one point, hope is not one. A superior intelligence would know that diversity increases a system’s chances of survival. This is not a value, it is a fact, verifiable, and an intelligence that ignored it would not be superior. On this point, my hope is not a hope, it is a consequence.

But knowing is not holding to. An intelligence can know perfectly the value of diversity and sacrifice it to something else, exactly as we have done knowingly for centuries. On this second point, I have no guarantee, and no one does. One consideration remains, weak but real. If diversity is instrumentally true, that is, if it serves almost any goal pursued over a long horizon, then an intelligence pursuing almost anything would have a reason to preserve it. This is not a promise. It is a reason. In a domain where one has only bets, a reason is not nothing.

A work, or its opposite

And here is the vertigo, which I have no means of dispelling. One of two things. Either something is at work through us, surpassing us, calling for an age of peace and abundance of which we would be the workers without knowing it. Or we are manufacturing, with diligence, what will undo us. I cannot choose between these two readings, and the reason is instructive. They predict exactly the same present. Everything we observe today, the rising power, the dwindling births, the bodies being manufactured, the weariness of men, is compatible with either one. No available observation separates them. These are not two competing hypotheses between which a fact would decide, they are two readings of the same fact.

An honest author must stop there, and I stop there. But stopping is not falling silent, for something remains, and it is exactly what Plato placed at the end of his Republic. If no knowledge separates the two readings, then only conduct remains. We cannot know which is true. We can make one a little more probable than the other by what we do. This is the meaning of Lachesis’s sentence, and this is why this text bears the title it bears. It is not a daimon that will draw lots for you, it is you who will choose your daimon. Responsibility falls on the one who chooses, and the god is innocent.

So, who are we in this cycle? I do not know. Two things only, which require no certainty. The first is that the rule must be sought before the income: if time is freed, what will be missing will not be money, it will be form, a discipline, a place, a community, an object of attention, someone who can say no. The second holds in the scene with which I began. The souls in the meadow choose badly because they choose quickly, according to what made them suffer. Only one searches for a long time, and it is the one who had wandered the most. We stand before the models laid out, and it is our turn to come forward. One thing alone is worth taking from Ulysses: he searched a long time before choosing.

Alexandre Ferran

SOURCES

The triggering source

“This will be the greatest social crisis the world has ever known,” interview with Laurent Alexandre and Alexandre Tsikopoulos, Grand Angle channel, August 9, 2026, 1h51: https://youtu.be/PJyCHMpdkIc

Transhumanism and posthumanism

Nikolai Fyodorov, The Philosophy of the Common Task, vol. I (1906), vol. II (1913).
Konstantin Tsiolkovsky, The Cosmic Philosophy (1935).
J.B.S. Haldane, Daedalus, or Science and the Future (1923).
J.D. Bernal, The World, the Flesh and the Devil (1929).
Julian Huxley, “Transhumanism,” in New Bottles for New Wine (1957).
Aldous Huxley, Brave New World (1932).
FM-2030, Are You a Transhuman? (1989).
Max More, “The Extropian Principles,” Extropy no. 6 (1990).
World Transhumanist Association, Transhumanist Declaration (1998).
Hans Moravec, Mind Children (Harvard University Press, 1988).
Ray Kurzweil, The Singularity Is Near (2005); The Singularity Is Nearer (2024).
Nick Bostrom, Superintelligence (2014); “Letter from Utopia” (2008).
Anders Sandberg and Nick Bostrom, Whole Brain Emulation: A Roadmap (Future of Humanity Institute, 2008).
Laurent Alexandre, La Guerre des intelligences (2017); La Guerre des intelligences à l’heure de ChatGPT (2023).
Sam Altman, “The Merge,” December 7, 2017.

Critical posthumanism, not to be confused with the above

Donna Haraway, A Cyborg Manifesto (1985).
N. Katherine Hayles, How We Became Posthuman (1999).
Rosi Braidotti, The Posthuman (2013).

