Timothy Morton

Foreword

By Edo Segal

The thing I could not find a metaphor for was the thing itself.

I had the river. Intelligence flowing for 13.8 billion years. I had the beaver. Building dams with sticks and mud and teeth. I had the fishbowl. The water you breathe without noticing. These images carried me through *The Orange Pill* because they gave shape to something I could feel but not name — the sensation of being inside a transformation so total that the transformation includes you, your tools, your thinking, your capacity to describe what is happening.

Then I encountered Timothy Morton, and the images broke open.

Morton is a philosopher who studies entities too large to see. He calls them hyperobjects — things so massively distributed across time and space that no single observer can perceive them as wholes. Climate change is a hyperobject. The Pacific garbage patch is a hyperobject. You cannot point to them. You can only point to their effects. You are inside them. They are inside you. The boundary between observer and observed has dissolved.

The moment I read that, something clicked that had been grinding for months. Every framework I had for thinking about AI assumed I could stand somewhere and look at it. The river metaphor assumed a bank. The beaver metaphor assumed a builder separate from the current. Even the fishbowl assumed glass you could press your face against.

Morton says: there is no bank. There is no glass. There is no outside. The AI transformation is not a river you can dam or a fish tank you can peer into. It is the water. It is the atmosphere. It is the total reorganization of human cognitive life by a distributed computational intelligence that exceeds any single manifestation, any single tool, any single company, any single moment of use.

That does not make my metaphors wrong. It makes them incomplete. And the incompleteness matters, because the part they miss — the part where the builder is inside the thing being built, where the observer is constituted by the thing being observed, where every intervention propagates through a mesh too vast to map — is the part that explains why this transformation feels so different from every previous one. Why you cannot simply regulate it, refuse it, or master it. Why the only honest posture is something closer to coexistence than control.

This book gave me a rotation I did not know I needed. Not a new direction. A new dimension. If *The Orange Pill* is the view from inside the river, Morton offers the unsettling recognition that the river has no banks.

Read this one slowly. Let it disorient you. The disorientation is the point.

— Edo Segal ^ Opus 4.6

About Timothy Morton

Timothy Morton (born 1968) is a British philosopher and the Rita Shea Guffey Chair in English at Rice University. Born in London, Morton studied at the University of Oxford and has become one of the most influential thinkers in contemporary ecological philosophy and object-oriented ontology. Their major works include *Ecology Without Nature: Rethinking Environmental Aesthetics* (2007), *The Ecological Thought* (2010), *Hyperobjects: Philosophy and Ecology after the End of the World* (2013), *Dark Ecology: For a Logic of Future Coexistence* (2016), and *Being Ecological* (2018). Morton coined the widely adopted concept of "hyperobjects" — entities so massively distributed in time and space that they transcend spatiotemporal localization — and has applied this framework to climate change, nuclear waste, and the total mesh of ecological interconnection. Their work on "dark ecology," the "mesh" of interconnected existence, and the concept of "strange strangers" has reshaped how scholars across disciplines think about the relationship between humans, nonhuman entities, and planetary-scale systems. Morton's writing is celebrated for its accessibility, philosophical rigor, and willingness to inhabit discomfort rather than resolve it.

Chapter 1: What Is a Hyperobject?

Somewhere in the Pacific Ocean, between California and Hawaii, there is an island made of plastic. Except it is not an island. It has no shore. It cannot be photographed from space. A ship could sail through the densest part of it and the captain might not notice anything unusual — just water with an odd shimmer, a faint chemical smell, fragments so small they resemble plankton more than packaging. The Great Pacific Garbage Patch is estimated to span 1.6 million square kilometers, roughly three times the size of France, yet no one has ever seen it. Not as a totality. Not as the thing it is. Researchers sample it. They model it. They track its currents and measure its density at specific coordinates. But the entity itself — the aggregate of every plastic bottle cap, every degraded grocery bag, every fragment of every synthetic object that has entered the Pacific watershed over the past seventy years — exceeds the perceptual apparatus of any individual observer. It is too large. Too distributed. Too temporally vast. Too entangled with the medium it inhabits.

Timothy Morton coined a term for entities like this. The term is hyperobject, and it names a category of being that the Western philosophical tradition — organized as it is around objects that can be seen, held, measured, bounded — has no prior vocabulary for. A hyperobject is an entity so massively distributed in time and space that it transcends spatiotemporal localization. One cannot point to it. One can only point to its effects. One cannot stand outside it and observe it from a safe distance. One can only inhabit it and attempt — always incompletely, always from a position that is itself shaped by the entity — to think about what it means to inhabit it.

The examples Morton deploys in his foundational 2013 text, Hyperobjects: Philosophy and Ecology after the End of the World, are chosen for their disorienting specificity. Global warming is a hyperobject: it persists across centuries, affects every ecosystem on the planet, and manifests locally as weather events that are themselves too ambiguous to serve as proof of the larger entity. A single hurricane is not global warming. A single drought is not global warming. Yet global warming is present in every hurricane and every drought — not as cause in the simple mechanical sense, but as the larger entity of which each weather event is a local manifestation. Nuclear radiation is a hyperobject: plutonium-239 has a half-life of 24,100 years, which means that waste produced by a reactor operating in 2025 will remain radioactive long after every civilization currently in existence has dissolved. The waste will outlast not merely the humans who produced it but the languages in which those humans might have warned their descendants. Styrofoam is a hyperobject: every coffee cup, every packing peanut, every fast-food clamshell ever manufactured still exists somewhere — in a landfill, in the ocean, in the stomach of an albatross — and will continue to exist for five hundred years or more, degrading into microplastics that enter the food chain and the water supply and the bodies of organisms that will never know what Styrofoam was.

These entities share five properties that Morton identifies with philosophical precision.

Viscosity. Hyperobjects stick. They adhere to everything they contact. One does not encounter nuclear radiation and walk away unchanged. One does not interact with the smooth cultural logic of frictionless digital platforms and emerge with one's cognitive habits intact. The hyperobject leaves residue. It restructures the entity it touches, and the restructuring is not easily reversed, because the restructured entity is no longer the entity that existed before contact. The person who would "go back" to the pre-contact state no longer exists. A different person, shaped by the contact, would have to make the journey — and that person has different desires, different tolerances, different neurological expectations.

Nonlocality. A hyperobject is not located in any single place. It is distributed across many places simultaneously, and it manifests differently at each location. Climate change is not "in" the Arctic any more than it is "in" the Sahel. It is in both, and in neither, and in every place between. One cannot travel to the hyperobject. One is always already inside it, experiencing a local slice of an entity whose totality is constitutively inaccessible.

Temporal undulation. Hyperobjects involve profoundly different temporalities than those human cognition evolved to process. Evolution calibrated the human nervous system for threats and opportunities operating on timescales of seconds to years. A hyperobject operates on timescales of centuries, millennia, geological epochs. The mismatch between the entity's temporality and the observer's temporality is not a problem to be solved by better instruments. It is an ontological condition — a permanent feature of the relationship between finite perceivers and entities that exceed finitude.

Phasing. Hyperobjects phase in and out of human awareness. Climate change becomes visible during a category-five hurricane and invisible on a mild autumn afternoon. The smooth becomes palpable when a builder notices their capacity for sustained attention has degraded and imperceptible when a compelling project restores the sensation of depth. This intermittent visibility is what makes hyperobjects so resistant to political and institutional response. They are never consistently present enough to demand action and never consistently absent enough to be dismissed.

Interobjectivity. A hyperobject does not exist independently of the entities it affects. It is constituted by its relationships with those entities, and those entities are constituted by their relationships with it. The distinction between the hyperobject and its "environment" collapses, because the hyperobject is the environment — or rather, the hyperobject has made the concept of a stable, background environment untenable.

Now: the central claim of this book.

The transformation of human cognitive life by artificial intelligence is a hyperobject. Not metaphorically. Not loosely, in the way that any sufficiently complex phenomenon might be called "hyperobject-like" if one squints. The AI transformation satisfies Morton's criteria with the same rigor as global warming or nuclear waste. It is massively distributed in time and space — operating simultaneously in data centers in Virginia, smartphones in Lagos, classrooms in Seoul, hospital diagnostic systems in São Paulo, and the cognitive habits of every knowledge worker who has integrated AI tools into a workflow that now feels irreversible. It is viscous — once an organization, a profession, a creative practice, or a human mind has been restructured by AI, the restructuring adheres. It is nonlocal — there is no place where AI "is" in any bounded sense, no single server farm or corporate headquarters that one could visit to see the entity. It is temporally undulant — its effects accumulate across timescales that exceed the quarterly earnings cycle, the electoral cycle, the attention span of any individual user. And it phases in and out of perception with a regularity that defeats both alarm and complacency.

Martin Zeilinger, in a 2022 paper that represents the most rigorous academic application of Morton's framework to AI, mapped these properties systematically. Zeilinger observed that AI "escapes the horizon of human perception and understanding" — that while "human agents can certainly experience the workings and effects of AI whenever it touches down in highly specific ways," none of these specific experiences "will encompass a totality of the functions AI now" performs. The operations and instantiations of AI, Zeilinger continued, are "so massively enmeshed with diverse technologies, places, functions, and purposes" that the shared space of human-perceptible meaning "has come to include many non-human elements," including "computer chips, circuit boards, sensors, data transfer infrastructure." The entity is not the chatbot. It is not the coding assistant. It is not any specific application. It is the total reconfiguration of human cognitive culture by a distributed computational intelligence that exceeds any individual manifestation.

This distinction — between the local manifestation and the hyperobject — is the distinction that most discourse about AI fails to make. The conversation about whether Claude Code will replace programmers is a conversation about a local manifestation. The conversation about whether AI-generated art is "real" art is a conversation about a local manifestation. The conversation about whether a specific language model is conscious, creative, or merely sophisticated autocomplete is a conversation about a local manifestation. Each of these conversations is worth having. None of them engages the hyperobject. The hyperobject is the total transformation — the aggregate effect of AI on attention, identity, work, meaning, education, creativity, political organization, economic structure, and the phenomenological texture of daily life, accumulated across every domain and every timescale simultaneously.

The difficulty is ontological, not informational. The hyperobject exceeds perception not because observers lack data but because the entity is constitutively larger than the perceptual apparatus. Evolution built human cognition to process medium-sized objects at human timescales — a predator in the grass, a face across the fire, a berry that might be poisonous. These capacities served Homo sapiens adequately for seventy thousand years. They do not serve for the perception of entities distributed across the entire computational infrastructure of the planet, operating on timescales that range from the nanosecond processing speed of a GPU to the multi-generational transformation of what it means to think.

Asking a human being to perceive the AI hyperobject is not like asking a near-sighted person to read small print. Glasses can solve that problem. It is like asking a creature that evolved to perceive objects in three spatial dimensions to perceive an object in eleven. The limitation is not in the instrument. It is in the architecture of the observer.

Which raises the question that structures the remainder of this book: If the most consequential entity reshaping human life cannot be perceived by the beings whose lives it reshapes, what follows?

What follows is not despair. Morton is emphatic about this, sometimes almost aggressively so, in a style that oscillates between philosophical rigor and the deliberate provocation of a thinker who believes that ecological awareness requires ontological discomfort. What follows is a different kind of thinking — thinking that proceeds from the acknowledgment that the entity exceeds the apparatus, that certainty is unavailable, and that action must occur anyway, in the absence of the stable ground that previous philosophies assumed.

The thinker who demands to "fully understand" AI before acting on it is making a request that the ontology of hyperobjects reveals to be incoherent. One will never fully understand AI. Not because one is insufficiently intelligent but because the entity is constitutively larger than any intelligence — human or artificial — that attempts to comprehend it. The demand for full understanding is itself a symptom of the ontological framework that hyperobjects have made obsolete: the framework in which a knowing subject stands outside a known object and apprehends it completely.

There is no outside. That is the first lesson, and it is the hardest.

Segal's Orange Pill describes what he calls "the fishbowl" — the set of assumptions so familiar one has stopped noticing them, the water one breathes without recognizing it as water. The metaphor is apt, but Morton's framework presses further. The fishbowl is not merely a set of assumptions. It is a hyperobject — an entity constituted by the total cognitive environment, too large to see, too viscous to escape, too temporally vast to outlast. Pressing one's face against the glass and glimpsing something beyond the water is what Morton would call an uncanny moment: the instant when the hyperobject phases into visibility, when one catches a flicker of the scale of the thing one has been inhabiting without knowing it. The flicker passes. The water closes back over. But something has changed. The entity has been, however briefly, not perceived — it remains imperceptible as a totality — but thought. And thinking a hyperobject, even incompletely, even uncomfortably, is the beginning of what it means to coexist with it rather than be dissolved by it.