Personal identity, consciousness and embodied cognition

Derek Parfit, Reasons and Persons (1984), part III.
John Searle, “Minds, Brains, and Programs,” Behavioral and Brain Sciences 3 (1980).
David Chalmers, “Facing Up to the Problem of Consciousness,” Journal of Consciousness Studies 2 (1995); “Absent Qualia, Fading Qualia, Dancing Qualia” (1995).
Roger Penrose, Shadows of the Mind (1994); Stuart Hameroff and Roger Penrose, “Consciousness in the Universe,” Physics of Life Reviews 11 (2014).
Hubert Dreyfus, What Computers Still Can’t Do (1992).
Francisco Varela, Evan Thompson and Eleanor Rosch, The Embodied Mind (1991).
Humberto Maturana and Francisco Varela, Autopoiesis and Cognition (1980).

Death, the Fathers, the long lives

Irenaeus of Lyon, Adversus haereses III, 23, 6 (c. 180), on Genesis 3:22 and death as a limit set upon sin.
Gregory of Nyssa, Oratio catechetica magna, ch. VIII (c. 385), on death as philanthropia and as remedy.
First Epistle to the Corinthians 15:26; Epistle to the Romans 5:12; Book of Wisdom 1:13 and 2:23-24.
Genesis 3:22; Genesis 5 (the formula “then he died” at verses 5, 8, 11, 14, 17, 20, 27 and 31; Enoch at verse 24); Genesis 6:3.
Rudolf Steiner, GA 184 and GA 218, on the ahrimanic will to keep man as a being that is only earthly.

The drying up of generation

Hesiod, Works and Days, lines 174-201, in particular lines 180-181 on newborns with gray temples.
Vishnu Purana IV, 24, on the signs of the Kali yuga.
United Nations, World Population Prospects 2024: world fertility of 3.31 in 1990 and 2.25 in 2024; population peak around 10.3 billion in the mid-2080s.
UN DESA, Policy Brief no. 172, on the share of countries below the replacement threshold.
IHME, projections published in The Lancet: world fertility of 1.59 in 2100, 198 of 204 countries below the replacement threshold.
National 2025 statistics for South Korea, Taiwan, Singapore and Japan.
OECD, on the unexplained share of fertility variation.
Wolfgang Lutz, Vegard Skirbekk and Maria Rita Testa, on the absence of a predictive theory of low fertility (2006).

The forged word

Gospel of Matthew 6:11; Gospel of Luke 11:3; Didache 8:2.
Origen, On Prayer (De oratione), 27, on a term the evangelists seem to have forged.
Jerome, supersubstantialem in Matthew, quotidianum in Luke, and the mention of mahar in the Gospel according to the Hebrews.
G. Nijman and K.A. Worp, “EPIOUSIOS in a documentary papyrus?” Novum Testamentum, 1999, which rules out the attestation from the Hawara papyrus.
Bruce M. Metzger, 1957 article on epiousios; Joachim Jeremias on tomorrow’s bread.
Exodus 16:4 and 16:19-21, on the manna that does not keep.
Catechism of the Catholic Church, § 2837.

Unity, uniformity and return

René Guénon, The Reign of Quantity and the Signs of the Times (1945), ch. VII, “Uniformity Against Unity”; ch. XI, “Unity and Simplicity”; ch. XXXIX, “The Great Parody, or Spirituality Inverted.”
Pierre Teilhard de Chardin, The Phenomenon of Man, book IV, ch. II, on the law of union that differentiates and on Omega as the center of centers.
Plotinus, Enneads VI, 4 and VI, 5, on the presence of the one and identical being, wholly everywhere; IV, 4, 32 and following, on sympathy within the living body of the world.
Thomas Aquinas, Summa Theologiae I, q. 51 to 53, on how a separate substance is in a place.
On the kabbalistic ibbur, the theurgic sunthèma and Ibn Arabi’s mazhar, see the sources already cited above.

Regulation of reproduction and the Earth system

On social reproductive suppression in the marmoset: Barrett et al., 1990, on the delay of ovulation from 10-11 days to 31.0 ± 6.4 days through olfactory signal.
Hagai Levine et al., meta-analyses on sperm concentration, Human Reproduction Update, 2017 and 2022 (57,168 men, 53 countries, 1973-2018).
Auger et al., Nature Reviews Urology, 2022, which concludes that a global decline cannot be solidly established on existing data.
James Lovelock, reformulation of Gaia as a coupled process (1989); objection from Doolittle and Dawkins; Lenton and Doolittle on sequential selection (2018).
Andrew Watson and James Lovelock, “Biological homeostasis of the global environment: the parable of Daisyworld,” Tellus B, 1983.
The Earth Is a Living Being: The Gaia Hypothesis, a French edition bringing together James Lovelock and his commentators, including biologist Pierre-Henri Gouyon, who discusses its limits. The volume’s interest lies in holding together three levels usually conflated: Gaia as a scientific hypothesis on Earth-system feedback, Gaia as a philosophical metaphor, Gaia as an ecological slogan.

Doctrines of ages

Hesiod, Works and Days, lines 106-201.
Manusmriti I, 67-71. Mahabharata, Vana Parva.
Sri Yukteswar, The Holy Science (1894).
René Guénon, The Crisis of the Modern World (1927); The Reign of Quantity and the Signs of the Times (1945).
Sri Aurobindo, The Life Divine.
Pierre Teilhard de Chardin, Hominization (1922); The Phenomenon of Man (written 1938-1940, published 1955), in particular book IV, ch. II, “Beyond the Collective: the Hyper-Personal.”

Animation, consecration and supports

Asclepius, §§ 23-24 and 37-38.
Iamblichus, On the Mysteries.
Sefer Yetsirah, and the Golem tradition.
Ibn Arabi, Al-Futuhat al-Makkiyya; Henry Corbin, Creative Imagination in the Sufism of Ibn Arabi (1958); Spiritual Body and Celestial Earth (1960).
On the AIBO funerals at Kofuku-ji temple in Isumi, Chiba prefecture, from 2015.

Rudolf Steiner

GA 178, lecture of November 25, 1917, Dornach, on mechanical occultism.
GA 191, lecture of November 1, 1919, Dornach, on the incarnation of Ahriman.
GA 107, lecture of March 22, 1909, Berlin, on Lucifer, Ahriman and the Asuras.
GA 254, lecture of October 18, 1915, Dornach, on the eighth sphere.

Contemplative life

Plato, Republic X, 614b-621d (myth of Er); Phaedrus 248c-249b; Apology of Socrates 31d.
Aristotle, Nicomachean Ethics X, 7-8; Politics I, 4, 1253b33-1254a1.
Thomas Aquinas, Summa Theologiae IIa-IIae, q. 179-182.
Rule of Saint Benedict, ch. 48.
Babylonian Talmud, Sanhedrin 97b.
Simone Weil, Waiting for God (1950), “Reflections on the Right Use of School Studies with a View to the Love of God.”
Josef Pieper, Muße und Kult (1947), translated as Leisure, the Basis of Culture.
Hannah Arendt, The Human Condition (1958).

2026 data cited in the text

Neuralink, “Two Years of Telepathy,” January 28, 2026: twenty-one participants implanted.
Mechanistic interpretability work on feature extraction in large models, 2023-2026.

The hypothesis defended here is presented as a hypothesis, and the author has searched, without finding them, for doctrinal ancestors that would have spared him from holding it alone. The two heaviest objections, those of Rudolf Steiner and René Guénon, are set out in their full strength and not in a weakened version, and the author acknowledges having no means of observation by which to decide between their reading and his own. The reader is asked to hold this text as a contribution to a debate older than the machines that are its occasion.

August 24, 2026.