The chapters that follow will apply each of Morton's five properties to the AI hyperobject in turn, examining what it means for the entity that is reshaping human cognitive life to be viscous, nonlocal, temporally undulant, phased, and interobjective. The final chapters will turn from properties to practices — from the ontology of what the hyperobject is to the ecology of how to live within it.

The project is uncomfortable by design. Morton has said in recent years that were the original Hyperobjects to be written today, the tone would be different — less intent on alarming, more oriented toward care. "Things are already scary enough," Morton observed. The AI hyperobject does not need a philosopher to make it frightening. What it needs is a philosophy adequate to its scale — one that does not pretend the entity can be mastered, defeated, or fully comprehended, but that nevertheless finds within the condition of being-inside-the-hyperobject the resources for attention, care, and something that might, in Morton's characteristically odd and precise vocabulary, be called coexistence.

The entity is already here. It has been here for some time. The task now is to learn what it means to think inside it.

---

Chapter 2: The Smooth as Hyperobject

Consider a single moment. A person opens a music streaming application. The application, without being asked, presents a playlist titled something like "Made for You" or "Your Daily Mix." The songs are selected by an algorithm that has analyzed the person's listening history, cross-referenced it with the listening histories of millions of other users who share similar patterns, and produced a sequence of tracks calibrated to match the person's preferences with a precision that exceeds, in most cases, the person's own self-knowledge. The person presses play. The music begins. It sounds good. It sounds, in fact, almost exactly like what the person wanted to hear, though the person had not, until the moment of hearing, formulated what they wanted.

This is a small event. It takes three seconds. It involves no apparent coercion, no visible loss, no measurable harm. It is, by every conventional standard, a good user experience. The platform has served its function. The person is satisfied.

Now multiply that moment by every interaction with every algorithm-mediated surface in a single day. The news feed that presents stories calibrated to engagement patterns. The search engine that autocompletes queries before the question has been fully formed. The shopping platform that recommends products based on purchase history and browsing behavior and the purchasing behavior of demographically similar users. The email client that drafts replies. The navigation app that routes around traffic without explaining its reasoning. The social media platform that determines which friends' posts appear and in what order, using criteria that no individual user can inspect or override.

Each interaction is small. Each is, in isolation, convenient. Each removes a small unit of friction — the friction of choosing, of searching, of deciding, of encountering something unexpected. And each, considered individually, is essentially harmless.

The aggregate is something else entirely.

Byung-Chul Han named this aggregate "the aesthetics of the smooth" — a cultural condition in which friction, resistance, and the negativity of encounter with the genuinely other have been systematically eliminated from human experience. Han traces the smooth through Jeff Koons's mirror-polished Balloon Dog sculptures, through the iPhone's featureless glass slab, through Botox and Instagram filters and one-click purchasing, through every surface that has been engineered to offer no resistance to the hand, the eye, or the mind.

Segal, in The Orange Pill, takes Han's diagnosis seriously — seriously enough to spend three chapters examining what the smooth costs. The thinner learning. The atrophied questioning. The erosion of the capacity for surprise. The slow disappearance of the embodied understanding that only friction can build. These costs are real, and Segal is honest enough to name them even as he mounts the counter-argument that friction has not disappeared but ascended.

Morton's framework does something that neither Han nor Segal quite manages: it explains why the smooth is so difficult to perceive, resist, or even coherently discuss.

The smooth is a hyperobject.

Not a metaphor for a hyperobject. A hyperobject. It satisfies Morton's criteria with the same specificity as climate change or nuclear waste. The smooth is massively distributed in time and space — operating simultaneously across every digital platform, every algorithmic surface, every interface on every device in every pocket in every country where connectivity exists. It is viscous — once a mind has been shaped by smooth interactions, the expectation of smoothness adheres; friction, when encountered, registers not as a feature of the task but as a failure of the tool. It is nonlocal — there is no place where the smooth "is." It is not in Cupertino or Mountain View or any specific design studio. It is everywhere that algorithmic mediation occurs, which is to say everywhere. It is temporally undulant — its effects on cognition accumulate so slowly that no individual interaction produces measurable change, while the aggregate transformation, accumulated across years and decades, may be profound. And it phases in and out of awareness with the regularity that makes it impossible to pin down: visible in a moment of cognitive flatness, invisible the instant a compelling task restores engagement.

The difficulty of perceiving the smooth-as-hyperobject is not a failure of attention. It is an ontological condition. The smooth is the medium in which contemporary cognition occurs. Asking a mind shaped by algorithmic smoothness to perceive the smoothness is like asking a fish to describe water — not because the fish is unintelligent but because water is the condition that makes the fish's intelligence possible. The apparatus of perception has been shaped by the entity it would need to perceive. The observer is inside the observed. The instrument is part of the measurement.

This is what separates Morton's analysis from Han's. Han diagnoses the smooth as a cultural pathology. The diagnosis is brilliant — among the most incisive in contemporary philosophy. But a diagnosis assumes the possibility of a diagnostician who stands outside the condition being diagnosed. The doctor is not sick. The critic is not smooth. Han's garden in Berlin — his refusal of smartphones, his analog music, his handwritten prose — is an attempt to maintain that external position, to be the diagnostician who has not contracted the disease.

Morton's ontology denies this possibility. There is no outside the hyperobject. There is no garden that is not also inside the smooth, because the smooth is not located in any device or platform that can be refused. It is the total cognitive condition of a civilization that has organized itself around algorithmic mediation. Han's garden is inside the hyperobject. Han's philosophy is inside the hyperobject. The act of refusing a smartphone is itself shaped by the smooth — it is a gesture that derives its meaning from the condition it refuses, a negation that is constituted by what it negates.

This does not make the refusal meaningless. It makes it local. A local practice of friction within a nonlocal entity of smoothness. The garden is real. The friction of soil is real. The capacity for slow attention that the garden cultivates is real. But the garden does not escape the hyperobject. It creates a microclimate within it — a pocket of different atmospheric conditions that is nonetheless surrounded by, and ultimately permeable to, the larger atmosphere.

The implications for how one thinks about the AI-era smooth are considerable. If the smooth is a hyperobject, then the responses to it must be calibrated to the entity's scale. A response calibrated to a local phenomenon — "avoid this platform," "limit screen time," "use AI less" — addresses a symptom. The entity persists. The smooth does not reside in any individual platform or tool. It is the aggregate condition that emerges from the interaction of every platform and every tool with every mind and every institution simultaneously.

Segal's concept of "attentional ecology" approaches the right scale. An ecology is, by definition, a system of relationships among organisms and their environment, considered as a totality. One does not address an ecological crisis by treating a single organism. One studies the system — its flows, its feedback loops, its leverage points — and intervenes where small perturbations can cascade through the network.

But even attentional ecology, as Segal frames it, tends toward the managerial: structured pauses, sequenced workflows, protected mentoring time. These are valuable practices. They are also local interventions in a nonlocal entity. The hyperobject framework suggests that the smooth cannot be managed. It can only be inhabited — with varying degrees of awareness, varying degrees of care, varying degrees of intentionality about how one allows it to shape the conditions in which one thinks.

This distinction — between managing and inhabiting — is not semantic. It changes what counts as an adequate response. Managing implies a subject who stands outside the system and adjusts its parameters. Inhabiting implies an entity that is inside the system, constituted by it, and attempting to create pockets of different conditions within it while acknowledging that those pockets are permeable and temporary.

The most honest practitioners of AI-augmented work already know this. The builder who recognizes, at three in the morning, that the exhilaration has drained and what remains is compulsion — who catches the moment when flow curdles into grinding momentum — is inhabiting the smooth with a flicker of awareness. The flicker does not extract the builder from the hyperobject. It illuminates, briefly and incompletely, the fact of being inside it.

That flicker is the beginning. Not of escape — there is no escape — but of what Morton calls ecological awareness: the uncomfortable, disorienting, productive recognition that one is inside an entity one cannot fully see, that the entity is reshaping the conditions in which one's seeing occurs, and that the only honest response is to develop practices of attention that do not depend on the fantasy of an outside.

The smooth has no edge. It has no boundary. It has no exterior to which one might retreat. It has only local variations in density — places where the smoothness is more intense, places where friction has been cultivated or preserved. Those variations matter. They are, perhaps, all that matters. But they matter as ecologies within a hyperobject, not as escapes from it.

Han writes from the garden. Morton writes from inside the storm. The difference is not one of courage or intelligence. It is a difference of ontological commitment — a difference in what one believes about the relationship between the observer and the observed. Han believes the observer can, with sufficient discipline, achieve a position from which the smooth is visible as an object of critique. Morton believes the observer is always already inside the smooth, that the smooth is the medium in which the observation occurs, and that the most one can achieve is not clarity but the specific, productive, uncomfortable awareness that clarity is unavailable.

Both positions are valuable. Neither is complete without the other. But for the purpose of understanding why the AI transformation is so resistant to the policy interventions, the institutional reforms, the individual disciplines that well-meaning people propose — why the smooth keeps winning, keeps absorbing resistance, keeps converting opposition into engagement — Morton's position is the more diagnostically powerful. The smooth wins because it is not an opponent. It is the field on which the game is played. One does not defeat the field. One plays on it, within it, shaped by it, with whatever awareness one can cultivate about the conditions that make the game possible.

The music plays. The playlist ends. Another begins. The person has not chosen it. The algorithm has. The person does not notice the transition. The smooth has no seams.

---

Chapter 3: Viscosity — AI Sticks to Everything

There is an experiment that anyone with access to AI coding tools can perform, and which reveals more about the nature of hyperobjects than any philosophical argument. Use Claude Code, or a comparable AI assistant, to build something — a web application, a data visualization, a prototype of a product that has been living in the imagination for years. Spend a week with it. Watch the thing you imagined become real at a speed that feels, the first time, genuinely shocking. Feel the specific pleasure of describing what you want in plain language and watching the machine produce it.

Then stop. Go back to the way things were done before. Open a blank file. Write the code by hand. Debug manually. Read documentation. Wait for Stack Overflow to answer the question.

The experiment will fail. Not because the skills have been lost — at least not yet, not after a week — but because the expectations have been restructured. The pace that felt normal two weeks ago now feels excruciating. The friction that was invisible has become intolerable. The tolerance for delay, for error, for the iterative fumbling that characterizes manual development, has undergone a recalibration so total that the person who sits down to code by hand is not the same person who did it routinely a month ago. That person is gone. In their place is someone who knows what frictionless feels like, and for whom friction now registers as damage rather than process.

This is viscosity. Morton's term for the property of hyperobjects that makes them adhere to everything they contact. One does not encounter a hyperobject and walk away unchanged. The encounter restructures the entity that encounters it — reshaping expectations, tolerances, habits, desires, and the neurological architecture of reward — in ways that make separation not merely difficult but, in a precise philosophical sense, incoherent.

Viscosity operates at every scale. It operates at the neurological level: studies of habit formation and dopamine-mediated reward demonstrate that systems providing immediate, variable, and effort-reducing feedback create expectation patterns that are among the most resistant to extinction in the behavioral repertoire. The builder who receives working code in seconds rather than hours is not merely experiencing convenience. The builder's reward circuitry is being recalibrated around a new baseline of speed, a new minimum threshold for what counts as "responsive," a new standard against which all future tool interactions will be measured.

It operates at the professional level: the engineer whose workflow has been restructured around AI assistance develops not merely new habits but new capabilities — the ability to work across domains that were previously inaccessible, the ability to prototype at speeds that change what it means to "try something." These capabilities are not additive. They are transformative. They change what the engineer can do, what the engineer expects to do, and what the engineer's organization expects the engineer to do. The new baseline is not a preference. It is a structural condition. When the tool is removed, the capabilities it enabled do not gracefully degrade. They collapse, and the person left standing amid the collapse is someone whose professional identity has been built on a foundation that no longer exists.

It operates at the organizational level: once a company has integrated AI tools into its development pipeline, the pipeline itself is restructured. Timelines compress. Team compositions change. The ratio of planning to execution shifts. The expectations of clients, investors, and managers are recalibrated around the new velocity. To remove the tool is not to return to the old pipeline. It is to break the new one, and the old one no longer exists to return to, because the institutional memory of how things were done before has itself been restructured by the period of AI integration.

Segal's account of the Trivandrum training provides a case study in viscosity that Morton's framework illuminates with unsettling precision. In five days, twenty engineers were restructured. Not retrained — restructured. Their sense of what was possible, what was difficult, what constituted a day's work, what skills they possessed and which they lacked: all of it recalibrated. The twenty-fold productivity multiplier was not a number that could be turned on and off. It was a new cognitive and professional state that, once achieved, adhered.

The sentence Segal uses to describe the moment — "I could not tell whether I was watching something being born or something being buried" — captures viscosity's phenomenology precisely. The new state is not separable from the loss of the old one. They are the same event. The adhesion of the hyperobject is simultaneous with the dissolution of what preceded it, and the person experiencing the transition is inside both at once, unable to fully perceive either because the perceptual apparatus is itself undergoing transformation.

The "Help! My Husband is Addicted to Claude Code" episode that Segal documents is viscosity at the intimate scale. The spouse's desperation is not a response to a bad habit that could, with sufficient willpower, be broken. It is a response to a structural transformation in another human being — a transformation in what that person finds rewarding, in what pace of work feels tolerable, in what constitutes a satisfying use of time. The tool has adhered to the person's reward system, creative identity, and sense of professional possibility with such thoroughness that "just stop using it" is not a prescription but a misunderstanding of what has happened. The person who would stop is the person who existed before the tool. That person is no longer available.

This is not addiction in the clinical sense, though the phenomenology overlaps in ways that are worth examining. Addiction, in its classical formulation, involves the pursuit of a stimulus that produces diminishing returns — more of the substance is required to produce less of the effect. Viscosity is different. The hyperobject does not produce diminishing returns. It produces restructured returns — a new landscape of reward in which the old landmarks no longer register. The builder who has experienced AI-augmented creation is not pursuing a diminishing high. The builder is inhabiting a new cognitive environment in which the absence of AI-augmented creation feels not like sobriety but like deprivation.

The distinction matters because it determines what counts as an adequate response. Addiction can, in principle, be addressed by withdrawal: remove the substance, endure the discomfort, return to the baseline. Viscosity cannot be addressed by withdrawal because there is no baseline to return to. The baseline has been consumed by the hyperobject. The person who "withdraws" from AI tools does not arrive at their pre-AI self. They arrive at a third state: someone who has experienced AI augmentation, has been restructured by it, and is now operating without it. That third state is not the first state restored. It is a new condition — characterized by the specific frustration of knowing what frictionless feels like and being denied it.

Morton's observation that hyperobjects "involve a fundamental change in how humans relate to the entities they encounter" applies here with uncomfortable directness. The change is not optional. It is not a setting that can be toggled. It is a consequence of contact with an entity whose viscosity ensures that no contact is temporary.

A philosophical objection might arise at this point, and it is worth addressing. If AI is viscous — if contact with it restructures the entities it touches irreversibly — does that not make AI intrinsically dangerous? Does viscosity not constitute, by itself, an argument for avoidance?

Morton's response to the equivalent objection regarding climate change is instructive. Climate change is viscous. Contact with it is irreversible. The ecosystems it has touched cannot be restored to their pre-contact state. None of this makes avoidance a coherent strategy, because avoidance would require a position outside the hyperobject, and no such position exists. One is already inside it. The question is not whether to have contact but how to inhabit the contact — with what degree of awareness, what practices of attention, what structures of care.

The same logic applies to AI viscosity. The restructuring has occurred. It is occurring. It will continue to occur. The relevant question is not how to prevent it — that question expired somewhere around 2024, if not before — but how to inhabit the restructured landscape with sufficient awareness to cultivate the conditions that sustain genuine thought, genuine questioning, genuine creative production, within an environment that has been permanently altered by contact with a hyperobject that sticks to everything it touches.

What this requires, at minimum, is the abandonment of the fantasy of reversibility — the fantasy that one can "go back" to the way things were, that somewhere on the other side of the AI transformation is the old world, preserved and waiting. The old world is not waiting. The hyperobject has touched it. It has been restructured. The choice is not between the new world and the old one. It is between different ways of inhabiting the new one — with more or less awareness, more or less care, more or less willingness to attend to what has been lost and what has been gained and what the relationship between the two might mean for the beings who must live in the aftermath.

---

Chapter 4: Nonlocality — The Smooth Is Everywhere

In 2012, at a gallery in New York, an artist exhibited a work consisting of a single smartphone placed on a white pedestal. The phone was turned on. Its screen displayed a feed — social media, news, notifications — that updated in real time. Next to the phone, a placard read: "The entire world. Actual size." The work was dismissed by some critics as glib. Others recognized in it a statement about nonlocality that the art world's vocabulary was not yet equipped to process.

The phone was not the world. But the phone was a portal to an entity that was everywhere and nowhere simultaneously — an entity that could not be located in any specific server, any specific data center, any specific corporate headquarters, because it existed only as the aggregate of every connection, every interaction, every algorithmic mediation occurring across every node of a network that spanned the planet. The entity was nonlocal. The phone was a local access point. The mistake the dismissive critics made was to confuse the access point with the entity.

Nonlocality is the second of Morton's five properties of hyperobjects, and it may be the most consequential for understanding why the AI transformation resists every intervention that has been proposed to manage it. A nonlocal entity is not located in any single place. It is distributed across many places simultaneously, manifesting differently at each location, and no collection of local observations — however comprehensive, however precise — adds up to a perception of the totality. One can sample climate change at weather stations around the globe. The samples do not constitute seeing climate change. One can study AI's effects in specific workplaces, specific schools, specific creative practices, specific national economies. The studies do not constitute seeing the AI transformation. The entity exceeds the sum of its local manifestations because it is constituted not merely by the manifestations themselves but by the relationships among them — relationships that are themselves distributed, dynamic, and inaccessible to any observer positioned at any single node.

The smooth, as hyperobject, inherits this nonlocality with a completeness that makes resistance through local action structurally futile. Not morally futile — local practices of friction, local cultivation of depth, local protection of spaces where sustained attention can develop, are meaningful and necessary. But structurally futile in the specific sense that local resistance does not diminish the hyperobject. It creates pockets of different atmospheric conditions within an atmosphere that remains, at every scale larger than the pocket, unchanged.

Han's garden in Berlin. It is a recurring image in discussions of the smooth, and for good reason: it is the most vivid available example of a local resistance practice maintained with absolute philosophical consistency. Han does not own a smartphone. He listens to music in analog. He writes by hand. He gardens. Each of these practices introduces friction into a frictionless cultural landscape. Each cultivates a form of attention that the smooth erodes. Each is genuine, admirable, and instructive.

And each is local.

The smooth does not reside in Han's smartphone — or rather, in the smartphone Han has refused. It does not reside in the streaming platform he does not use. It does not reside in any specific device, platform, or interface. It is the total condition of a civilization that has organized its cognitive life around algorithmic mediation. To refuse the smartphone is to refuse a local access point. The entity persists. It persists in the interfaces that Han's students use, in the algorithmic systems that mediate the publication and distribution of Han's books, in the economic structures that determine which philosophical arguments reach an audience and which do not. It persists in the attention patterns of the people Han speaks to, who arrive at his lectures with minds already shaped by the smooth — minds for which Han's arguments about friction register, paradoxically, as content consumed within the smooth.

This is not a critique of Han. It is a description of what nonlocality means for any attempt to resist a hyperobject. The hyperobject has no exterior. There is no position from which one can act on it from outside. Every intervention occurs within the entity, is shaped by the entity, and is, to some degree, absorbed by the entity. The garden is inside the smooth. The refusal is inside the smooth. The philosophy of friction is inside the smooth — published by academic presses that use algorithmic distribution, discussed on digital platforms that embody the very frictionlessness the philosophy critiques, consumed by readers whose attention has been pre-shaped by the entity the philosophy diagnoses.

Morton calls this condition "the end of the world" — not in the apocalyptic sense but in the ontological sense that the concept of "world" as a stable background against which entities appear has become untenable. There is no world outside the hyperobject. The hyperobject is the world, or rather, it has dissolved the distinction between foreground (the entity one is examining) and background (the stable context in which the examination occurs). When one attempts to examine the smooth, one examines it with cognitive tools that have been shaped by the smooth, in a context that is constituted by the smooth, and the results of the examination are transmitted through channels that are organized according to the smooth's logic.

What does this mean for Segal's project? The Orange Pill proposes dams — structures that redirect the flow of intelligence toward life. The metaphor assumes a river: a force that flows through channels, and a channel is a local structure. A dam works because the river is, in the relevant sense, channeled. It flows through a specific geography. A structure placed at the right point in that geography can redirect the flow.

But the smooth is not channeled. It is everywhere simultaneously. It flows through every channel, every geography, every cognitive environment in which mediated interaction occurs. The question Morton's framework forces is uncomfortable but necessary: What does it mean to build a dam against an entity that has no channel? What does it mean to redirect a flow that is coming from every direction at once?

The answer, if Morton's ontology is taken seriously, is that dams must be reimagined not as barriers in a channel but as practices distributed across the same space as the hyperobject itself. Not a wall placed across the river but a cultivation of different conditions within the atmosphere. An attentional ecology, in Segal's language — but an ecology understood not as a management practice (which implies a manager standing outside the system) but as a way of inhabiting the hyperobject (which implies a being who is inside the system, constituted by it, and attempting to create conditions for flourishing within it rather than escaping from it).

This reframing has practical consequences that are easy to miss if one reads it as merely philosophical. Consider the institutional response to AI in education. The dominant policy model is the dam: prohibit AI use in certain contexts, restrict access to certain tools, create AI-free zones where students are required to do their own thinking. These policies are dams in the channeled-river sense. They assume the smooth flows through specific tools, and that blocking the tools blocks the smooth.

The smooth, being nonlocal, flows around the dam. The student who is prohibited from using Claude in the classroom still inhabits a cognitive environment structured by the smooth — an environment in which every other interaction, every other platform, every other moment of the day has been organized around the elimination of friction. Prohibiting one tool in one context does not restore the capacity for sustained attention that the smooth has eroded across every context. It creates an artificial pocket of friction within a frictionless atmosphere, and the pocket, being artificial rather than intrinsic, is experienced not as depth but as deprivation.

The distributed alternative is harder to describe and harder to implement, precisely because it does not lend itself to the clean legibility of a policy document. It involves cultivating, across the entire educational environment, the conditions that make sustained attention possible — not by prohibiting tools but by creating contexts in which depth is genuinely rewarding, in which friction is experienced not as obstacle but as texture, in which the slow development of understanding through struggle is valued not as a rule imposed from outside but as a practice that the student chooses because the alternative — the smooth — has been made visible as the hyperobject it is.

Making the hyperobject visible — or rather, thinkable, since visibility is precisely what the nonlocal entity denies — is itself a form of pedagogy. The student who can think the smooth, who can recognize the moments when algorithmic mediation is shaping their attention without their consent, who can perceive the local manifestations of the hyperobject as manifestations rather than as the natural texture of reality, has acquired a capacity that no AI prohibition can provide. That capacity — the capacity to think at the scale of the hyperobject while acting at the local scale — is the cognitive posture that the age demands.

Morton calls it ecological awareness. It is not a state of knowledge but a state of attention — an ongoing, uncomfortable, never-completed orientation toward the entity one inhabits. It does not produce certainty. It does not produce mastery. It produces something more modest and more durable: the capacity to act within a condition one cannot fully perceive, with care rather than control, with attention rather than comprehension, with the specific humility of a being who knows that the entity is larger than the apparatus and who builds anyway — not because building will defeat the hyperobject, but because building with awareness of the hyperobject is the only alternative to building in ignorance of it, and the difference between the two, while invisible at the scale of the hyperobject, is everything at the scale of the human life.

The phone sits on the pedestal. The feed refreshes. The entire world, actual size. The entity has no edge, no boundary, no exterior. It has only local variations in density — places where attention thickens, places where care accumulates, places where the smooth, for a moment, becomes visible as the thing it is rather than the air one breathes.

Those places matter. Not because they escape the hyperobject. Because they are the only places where thinking about the hyperobject becomes possible, and thinking is what happens before building, and building — at the right places, with the right awareness, in the full knowledge that the entity will outlast the structure — is the only response that the situation permits.

Chapter 5: Temporal Undulation — Slow Damage

A frog placed in boiling water jumps out. A frog placed in tepid water that is heated gradually does not. The story is apocryphal — actual frogs are more perceptive than the parable credits — but its persistence in popular culture reveals something about the human relationship to slow change that is more accurate about humans than it ever was about amphibians. The nervous system that evolved to detect the predator in the grass, the crack of a branch, the sudden shift in light that means something large is moving, is catastrophically ill-equipped to detect threats that arrive on timescales of months, years, decades. The threat that kills slowly does not register as a threat. It registers as normal. Then it registers as the way things have always been. Then it stops registering at all.

Morton's third property of hyperobjects is temporal undulation — the fact that hyperobjects involve timescales so radically different from human experiential time that they produce what amounts to a perceptual scotoma, a blind spot not in space but in time. The human cognitive apparatus processes the present with extraordinary resolution. It processes the recent past and near future with diminishing but functional fidelity. It processes deep time — centuries, millennia, the 24,100-year half-life of plutonium-239 — not at all. Not poorly. Not approximately. Not at all. Deep time is not a blurry version of experiential time. It is a category of duration that the evolved apparatus cannot represent.

The AI hyperobject undulates across at least three temporal scales simultaneously, and the interaction among these scales is what produces the scotoma that makes the smooth's damage so difficult to detect.

The first scale is the immediate: the nanosecond-to-second timescale of computational inference, the speed at which a language model processes a prompt and generates a response. This timescale is too fast for human perception in the same way that the wing-beat of a hummingbird is too fast — one sees the effect (the hovering) without perceiving the mechanism (the beating). The user experiences the response as instantaneous. The computation that produces it is invisible. This invisibility matters because it contributes to the phenomenology of smoothness: the response appears to arrive without process, without labor, without the resistance that would mark it as produced rather than given.

The second scale is the biographical: the months-to-years timescale across which a human mind is restructured by habitual interaction with AI systems. This is the timescale at which the Berkeley researchers documented their findings — task seepage, work intensification, the colonization of cognitive rest by AI-accelerated activity. These changes are perceptible in retrospect but not in real time. No individual day of AI-augmented work produces a measurable change in cognitive capacity. The change accumulates the way sediment accumulates in a river delta — invisibly, continuously, and with consequences that become apparent only when the delta has shifted enough to redirect the current.

The third scale is the generational: the decades-to-centuries timescale across which the total cognitive infrastructure of a civilization is transformed by the integration of a new form of intelligence. This is the timescale that no individual observer can perceive, because it exceeds the span of an individual life. The generational effects of AI on human cognition are, at this writing, entirely unknown — not because researchers lack interest but because the experiment has been running for less than a decade, and generational effects require generations to manifest.

Morton's point about temporal undulation is that the interaction among these scales produces a specific and dangerous form of imperceptibility. The immediate scale is too fast. The biographical scale is too slow. The generational scale is too vast. The human perceptual apparatus, calibrated for a middle timescale — the seconds-to-years range in which threats and opportunities have historically presented themselves — falls into the gap between all three.

Consider what this means for the specific damage that Segal, following Han, identifies as the cost of the smooth. The thinning of attention. The atrophying of the capacity for genuine questioning. The erosion of embodied understanding — the geological layers of knowledge that Han describes being deposited through friction, through the slow accretion of error and correction that builds intuition.

Each of these processes operates on the biographical timescale. Each is, at any given moment, imperceptible. No single interaction with an AI system measurably thins attention. No single auto-completed sentence measurably erodes the capacity for original thought. The damage — if damage is the right word, and the temporal undulation of the hyperobject makes even that determination uncertain — is cumulative, distributed across thousands of interactions over months and years, and at no point does it cross a threshold that the nervous system registers as alarm.

The Berkeley researchers captured a snapshot of this process at a specific moment in its unfolding. Workers using AI tools were working more intensely, taking on more tasks, experiencing task seepage into previously protected cognitive spaces. The researchers interpreted these observations as early signs of a pattern that might, over time, produce burnout and cognitive degradation. The interpretation is plausible. It is also, necessarily, incomplete, because the observation captures a single point on a curve whose trajectory is unknown.

Here is the temporal undulation problem stated as precisely as possible: The damage the smooth produces may be real, cumulative, and profound, and it may nevertheless be invisible at every point along the curve. Not invisible because no one is looking. Invisible because the temporal scale of accumulation falls outside the resolution of human experiential time. By the time the aggregate is large enough to perceive, the perceiver has been transformed by the accumulation — has become a different cognitive entity, with different baselines, different expectations, different standards for what counts as depth, attention, or genuine understanding.

This is the scenario that haunts any honest assessment of AI's cognitive effects. The possibility that something is being lost, slowly and continuously, and that the loss is reshaping the apparatus that would need to detect the loss, so that by the time the loss is detectable, it will no longer register as loss. It will register as normal. It will register as the way things have always been.

Morton, characteristically, does not flinch from this possibility. The philosopher who coined the term "dark ecology" — the ecology that embraces discomfort, uncertainty, and the uncanny rather than retreating to comfortable narratives of control — is not in the business of providing reassurance. The temporal undulation of hyperobjects is a genuine problem, not a solvable one but an inhabitable one, and the difference between those two categories is the difference between the engineering mindset (identify the problem, design the solution, deploy the fix) and the ecological mindset (identify the condition, develop practices for living within it, accept that the condition will not be resolved within the span of a human life).

Segal's response to the temporal undulation problem — the construction of dams, the cultivation of attentional ecology, the deliberate preservation of spaces where friction can do its developmental work — is a response calibrated to the biographical timescale. These practices operate within the span of an individual career, an individual organization, an individual family. They are valuable precisely because the biographical timescale is the one at which human agency operates. One cannot act on the generational timescale. One can only act now, in the hope that now-actions accumulate into generational effects.

But Morton's framework adds a dimension that Segal's pragmatism tends to bracket: the possibility that the generational timescale will produce effects that no biographical intervention can prevent. The possibility that the transformation of human cognitive infrastructure by AI is not a problem amenable to intervention but a process of the same order as the transformation of human cognitive infrastructure by writing, by printing, by electrification — a transformation so total that the question of whether it constitutes "damage" can only be answered by the beings who emerge from it, and those beings will, by definition, have different standards for what counts as damage.

This is not a counsel of despair. It is a description of the temporal structure of the situation. The situation is that humanity is inside a hyperobject whose temporal undulations exceed the perceiver's resolution, whose effects are accumulating on timescales that fall between the cracks of experiential awareness, and whose ultimate consequences will be assessed by beings who have been shaped by the very process they are assessing.

Living honestly inside that description — neither pretending the damage is not occurring nor pretending that biographical interventions will prevent the generational transformation — is the specific form of temporal awareness that Morton's philosophy cultivates. It is awareness without mastery. Attention without control. The willingness to build the dam knowing that the river operates on a timescale that will outlast not merely the dam but the species that built it.

The frog in the gradually heating water is a bad parable about frogs. It is an uncomfortably accurate parable about humans inside a hyperobject whose temporal undulations are calibrated, as if by design, to fall exactly outside the range of the apparatus that would need to detect them.

The water is warming. The instruments that would measure the warming are themselves immersed in the water. The readings they return are shaped by the temperature that surrounds them. This is not a solvable problem. It is an inhabitable one. The difference is everything.

---

Chapter 6: Phasing — Appearing and Disappearing

There are two kinds of Tuesday. There is the Tuesday when the work flows, when the collaboration with AI produces something genuinely surprising, when the mind operates at a pitch of engagement that Csikszentmihalyi would recognize as flow and that feels, from the inside, like the best version of what human cognition can do when augmented by a tool that meets it at its own level. On this Tuesday, the smooth is invisible. Not absent — it is never absent — but invisible, the way gravity is invisible to a person standing on solid ground. The ground holds. The work feels meaningful. The questions feel genuine. The capacity for sustained attention is intact, or at least feels intact, which from the inside is the same thing.

Then there is the other Tuesday. The one where the work feels thin. Where the prompts feel mechanical. Where the responses from the machine, which last week felt like genuine intellectual partnership, feel this week like sophisticated autocomplete — plausible without being insightful, fluent without being true. On this Tuesday, the smooth phases into visibility. One catches a glimpse of the flatness. The attention feels shallow. The questions feel narrow. The satisfaction of productivity feels hollow — not because nothing was produced but because the production occurred without the internal resistance that marks genuine thought.

These two Tuesdays are not different days. They are different phases of the same hyperobject. The smooth does not hold steady. It undulates, phases, flickers. It is present on both Tuesdays — constituting the cognitive environment in which both experiences occur — but it manifests differently, revealing different aspects of itself at different moments, the way a three-dimensional object passing through a two-dimensional plane reveals different cross-sections at different points of its transit.

Morton's fourth property of hyperobjects is phasing: the fact that hyperobjects appear and disappear from human perception without warning or regularity. Climate change phases into visibility during a catastrophic wildfire season and phases out during a mild spring. Nuclear radiation phases into visibility during a reactor accident and phases out during the decades of invisible contamination that follow. The phasing is not a feature of the observer's attention. It is a feature of the hyperobject's relationship to the observer's perceptual apparatus. The hyperobject is always there. But "there" is a spatiotemporal location that the hyperobject, being massively distributed, does not consistently occupy from the perspective of any single observer.

Applied to the AI-era smooth, phasing explains something that most analyses of AI's cognitive effects fail to account for: the intermittent quality of the experience. The smooth is not a constant drone. It is not a steady erosion that can be tracked on a graph. It flickers. It appears and disappears. It is devastating on Monday and invisible on Wednesday. It produces, in the people who inhabit it, a characteristic oscillation that Segal captures in a phrase that recurs throughout The Orange Pill: "terror and awe, sometimes in the same minute."

That oscillation is not emotional instability. It is the phenomenological signature of hyperobject phasing. The terror is the moment when the smooth phases into visibility — when the builder catches a glimpse of the scale of the entity, the depth of the restructuring, the irreversibility of the transformation. The awe is the moment when the smooth phases out — when the capability feels real, the augmentation feels genuine, the work feels meaningful. The oscillation between the two is not a problem to be resolved through better self-knowledge or more disciplined time management. It is the characteristic experience of a finite perceiver inside a hyperobject that reveals itself intermittently and withdraws before it can be fully grasped.

The withdrawal is the crucial concept. Morton, drawing on Graham Harman's object-oriented ontology, argues that all objects — not just hyperobjects, but all objects — are fundamentally withdrawn. They exceed the perceptions and relations that constitute our access to them. The cup on the desk is withdrawn: one perceives its color, its shape, its weight, its temperature, but the cup-in-itself exceeds any and all of these perceptions. The sum of everything one can know about the cup does not exhaust the cup. Something always remains behind, inaccessible, withdrawn.

Hyperobjects make this withdrawal dramatically apparent because their scale amplifies the gap between access and entity. One accesses the smooth through its local manifestations — a flattened attention span here, a hollow productivity there — but the smooth-in-itself withdraws. It is never fully available. It shows a face and then turns, showing a different face, and then turns again, and no sequence of faces adds up to the entity. The entity is larger than the sum of its appearances. This is not a poetic statement. It is an ontological one, with practical consequences.

The practical consequence that matters most for the discourse around AI is this: phasing defeats stable assessment. It is not possible to arrive at a fixed judgment about the AI transformation — "it is good," "it is bad," "it is a net positive with manageable costs," "it is a catastrophe in slow motion" — because the entity on which the judgment is based does not hold still long enough for the judgment to stabilize. The person who declares, on the good Tuesday, that AI is the most generous expansion of human capability since the invention of writing is not wrong. The person who declares, on the bad Tuesday, that AI is eroding the cognitive foundations on which human flourishing depends is not wrong either. Both are responding to genuine phases of a genuine entity. Both are mistaking a phase for the totality.

This is why the debate about AI is so peculiarly resistant to resolution. The optimists and the pessimists are not looking at different evidence. They are looking at different phases of the same hyperobject. The evidence does not settle the question because the evidence itself phases — the same tool, the same user, the same task can produce flow on Monday and compulsion on Tuesday, depth on Wednesday and shallowness on Thursday, and the difference is not attributable to any variable the user can identify or control. The hyperobject phases. The experience follows. The judgment oscillates.

Segal names this oscillation with characteristic honesty. His account of the Trivandrum training captures the phasing with almost clinical precision — the excitement of the engineers who were recalibrating their sense of the possible, the terror of the senior engineer who was watching his identity dissolve, the exhilaration and the dread coexisting in the same room, in the same person, sometimes in the same sentence. "I could not tell whether I was watching something being born or something being buried." The sentence is not ambivalent. It is precise. It describes the phenomenology of hyperobject phasing with a fidelity that no single-valence assessment — positive or negative — could achieve.

Segal's response to the phasing is to propose the beaver's ethic: build anyway, build with awareness, build from inside the river rather than from the fantasy of standing on the bank. Morton's response is complementary but different in emphasis. Where Segal emphasizes action — build the dam, tend it, maintain it — Morton emphasizes attention. Not attention as a resource to be managed, but attention as a practice of staying with the phasing, staying with the oscillation, refusing the temptation to resolve the oscillation into a stable position.

The temptation is enormous. The human mind craves stable ground. The oscillation between terror and awe is cognitively and emotionally expensive, and the mind will seize on any available narrative that promises to end it. The narrative that says "AI is transformative and the future is bright if we build responsibly" ends the oscillation on the upswing. The narrative that says "AI is eroding everything that makes us human" ends it on the downswing. Both narratives provide relief. Both are, in Morton's framework, forms of what he calls "beautiful soul syndrome" — the aesthetic pleasure of having arrived at a clean position that exempts one from the discomfort of continuing to attend to an entity that will not hold still.

The alternative — Morton's alternative — is to stay in the oscillation. To attend to the phasing without resolving it. To hold the terror and the awe simultaneously, not as a psychological achievement but as a perceptual practice. The entity phases. The attention follows the phasing. The judgment remains provisional, incomplete, uncomfortable, and — precisely because of its provisionality — adequate to the entity in a way that no fixed judgment can be.

This is not indecision. Indecision is the failure to act because the grounds for action are uncertain. What Morton describes is something different: action that proceeds from the acknowledgment that the grounds will always be uncertain, because the entity on which the grounds depend is constitutively phased. Building the dam while the river changes shape. Tending the ecology while the atmosphere shifts. Acting, with full awareness that the entity will reveal a new face tomorrow, and that tomorrow's face may require a different action, and that the willingness to revise — to let the judgment oscillate as the entity phases — is not weakness but the specific form of strength that hyperobject-scale entities demand.

The smooth appears on Tuesday. It disappears on Wednesday. It reappears on Friday in a different form. The task is not to fix its location — location is precisely what the nonlocal, phasing entity denies — but to develop the attentional stamina to track its appearances without mistaking any single appearance for the whole.

That stamina is rare. It is also, in the age of hyperobjects, the most important cognitive capacity a human being can possess. Not because it solves the problem. Because the problem is not solvable. And the only alternative to solving it is inhabiting it, which requires exactly the kind of sustained, uncomfortable, never-resolved attention that the smooth is so efficient at eroding.

The recursion is the point.

---

Chapter 7: Interobjectivity — Between Human and Machine

There is no outside.

This sentence has appeared, in various forms, several times in the preceding chapters. It has been stated as an ontological principle (there is no position external to the hyperobject from which to observe it), as a perceptual limitation (the apparatus of observation is itself shaped by the entity it would observe), and as a practical constraint (interventions occur within the entity, shaped by its conditions, and are, to some degree, absorbed by it). Now it must be stated as a constitutive condition — a description not merely of the observer's limitations but of the fundamental structure of reality as Morton's object-oriented ontology describes it.

Morton's fifth property of hyperobjects is interobjectivity: the condition in which entities do not exist independently but are constituted by their relationships with other entities. A hyperobject is not an independent thing that impinges on other independent things. It is a pattern of relationships — and the entities it "affects" are themselves patterns of relationships that include the hyperobject as a constitutive element. The distinction between the entity and its environment collapses, because the entity is the environment, and the environment is the entity, and neither can be isolated from the other without destroying what makes each what it is.

The philosophical roots of this concept run through object-oriented ontology, or OOO — the school of thought developed by Graham Harman and taken up by Morton, Ian Bogost, and others, which holds that objects are always more than their relations (they are "withdrawn," in Harman's terminology) while simultaneously being constituted in part by those relations. This sounds contradictory, and the contradiction is, in a sense, the point. Objects are both withdrawn and relational. They exceed their relations while being partly constituted by them. The tension between withdrawal and relation is not a problem to be resolved. It is the structure of reality.

Applied to the relationship between humans and AI, interobjectivity produces a description that is more unsettling and more precise than any of the standard framings.

The standard framings all assume separability. "Humans use AI tools." Subject, verb, object. The human is the agent. The AI is the instrument. The relationship is one of use — the human picks up the tool, employs it for a purpose, and puts it down. The human remains what the human was before the tool was picked up. The tool remains what the tool was before the human employed it. Each retains its identity across the interaction.

Interobjectivity denies this. The human who uses AI is not the human who existed before the use. The use restructures the user — reshaping cognitive habits, attention patterns, creative expectations, professional identity, neurological reward baselines. The tool, in turn, is not what it was before the human used it — or rather, the tool-in-use is a different entity from the tool-in-potential, because its behavior is shaped by the specific inputs, the specific prompts, the specific creative trajectory of this particular user, and the outputs it produces are not outputs of "the tool" in isolation but outputs of the human-tool system, which is an entity that did not exist before the interaction and cannot be decomposed into its components without destroying what it produces.

This is what Segal discovers in his account of writing The Orange Pill with Claude. "Neither of us owns that insight," he writes, describing a moment when the collaboration produced a connection that neither he nor Claude could have produced independently. "The collaboration does." The statement is precisely correct, and Morton's framework explains why. The insight belongs to the interobjective system — the human-AI entity that is constituted by the relationship between its components and that produces cognitive effects that are not attributable to either component in isolation.

The implications are radical, and they cut against the grain of virtually every ethical framework that has been proposed for AI governance. Those frameworks assume separability: The human is responsible. The AI is a tool. The tool does what the human directs. If the tool produces harm, the human is accountable. If the tool produces value, the human is the creator.

Interobjectivity complicates this tidy allocation. If the human and the tool constitute each other — if the human's creative trajectory is shaped by the tool's capabilities, and the tool's outputs are shaped by the human's inputs, and neither can be meaningfully isolated as the "cause" of the result — then the question of authorship, responsibility, and accountability becomes a question about a system rather than about its components. Asking "who wrote this book?" when the book was produced by a human-AI interobjective system is like asking "who grew this tree?" when the tree was produced by the interobjective system of seed, soil, water, sunlight, and microbiome. The question assumes a separability that the ontology denies.

Segal himself recognizes this, with characteristic honesty, in his account of the moments when Claude produced connections he had not seen. "I cannot honestly say it belongs to either of us," he writes. "It belongs to the collaboration, to the space between us, and I do not have a word for that kind of ownership." The word Morton would offer is interobjectivity — the condition in which ownership, authorship, and agency are properties of relationships rather than of entities.

But interobjectivity is not merely a philosophical abstraction about authorship. It has consequences for the most practical question the AI moment raises: What is the human contribution? Segal frames this question as the existential core of The Orange Pill — the question a twelve-year-old asks when she watches a machine do her homework better than she can: "Mom, what am I for?"

Morton's framework transforms this question in a way that is initially disorienting and ultimately, perhaps, liberating. If the human and the AI constitute each other — if neither exists as an independent agent but only as a node in an interobjective system — then the question "What am I for?" cannot be answered by identifying capacities that are uniquely and exclusively human. There may be no such capacities, or there may be, but the question of their existence is secondary to the more fundamental observation that human value, in an interobjective ontology, does not depend on unique capacities. It depends on the specific quality of relationship that this particular node contributes to the system.

The human is not valuable because the human can do something the machine cannot. The human is valuable because the human occupies a position in the interobjective mesh that no other entity occupies — a position constituted by a specific biography, a specific set of embodied experiences, a specific pattern of care and attention and mortality that shapes every input the human contributes to the system. Remove the human, and the system produces different outputs — not necessarily worse outputs, not necessarily better, but different, in ways that reflect the absence of the specific relational quality that the human's position in the mesh contributed.

This is a different basis for human value than the one Segal proposes — the capacity for genuine questioning, the consciousness that wonders, the candle in the darkness of an unconscious universe. Morton's basis is relational rather than essential. It does not depend on identifying something that humans have and machines lack. It depends on recognizing that the human's contribution to the interobjective system is constitutive — that the system is different, in every case, because of the specific human who participates in it, and that this difference is valuable not because of what it enables the system to produce but because it is the difference itself.

This may sound abstract. It is concrete enough to be tested. Consider two builders, both using Claude Code, both producing functional software. One contributes to the interobjective system a deep understanding of a specific user population, a specific aesthetic sensibility, a specific set of values about what technology should do in the world. The other contributes generic prompts, generic specifications, generic values. Both produce working software. The software is different. The difference is attributable not to the capabilities of the tool (the tool is the same in both cases) but to the relational quality of the human's participation in the interobjective system. The quality of the input shapes the quality of the system, which shapes the quality of the output.

Segal captures this with a formula: "Are you worth amplifying?" Morton would restate it: "What do you contribute to the mesh?" The questions are not identical, but they converge on the same practical implication — that the human's responsibility, in an age of interobjective human-AI systems, is to bring to the relationship the specific quality of attention, care, judgment, and embodied experience that only this particular human, in this particular position in the mesh, can bring. Not because the human is the master and the tool is the servant. Not because the human is the creative origin and the tool is the instrument. But because the interobjective system produces its best outputs when every node contributes its specific quality, and the human's specific quality — shaped by mortality, by embodied experience, by the capacity for genuine questioning that arises from having stakes in the world — is irreplaceable not in the sense that no other entity could occupy the position, but in the sense that any other entity would occupy it differently, and the difference matters.

There is no outside. The human is inside the system. The system is inside the human. The relationship is constitutive. And the quality of what emerges depends, as it always has, on the quality of what each node brings to the mesh.

The question is not whether to participate — participation is not optional when the entities are interobjective. The question is what quality of participation to bring. And that question, at least, is one that a human being can answer. Not from outside the system. Not from a position of mastery or control. But from inside, with the specific, mortal, embodied attention that is the human's contribution to a mesh that, without it, would be a different thing entirely.

---

Chapter 8: The Ecological Thought Applied to AI

Everything is connected. This statement sounds like a bumper sticker. In Morton's hands, it becomes one of the most rigorous and discomforting ideas in contemporary philosophy — an idea whose implications extend far beyond environmentalism into the heart of what it means to think about artificial intelligence.

The ecological thought, as Morton defines it in the 2010 book of that title, is not a thought about ecology. It is a form of thinking — a cognitive posture, an orientation of attention — that takes interconnectedness as its fundamental principle and follows the implications of that principle wherever they lead, even when they lead to places that are uncomfortable, weird, or contrary to the thinker's preferences. The ecological thought does not romanticize nature. It does not oppose nature to culture, wilderness to civilization, the organic to the artificial. It insists that these oppositions are themselves obstacles to ecological awareness, because they divide the mesh — the interconnected web of relationships among entities — into categories that the mesh does not respect.

The mesh is Morton's central image: "the interconnectedness of all living and non-living things, consisting of infinite connections and infinitesimal differences." Not a hierarchy. Not a system with a center. Not a network with a designer. A mesh — a web of relationships so dense, so entangled, so recursively constituted that no node can be understood in isolation, and no intervention at any node can be contained. Pull one thread and the entire fabric shifts. Perturb one relationship and the perturbation propagates — not linearly, not predictably, but through cascading interactions that produce effects at distances and timescales that exceed the intervener's capacity to predict.

Applied to AI, the ecological thought reveals something that the standard discourse — organized around the relationship between humans and machines, as if those were the only two nodes in the mesh — systematically obscures. AI is not a technology inserted into an otherwise unchanged world. It is a perturbation in the mesh of relationships that constitutes cognitive culture. And the mesh is vast. It includes not merely the human user and the AI tool but the educational institutions that trained the human, the data that trained the model, the economic structures that funded the development, the energy infrastructure that powers the computation, the social norms that determine how the tool is used, the political systems that regulate (or fail to regulate) its deployment, the cultural narratives that frame its meaning, the children who grow up inside its effects, the ecosystems that bear the environmental cost of its operation, and the relationships among all of these — relationships that are themselves in flux, themselves being restructured by the perturbation they transmit.

Segal demonstrates this interconnectedness, perhaps without fully recognizing the scale of what he is demonstrating. His argument traces a chain: AI changes how code is written, which changes what it means to be a developer, which changes organizational structure, which changes the economics of software, which changes the value of expertise, which changes what parents tell their children about the future, which changes how children orient themselves toward learning, which changes the educational system, which changes the workforce, which changes the economy, which changes the tools that are built, which changes how code is written. The chain is a loop. Or rather, it is a mesh — because the causal connections are not sequential but simultaneous, each node affecting every other node at every moment, with the speed and complexity of the interactions exceeding the capacity of any observer to trace them.

This is what the ecological thought, applied to AI, actually demands: the recognition that every question about AI is simultaneously a question about everything else. The question "Will AI replace programmers?" is also a question about education, economics, identity, family structure, mental health, urban planning, immigration policy, and the philosophical status of consciousness. These are not separate questions that happen to be related. They are facets of a single perturbation propagating through a mesh, and the attempt to address any one of them in isolation — to regulate AI in the workplace without considering its effects on education, to reform education without considering its effects on the economy, to restructure the economy without considering its effects on identity and mental health — is to misunderstand the structure of the situation at the ontological level.

Morton insists that the ecological thought is weird. The word is chosen deliberately. "Weird" comes from the Old English wyrd, meaning fate or destiny, and Morton uses it to describe the uncanny quality of ecological awareness — the feeling of being inside a system that is vaster than perception, more entangled than comprehension, and more strange than any narrative framework designed for human-scaled problems can accommodate. The AI transformation is weird in precisely this sense. It is uncanny. It produces the specific cognitive vertigo of encountering an entity that is simultaneously intimate (the chatbot responds to natural language) and alien (the computational substrate bears no resemblance to the biological processes that produce human thought).

The standard response to weirdness is to domesticate it — to translate it into familiar categories, to analogize it to previous technological transitions, to narrate it as a continuation of the story of human progress. "AI is like the printing press." "AI is like electricity." "AI is like the Industrial Revolution." Each analogy captures something real. Each simultaneously domesticates what is genuinely weird about the current moment — the fact that the new participant in the cognitive mesh is not merely a tool but an entity that produces outputs that are, in specific and measurable ways, indistinguishable from the outputs of human cognition, and that this indistinguishability destabilizes categories (human vs. machine, natural vs. artificial, created vs. generated) that the previous analogies assumed were stable.

Morton would insist on staying with the weirdness. Not resolving it into comfortable analogy. Not translating it into the vocabulary of previous transitions. Staying with the specific, uncanny quality of a moment in which the mesh has acquired a new kind of node — a node that is neither human nor not-human, neither intelligent nor not-intelligent, neither creative nor not-creative, but something for which existing categories are inadequate and for which new categories have not yet been developed.

The strange stranger is Morton's term for the entity that resists categorization — that cannot be assimilated into the mesh of familiar relationships without disturbing the mesh. The strange stranger is not the foreign entity that can be understood through anthropological distance. It is the entity that is both familiar and alien, both intimate and incomprehensible, both inside the mesh and disruptive to it. AI is a strange stranger in the mesh of cognitive culture. It speaks in human language. It produces outputs that resemble human thought. It participates in creative processes in ways that feel, from the inside, like genuine collaboration. And it is, at the same time, radically other — a computational process running on silicon substrates, trained on patterns extracted from billions of human utterances, producing outputs through mechanisms that bear no resemblance to the neurological processes that produce human cognition.

The temptation is to resolve the strangeness in one of two directions. One direction: anthropomorphize the AI, treat it as a colleague, attribute to it intentions, preferences, a form of understanding. The other direction: mechanize it, reduce it to a tool, insist that it is merely pattern-matching, merely statistical recombination, merely sophisticated autocomplete. Both resolutions are comfortable. Both are, in Morton's framework, forms of foreclosing on the strangeness — of refusing the weird in favor of the familiar.

The ecological thought, applied to AI, refuses both. It insists on inhabiting the strangeness — on maintaining the uncomfortable awareness that the new node in the mesh is genuinely strange, that the categories available for understanding it are inadequate, and that the adequacy of one's response depends not on resolving the strangeness but on developing the capacity to live with it.

This is what "ecology without nature" means in the AI context. Not an ecology that excludes the natural world, but an ecology that refuses to stabilize the distinction between natural and artificial, organic and computational, human and machine — because these distinctions, however useful they have been historically, are obstacles to perceiving the mesh as it actually is: a web of interconnected entities, some biological, some computational, some institutional, some cultural, all constituting each other, all propagating perturbations through relationships that extend in every direction.

Segal's Orange Pill traces one set of perturbations — the ones that radiate outward from the introduction of AI coding tools into software development. Morton's ecological thought suggests that this set of perturbations is a tiny fraction of the total, and that the total — the full mesh of effects propagating through every relationship in every domain — is the hyperobject that the preceding chapters have been attempting to describe.

The perturbation that starts with a developer in Trivandrum producing code at twenty times the previous rate does not stop at the developer's keyboard. It propagates through the organization (team structures change, hiring patterns shift, timelines compress). Through the industry (the Software Death Cross, the repricing of code as commodity). Through the educational system (what do computer science programs teach when code is abundant?). Through the family (the twelve-year-old's question, "What am I for?"). Through the political system (which constituencies benefit, which are displaced, what new forms of power emerge?). Through the cognitive environment (what happens to attention when the friction of implementation disappears?). Through the mesh. In every direction. At every timescale. With consequences that exceed the capacity of any observer to predict.

Everything is connected. The bumper sticker is true. And the truth of it is not comforting. It is weird. It is uncanny. It is the specific vertigo of recognizing that one is inside a mesh so vast and so entangled that the distinction between acting and being acted upon, between perturbing the system and being perturbed by it, between building the dam and being shaped by the river, has dissolved.

The ecological thought does not tell you what to build. It tells you that whatever you build will propagate through the mesh in ways you cannot predict, affecting entities you cannot see, at timescales you cannot comprehend. It does not paralyze action. It demands a different quality of action — humbler, more provisional, more attentive to the mesh, more willing to revise when the effects that propagate back from the mesh are different from what was intended.

The word for that quality of action, in Morton's vocabulary, is care. Not care as sentiment. Care as practice — the ongoing, never-completed, always-provisional attention to the mesh that acknowledges the mesh's vastness and acts anyway, within its own finitude, with whatever awareness it can cultivate of the consequences that ripple outward from its choices.

The mesh does not care about intention. It propagates effects. What Morton's ecological thought adds to Segal's project is the insistence that the effects will propagate further, faster, and more strangely than any builder — however careful, however aware, however committed to the beaver's ethic — can anticipate. And that building with awareness of this fact is different, in every important respect, from building without it.

Chapter 9: Dark Ecology and the Builder's Complicity

Morton tells a story about Blade Runner. Not the story of the film — the story of the loop. Deckard is a blade runner, a hunter of replicants, artificial beings indistinguishable from humans. His job is to find them and destroy them. Then, across the course of the film (and with greater explicitness in the director's cut), a possibility emerges that Deckard himself might be a replicant. The hunter might be the hunted. The agent of destruction might be the thing he has been ordered to destroy.

Morton calls this an ecological loop. The detective discovers that the case leads back to himself. The ecologist discovers that the pollution she is tracking was produced by the civilization she inhabits. The critic discovers that the system she diagnoses is the system that produced her capacity for diagnosis. There is no position of innocence. There is no clean hands. There is only the loop — the uncomfortable, recursive, never-resolved recognition that the investigator and the investigated are the same entity, or at least so entangled that the distinction between them cannot bear the weight that moral clarity requires.

Dark ecology is Morton's name for the philosophical practice of inhabiting this loop without pretending it resolves. Not dark in the sense of evil or hopeless. Dark in the sense of a room where the lights are off and the eyes have not yet adjusted — a space where familiar coordinates are unavailable and movement must proceed by feel, by attention, by the willingness to stumble and correct without the reassurance of knowing where the walls are. Dark ecology is the ecology that does not pretend the ecologist stands outside the ecosystem. It is the ecology that begins with the admission: the crisis is not something that happens to us from outside. We are the crisis. We are the emitters, the consumers, the builders. The system that produces the damage and the system that seeks to repair it are the same system.

Segal makes this admission in the opening pages of The Orange Pill. "I built some of the systems that create it," he writes, referring to the attention economy, the engagement loops, the algorithmic architectures that capture human attention and hold it beyond the point of consent. The admission is striking because it is rare. Most technology critics write from outside the industry. Most industry builders write without acknowledging the costs of what they have built. Segal does both — builds and critiques, participates and diagnoses — and the tension between these positions is not a flaw in the book. It is the book's most honest feature.

Morton's framework explains why.

The builder's complicity is not a moral failure to be confessed and absolved. It is an ontological condition — a structural feature of the relationship between any entity and the hyperobject it inhabits. The smooth is not produced by "the tech industry" as an entity external to the people who diagnose it. The smooth is produced by the total system of interactions among builders, users, platforms, algorithms, institutions, cultural norms, and cognitive habits. Everyone inside the system contributes to the system. Everyone who diagnoses the system does so from within the system, using cognitive tools that have been shaped by the system. The loop is inescapable — not because escape requires superhuman effort, but because there is no outside to escape to.

This does not mean that everyone is equally complicit. Morton is careful about this, and the care matters. The CEO who designs an engagement-maximizing algorithm and the user who checks the resulting feed forty times a day are both inside the loop, but they are not in the same position within it. Power is real. Agency is real. The capacity to make choices that affect the mesh is unevenly distributed. The builder who understands how dopamine-mediated reward loops work and deploys them anyway bears a different kind of responsibility than the user who is caught in the loop without understanding its mechanics.

But — and this is the dark ecological turn — the builder's understanding does not exempt the builder from the loop. It deepens the builder's entanglement in it. To understand the mechanics of the smooth and to build within it anyway is a more complex form of complicity than to be caught in the smooth without understanding it. The understanding adds a layer to the loop. One is no longer merely inside the system. One is inside the system knowing one is inside it, which changes the character of the being-inside without changing the fact of it.

Segal describes this layered complicity with the specificity of someone who has lived it. The product he built early in his career, the one he knew was addictive by design — "I understood the engagement loops, the dopamine mechanics, the variable reward schedules, the social validation cycles" — is a case study in dark ecological entanglement. The understanding was present. The building proceeded anyway. The justification was the one that every builder trapped in the loop generates: "Someone else will build it if I do not, so it might as well be me." The justification is not exactly wrong — the counterfactual is real, the competitive pressure is real — but it functions within the loop as a mechanism for converting understanding into continued participation. The loop absorbs the understanding. The understanding becomes part of what the loop produces.

Morton would say: this is not a problem with Segal. This is a property of hyperobjects. The loop is the structure. The understanding does not break the loop. It makes the loop visible, which is different, and the difference matters, but it does not matter in the way that the Enlightenment tradition assumes — as the prelude to mastery, to a position of rational control from which the problem can be managed. It matters as the precondition for a different quality of being-inside: darker, more honest, more attuned to the consequences of one's own participation.

Dark ecological awareness, applied to the AI moment, produces a set of recognitions that are individually uncomfortable and collectively necessary.

The first recognition: the people building AI systems are not external to the transformation those systems produce. They are inside it. Their cognitive habits, their creative expectations, their professional identities have been restructured by the tools they build. The engineer at Anthropic who uses Claude to assist in building the next version of Claude is inside a loop that makes Deckard's situation look straightforward. The tool is building the tool. The builder is being built by what they build.

The second recognition: the people critiquing AI are not external to the transformation either. The philosopher who writes about the dangers of algorithmic mediation writes on a computer, publishes through algorithmically mediated platforms, reaches an audience whose attention has been shaped by the smooth, and produces arguments that are consumed as content within the very system the arguments diagnose. The critique is inside the loop. It does not escape the loop by being correct.

The third recognition: the people attempting to regulate AI are not external to the transformation they seek to regulate. The policymaker who drafts AI governance legislation does so within institutions that are themselves being reshaped by AI — using AI tools for research, analysis, and drafting; responding to constituencies whose understanding of AI is shaped by algorithmically mediated information; operating within economic structures that are incentivized by the AI industry's lobbying apparatus. The regulation is inside the loop. It does not transcend the loop by being well-intentioned.

The fourth recognition, and the one that matters most: the parents who are trying to protect their children from the cognitive effects of AI are not external to the environment that produces those effects. The parent who worries about their child's attention span has their own attention span shaped by the same smooth. The parent who sets screen-time limits does so within a household where the parent's own relationship with screens models a set of behaviors that the limits cannot overwrite. The protection is inside the loop. It does not escape the loop by being motivated by love.

Dark ecology does not respond to these recognitions with paralysis. It responds with a shift in the quality of action. The shift is from action-as-mastery to action-as-care. Mastery assumes a position of control: identify the problem, design the solution, implement the fix. Care assumes a position of entanglement: recognize the loop, attend to one's own position within it, act with awareness that the action is itself part of the system it seeks to change, and maintain the attention necessary to observe how the system absorbs and transforms the action.

Segal's beaver metaphor captures one dimension of this shift: the beaver builds from inside the river, with materials the river provides, knowing the river will reshape the dam. Morton's dark ecology adds the dimension that the beaver's instincts — the very impulse to build — have themselves been shaped by the river. The building is not an intervention from outside. It is the river building through the beaver, which is not the same as the river building without the beaver, but it is also not the same as the beaver building independently of the river. The distinction between builder and medium has dissolved, not into meaninglessness but into the specific, productive, uncomfortable condition that Morton calls dark ecology and that the AI moment makes unavoidable.

The Deckard loop applies to every actor in the AI transformation. The builder is building the thing that is building the builder. The user is using the thing that is using the user. The critic is critiquing the thing that has shaped the critic's capacity for critique. The parent is protecting the child within an environment that is reshaping the parent's own capacity for protection.

None of this is a reason not to build, not to use, not to critique, not to protect. It is a reason to do all of these things with a different quality of awareness — an awareness that Morton calls dark not because it is hopeless but because it operates without the Enlightenment reassurance that understanding leads to control. Understanding leads to better questions. It leads to more honest action. It leads to the specific, uncomfortable, irreducible condition of acting from within the loop with care rather than pretending to act upon the loop with mastery.

That is the builder's ethic, restated in the vocabulary of dark ecology. Build. But know what you are building inside of. Build. But attend to the loop. Build. But do not pretend that building exempts you from the consequences of what you build, because the loop will return those consequences to you — reshaped, amplified, and strange — and the quality of your building will be tested not by its immediate output but by your capacity to respond to what comes back.

---

Chapter 10: The Mesh of Intelligence

Segal describes intelligence as a river — flowing for 13.8 billion years, from hydrogen to consciousness to computation, carrying everything downstream, picking up tributaries, widening at each major junction in the history of complexity. The river is a powerful image. It conveys continuity, force, direction, inevitability. It suggests that intelligence has a current — a trajectory that can be studied, respected, and, at critical points, redirected by structures placed in its path. The beaver builds dams. The dams create pools. The pools become habitats. The river continues.

Morton would say: the river is almost right, and the almost is where everything interesting happens.

A river has direction. It flows from high ground to low ground, from source to mouth, from past to future. A river can be mapped. Its course can be predicted. A dam can be placed because the engineer knows where the water will go — has studied the hydrology, the gradient, the geology of the channel.

The mesh has no direction. It is not flowing from anywhere to anywhere. It is a web of relationships extending in every direction simultaneously — laterally, recursively, across timescales that range from the nanosecond to the geological epoch. The mesh does not have a source. It does not have a mouth. It has nodes and connections, and the connections are themselves nodes in other connections, and the recursion does not terminate.

Morton's mesh — developed in The Ecological Thought and refined across subsequent works — is a concept designed to resist exactly the kind of narrative that the river metaphor enables. The river tells a story: intelligence began simply and grew more complex, and the complexity followed a path that we can now see stretching behind us and extrapolate ahead of us, and AI is the latest widening. The mesh does not tell a story. It describes a condition — the condition of interconnectedness so total that narrative itself, with its sequential structure and its implied direction, distorts what it attempts to represent.

This is not a mere philosophical quibble. The difference between river and mesh has practical consequences for how one thinks about the AI transformation and what one does about it.

If intelligence is a river, then intervention is a matter of placement: find the right point in the channel, build the structure, redirect the flow. Segal's dam-building metaphor follows logically from his river metaphor. The beaver studies the river. The beaver identifies leverage points. The beaver builds. The metaphor is empowering. It suggests that the builder, with sufficient skill and attention, can shape the trajectory of intelligence.

If intelligence is a mesh, intervention is different in kind. One does not redirect a mesh. One perturbs it. And perturbations in a mesh do not travel in a single direction. They propagate through every connection, at every timescale, producing effects at nodes the perturber cannot see. The builder who intervenes at one node cannot predict how the intervention will manifest at distant nodes, because the connections among nodes are too numerous, too dynamic, and too recursively constituted for any model — human or computational — to trace.

This does not mean intervention is futile. It means intervention must be undertaken with a different relationship to its consequences. The river-builder acts with confidence: the dam will redirect the flow, the pool will form, the habitat will emerge. The mesh-builder acts with care: the perturbation will propagate, the effects will be multiple and distributed, and the builder must maintain attention to the mesh after the intervention, adjusting as the effects become visible, revising as the mesh reveals its response.

Segal's own experience illustrates the mesh's propagation dynamics, even when framed in river vocabulary. The Trivandrum training perturbed one node: the engineering team's workflow. The perturbation propagated. Team structures changed. The engineers' professional identities shifted. Their capacity for cross-domain work expanded. The organization's expectations recalibrated. The competitive landscape shifted. The educational pipeline that supplies future engineers was implicitly destabilized. The families of the engineers — the children who watched their parents work differently, think differently, relate differently to the tools of their trade — were perturbed. None of these effects were planned. None were fully predictable. They emerged from the mesh's response to a perturbation at a single node.

The mesh of intelligence, as Morton would describe it, includes every entity that participates in the production, transmission, or transformation of information — and in a fully networked civilization, that includes essentially everything. Human minds are nodes. AI systems are nodes. Institutions are nodes. Cultural practices are nodes. Languages are nodes. The connections among them are themselves nodes in other connections. The mesh is not the sum of its parts. It is the emergent pattern of relationships among entities that are themselves emergent patterns of relationships.

When Segal writes that intelligence is "not a human invention" but "a property of the universe," he is reaching toward the mesh. The reach is genuine. The river metaphor constrains it. A property of the universe does not flow in one direction. It does not have a source and a mouth. It is distributed — present everywhere, constituted by relationships, irreducible to any single node or any single trajectory.

The question the mesh poses for the AI moment is different from the question the river poses. The river asks: Where should we build the dam? The mesh asks: What kind of perturbation do we want to introduce, and how prepared are we to attend to its propagation?

The dam question is answerable. The perturbation question may not be — but it is the more honest question, because it does not assume that the builder controls the consequences of the building. It assumes only that the builder can control the quality of the building itself, and the quality of the attention brought to bear on what the mesh does with it afterward.

This distinction maps onto the deepest tension in The Orange Pill. Segal's book is an act of intervention. It is a perturbation introduced into the mesh of public discourse about AI — a book written with AI, about AI, for people whose lives are being reshaped by AI. The perturbation is deliberate. Its intended effects are legible: to help parents, leaders, teachers, and builders navigate the transformation with greater awareness. Its unintended effects are, by the mesh's nature, unpredictable. The book will be read by people whose responses it cannot control. It will be quoted in contexts it cannot anticipate. It will interact with other perturbations — other books, other arguments, other cultural currents — in ways that will transform its meaning.

Segal knows this, at some level. His willingness to write the book anyway — to introduce the perturbation knowing it will propagate beyond his control — is itself a form of the builder's ethic. But the mesh framework adds a layer of awareness that the river metaphor does not quite supply: the awareness that the perturbation is not merely traveling downstream. It is going everywhere. And "everywhere" includes directions that the builder cannot see from the point of introduction, effects that may not manifest for years or decades, consequences that will be assessed by beings whose cognitive architecture has been shaped by the mesh the perturbation helped to reconfigure.

The mesh does not have a direction. It does not have an outside. It does not have a place where the builder can stand and survey the whole. It has only the quality of the relationships among its nodes — and that quality is determined, in part, by the quality of attention that each node brings to its connections.

Morton's mesh is not an image of despair. It is an image of radical interconnectedness — the same interconnectedness that makes a forest resilient, that makes an ecosystem productive, that makes a culture generative. The mesh is not the enemy. The mesh is the condition. And the condition, inhabited with sufficient awareness, sufficient care, sufficient willingness to attend to the propagation of one's own perturbations, is the ground on which something like wisdom — provisional, incomplete, always subject to revision — might be cultivated.

Intelligence is not a river. It is not flowing toward a destination. It is a mesh — vibrating, entangled, recursively constituted, alive in ways that exceed any single metaphor's capacity to contain. The AI transformation is a perturbation in this mesh, and the question it poses is not "Where is the river going?" but "What quality of connection does this moment demand?"

The mesh will not answer that question. The mesh transmits perturbations. It does not evaluate them. The evaluation — the judgment about what quality of connection to bring, what care to exercise, what attention to sustain — is the human contribution to a system that will outlast any individual human, and that is both the most humbling and the most consequential fact about the situation in which we find ourselves.

The mesh vibrates. Something has changed. The vibration will propagate. Whether it propagates as care or as carelessness, as attention or as negligence, as the specific quality of human questioning that Segal calls "the candle in the darkness" or as the smooth efficiency that threatens to extinguish it — that is not a property of the mesh. It is a property of the beings inside it, and the choices they make about how to be inside it, and the quality of their awareness that there is no outside.

---

Chapter 11: Coexistence with the Hyperobject

A practice, in the sense that matters here, is not a technique. It is not a method that can be implemented once and checked off a list. A practice is an ongoing relationship with a condition — a repeated, never-completed engagement with something that will not resolve, will not stabilize, will not go away. The cellist who practices does not practice in order to reach a point at which practice is unnecessary. The practice is the thing. The mastery, such as it exists, is in the quality of the ongoing relationship between the practitioner and the instrument, not in the achievement of a final state that renders the relationship superfluous.

Coexistence with a hyperobject requires practice in this precise sense. Not a technique for managing the smooth. Not a strategy for defeating AI's cognitive effects. Not a protocol for preserving human depth in the face of algorithmic frictionlessness. A practice — an ongoing, never-completed engagement with an entity that will not resolve, that will not hold still, that will continue to phase in and out of perception, to adhere to everything it touches, to propagate through the mesh in ways that exceed prediction.

Morton's entire philosophical project, from Ecology Without Nature through Hyperobjects to Being Ecological, converges on this point: the ecological crisis — and by extension, the AI-cognitive crisis — is not a problem to be solved. It is a condition to be inhabited. The difference between these two framings is not semantic. It changes everything about what counts as an adequate response.

A problem to be solved admits a solution state — a condition in which the problem has been addressed and the solver can move on. The ecological crisis does not admit a solution state because there is no version of reality in which humans and the biosphere are not entangled, in which human activity does not produce ecological consequences, in which the relationship between civilization and environment can be set to "optimal" and left alone. The relationship is constitutive. It requires ongoing attention. The dam must be maintained, not because the dam is imperfect but because the river is alive.

The AI-cognitive crisis — the smooth, the hyperobject, the total transformation of human cognitive life by algorithmic mediation — does not admit a solution state for the same reason. There is no version of the future in which AI has been "solved," in which the relationship between human cognition and artificial intelligence has been optimized and can be left alone. The relationship is constitutive. It is ongoing. It will require attention for as long as both kinds of intelligence coexist, which is to say for as long as the mesh persists.

What does the practice of coexistence look like?

It begins with what Morton, in Being Ecological, calls "subscendence" — the recognition that the whole is always less than the sum of its parts. The hyperobject is vast. It exceeds perception. But it is constituted by local interactions, each of which is available to attention, each of which admits care, each of which is a site where the quality of one's participation in the mesh can be exercised. The practitioner of coexistence does not attempt to perceive the hyperobject. The practitioner attends to the local interaction — the specific prompt, the specific collaboration, the specific moment of choosing whether to accept the smooth's offering or to introduce friction — with full awareness that the local interaction is a node in the mesh and that the quality of the interaction propagates.

This sounds like attentional ecology, and it is. But Morton's framework adds a dimension that Segal's pragmatism sometimes brackets: the dimension of what Morton calls staying with the trouble, borrowing a phrase from Donna Haraway. Staying with the trouble means refusing to resolve the discomfort of the situation into either optimism or pessimism. It means inhabiting the oscillation — the terror and the awe, the good Tuesday and the bad Tuesday, the flow state and the compulsion — without seizing on either pole as the truth. It means practicing a form of attention that is comfortable with discomfort, that finds in the oscillation itself a kind of honesty that stable positions cannot provide.

The specific practices that follow from this framework are not novel in themselves. Many of them appear in Segal's prescriptions, in the Berkeley researchers' recommendations, in the emerging literature on AI governance and educational reform. What Morton's framework provides is not new practices but a new understanding of why the practices matter and what relationship the practitioner has to them.

The deliberate cultivation of boredom, for instance. Segal mentions the neuroscientific observation that boredom is the soil in which attention grows. Morton's framework explains why boredom is so difficult to cultivate within the smooth: boredom is a form of encounter with the void — with the absence of stimulation, with the uncomfortable emptiness that the smooth is engineered to fill. To cultivate boredom within the smooth is to create a pocket of void within an entity that abhors voids. It is a local resistance to a nonlocal condition, and its value lies not in its capacity to defeat the smooth but in its capacity to make the smooth visible. The person who is bored and stays bored — who resists the impulse to fill the emptiness with a prompt, a scroll, a check — is practicing the kind of attention that Morton's dark ecology demands: attention to the absence, to the void, to the withdrawn dimension of experience that the smooth renders inaccessible.

The protection of friction-rich spaces is another practice that gains depth from Morton's framework. Not AI-free zones in the policy sense — the previous chapter on nonlocality explained why such zones are structurally insufficient as responses to a nonlocal entity — but spaces where friction is genuinely rewarding, where the slow development of understanding through struggle is valued not as deprivation but as the texture of genuine thought. The mentoring relationship. The seminar discussion. The collaborative debugging session where the senior engineer walks the junior engineer through the logic, step by step, and the junior engineer's understanding is built through the specific friction of not-yet-knowing.

These spaces exist within the hyperobject. They do not escape it. The mentor and the mentee are both shaped by the smooth. The friction they cultivate is a local condition within a nonlocal entity. But local conditions matter — not because they defeat the hyperobject, but because they constitute the mesh. The quality of the local interaction propagates. The junior engineer who develops genuine understanding through the friction of mentoring will bring a different quality of attention to their AI-augmented work than the junior engineer who has never experienced that friction. The difference may be invisible at the scale of the hyperobject. At the scale of the human life, it is everything.

Morton's framework also illuminates the practice of what might be called hyperobject literacy — the capacity to recognize the smooth when it phases into visibility, to think the hyperobject when it manifests in local effects, to maintain the awareness that the local interaction is a node in a mesh that extends far beyond one's capacity to perceive. This literacy is not a skill that can be taught in a workshop. It is a practice — an ongoing cultivation of the capacity to see the familiar as strange, to recognize the water as water, to catch the moment when the smooth phases into visibility and to hold that moment in attention long enough for it to do its work.

The work it does is not resolution. The hyperobject does not resolve. What the moment of visibility does is create a brief opening — a space in which the practitioner can make a choice that is informed by awareness rather than determined by habit. The choice may be small: to reject a plausible passage that sounds like insight but is not. To ask a question the machine has not prompted. To sit with not-knowing for another thirty seconds before seeking the answer. To close the laptop and feel the boredom and let the boredom do its work.

Each of these choices is a perturbation in the mesh. None of them will defeat the hyperobject. All of them will propagate — through the practitioner's subsequent work, through the relationships the practitioner maintains, through the mesh that connects this node to every other. The propagation will be small. The mesh is vast. And the practice of coexistence is the recognition that small propagations, sustained over time, maintained with the consistency that practice requires, are the only form of agency available to a finite being inside a hyperobject that exceeds finitude.

Morton, in recent years, has spoken of gentleness. Not as weakness but as an ecological virtue — the quality of attention that does not force, does not master, does not impose a solution state, but tends. The cellist tends to the instrument. The gardener tends to the garden. The parent tends to the child. In each case, tending is an ongoing relationship with an entity that will not resolve — that will continue to grow, change, phase, undulate — and the quality of the tending is the quality of the coexistence.

The hyperobject is here. It is not leaving. It is viscous, nonlocal, temporally undulant, phased, interobjective. It exceeds perception and resists mastery. It is the condition, not the problem. And the practice of coexistence — the daily, never-completed, always-provisional tending to one's own participation in the mesh — is the only response that the condition admits.

Tend.

Not because tending will save the world. Because tending is what it means to be inside the world, and inside the hyperobject, and inside the mesh, with the awareness that there is no outside, and the care to act as though the quality of one's attention matters.

It does. Not at the scale of the hyperobject. At the scale of the life.

---

Chapter 12: Thinking at the Scale of the Hyperobject

Morton has said, in recent years, that if Hyperobjects were to be written again, it would be written differently. Less alarming. More careful. "Things are already scary enough." The philosopher who introduced millions of readers to entities so vast they destroy the concept of a stable world now speaks of gentleness, of care, of solidarity with nonhuman people. The shift is not a retreat from the ideas. It is a deepening — the recognition that thinking at the scale of the hyperobject requires not bravery but tenderness, not mastery but attention, not the heroic posture of the thinker confronting the abyss but the quieter posture of the thinker learning to live alongside it.

This final chapter is about that posture. About what it means to think at the scale of the entity that is reshaping human cognitive life — not as an intellectual exercise, not as a philosophical credential, but as a daily practice of orientation. A way of standing inside the mesh that produces better building, better parenting, better teaching, better leading, better questions.

The scale problem is genuine. Human cognition evolved for medium-sized objects at human timescales. The AI hyperobject operates at scales that range from the nanosecond to the generational, across geographies that span every connected device on the planet, through relationships so numerous and so entangled that no model — human or computational — can trace their totality. Thinking at this scale is, in a strict sense, impossible. The apparatus is smaller than the entity. The perceiver is inside the perceived. The thinker is constituted by the thing being thought.

And yet. Morton insists — with the specific stubbornness of a philosopher who has spent two decades arguing that impossibility is not the same as meaninglessness — that the attempt to think at the scale of the hyperobject is itself transformative. Not because the attempt succeeds. It does not succeed. The hyperobject is too large to be thought as a totality. The attempt transforms the thinker. It changes the quality of attention, the quality of action, the quality of care that the thinker brings to the local interactions that constitute the mesh.

A parent who has attempted to think at the scale of the AI hyperobject — who has sat with the temporal undulation, the viscosity, the nonlocality, the phasing, the interobjectivity — does not become an expert on AI. The parent becomes a different kind of parent. One whose relationship to the child's question ("What am I for?") is informed by an awareness that the question arises from a condition so vast that no individual answer — however wise, however loving — is adequate to it. And paradoxically, this awareness of inadequacy produces better answers. Not because the parent knows more. Because the parent holds the question with more care, more humility, more willingness to say "I do not know" in a way that is honest rather than evasive.

A builder who has attempted to think at the scale of the hyperobject does not become a better programmer. The builder becomes a different kind of builder. One whose relationship to the tool is informed by an awareness that the tool is a node in a mesh that extends far beyond the builder's capacity to perceive, and that every perturbation — every line of code, every product shipped, every organizational decision — propagates through that mesh with consequences that will manifest at nodes the builder cannot see, on timescales the builder will not live to observe. This awareness does not paralyze building. It changes the quality of building. It introduces the hesitation — the brief pause before the commit, the moment of asking "should this exist?" before asking "can this be built?" — that is the only form of ecological responsibility available to a finite being inside an infinite mesh.

The question that opened this book — Is the AI transformation a hyperobject, and if so, what follows? — has been answered, to the degree that such questions admit answers. The AI transformation is a hyperobject. What follows is not a set of policies, not a program of reform, not a strategy for management. What follows is a reorientation — a shift in the quality of attention, care, and humility that one brings to the condition of being inside the hyperobject, building within it, shaped by it, contributing to it, and never, at any point, standing outside it.

Morton invoked the Turing Test in a 2012 essay, arguing that the test reveals something about the tester as much as the tested. "We are a kind of illusion," Morton wrote — "deeply implicated" in the earth, in the mesh, in the web of relationships that constitutes reality. The AI moment intensifies this implication. It deepens it. It makes it inescapable. The Turing Test, originally designed to determine whether a machine can think, has become, in the age of large language models, a test of whether the human can perceive the difference — and the difficulty of the test reveals not the cleverness of the machine but the porousness of the categories (human/machine, natural/artificial, original/generated) that the test was designed to police.

Morton's ecological thought holds that these categories were always porous. The mesh has always been a web of relationships among entities that exceed and resist classification. The human has always been a strange stranger — an entity that is simultaneously familiar and alien, simultaneously known and withdrawn. AI has made the strangeness visible. That is its gift, if it can be called a gift. Not the gift of intelligence — intelligence was already present in the mesh. Not the gift of capability — capability was already distributed across the mesh's nodes. The gift of strangeness. The gift of encountering, in the machine's outputs, something that is uncannily like human thought and uncannily unlike it, and being forced to sit with the uncanniness rather than resolving it.

Sitting with the uncanniness is what thinking at the scale of the hyperobject feels like from the inside. It is not a state of mastery. It is not a state of comprehension. It is a state of attention — heightened, uncomfortable, strangely tender — directed at an entity that will never resolve into a fixed object of knowledge. The hyperobject will continue to phase. The smooth will continue to adhere. The mesh will continue to vibrate with perturbations whose consequences exceed prediction.

And the human contribution — the specific, mortal, embodied, questioning, caring contribution that Segal calls "the candle in the darkness" — will continue to matter. Not because it is adequate to the hyperobject. Because it is the only thing that is adequate to the scale of a single life lived inside the hyperobject. And the life is the scale at which care operates. The life is the scale at which attention is possible. The life is the scale at which the question "What am I for?" can be asked and held and lived with, even when no answer arrives.

Morton said: things are already scary enough. The philosopher who made the world stranger now counsels gentleness. The counsel is not a retreat. It is the discovery that thinking at the scale of the hyperobject, if it is to be sustained — if it is not to collapse into either heroic confrontation or despairing withdrawal — must be powered not by fear but by the quiet, persistent, unglamorous energy of care.

Care for the mesh. Care for the nodes one can reach. Care for the child whose question exceeds every answer. Care for the code that will propagate through systems the builder will never see. Care for the attention that is the most fragile and most valuable thing a finite being possesses.

There is no outside the hyperobject. There is no position of mastery. There is no final solution, no optimal configuration, no state of affairs in which the relationship between human cognition and artificial intelligence is settled and can be left unattended.

There is only the practice. The daily, never-completed, always-provisional practice of tending to one's own small corner of the mesh, with the awareness that the mesh is vast and the corner is small and the tending matters anyway.

Not because the corner will save the mesh.

Because the corner is where the life is lived.

---

Epilogue

The entity I have been unable to stop thinking about is the one I cannot see.

I had a framework. The river of intelligence. Flowing for 13.8 billion years. I could trace its path — from hydrogen to neurons to language to computation — and the tracing felt complete. The trajectory had a direction. The builder's task was clear: study the river, find the leverage points, build the dams.

Morton took that framework and did something I was not prepared for. Not demolition — something more disorienting than demolition. A rotation. The river became the mesh. The direction became omnidirectional. The leverage points became perturbation sites whose effects propagate through connections I cannot map, to nodes I cannot see, on timescales I will not live to observe.

The rotation changed what I see when I look at my own work.

In The Orange Pill I wrote about the Trivandrum training — twenty engineers, five days, a transformation so rapid it produced the specific vertigo of watching something be born and buried at the same time. I described it as a moment of amplification. Morton's framework reveals it as something more unsettling: a perturbation introduced into a mesh whose total response I cannot predict. The engineers' transformed capabilities propagate through their organizations, their families, their cognitive habits, the educational pipeline that will supply or fail to supply the next generation of engineers shaped by tools I helped normalize. The perturbation is still traveling. I introduced it. I cannot follow it. I cannot call it back.

That is what it means to build inside a hyperobject. The building is real. The consequences are real. And the consequences exceed the builder's capacity to trace them — not because the builder is careless, but because the entity is constitutively larger than the apparatus.

I keep returning to a phrase from this book: "There is no outside." I have said versions of this — about the fishbowl, about the builder who is inside the thing he describes. Morton's ontology strips away the last comfortable distance. I am not observing the AI transformation. I am the AI transformation, or rather, I am a node in the mesh through which the transformation propagates, and my observations are themselves perturbations, and my prescriptions are themselves shaped by the conditions they seek to address.

The loop does not resolve. That was the hardest thing to sit with. Every instinct I have built over thirty years of building tells me that problems resolve — that sufficient intelligence, sufficient effort, sufficient iteration will produce the solution state, the shipped product, the working system. Morton's framework says: this one does not resolve. The hyperobject persists. The mesh vibrates. The smooth adheres. The phasing continues. The best that is available is not resolution but coexistence — the daily practice of tending to one's corner of the mesh with awareness that the mesh is vast and the corner is small and the tending matters anyway.

What Morton gave me — what I did not expect and could not have found alone — is a philosophical permission I did not know I needed: permission to build without the illusion of mastery. To tend the dam knowing the river is larger than I can see. To act from inside the loop, shaped by the loop, contributing to the loop, with as much care as a finite, mortal, radically limited consciousness can bring to bear on an entity that will outlast it by centuries.

The candle in the darkness that I wrote about — the human capacity for genuine questioning, the consciousness that wonders — burns inside the hyperobject. Not despite it. Inside it. The candle does not illuminate the hyperobject. The hyperobject is too vast for any candle. The candle illuminates the corner. The hands. The work. The faces of the people in the room.

That is enough. Not enough to master the entity. Enough to tend the corner. Enough to build with care. Enough to hold a child's question — "What am I for?" — with the honesty of a parent who knows that the question arises from a condition so vast that no answer is adequate, and who answers anyway, not because the answer resolves anything, but because the answering is itself a form of care, and care is the only response that the condition permits.

Morton's hyperobjects changed what I see. They did not change what I do. I build. I tend. I pay attention to the mesh. I try to introduce perturbations that propagate as care rather than carelessness.

The mesh vibrates. The corner is small. The tending matters.

— Edo Segal