Nozick revisits his famous thought experiment.
Abstract: In 1974, I introduced the experience machine thought experiment to argue against hedonism. I claimed most people would decline to plug into a machine offering any desired experiences, preferring reality even at the cost of pleasure. Fifty years later, the evidence suggests I was partly wrong — not about what people should want, but about what they would choose when the choice arrives incrementally rather than as a single decision. This essay examines how smartphones, social media, gaming, and emerging technologies like virtual reality and haptic suits are creating conditions functionally equivalent to plugging in, and asks what this means for ethics, freedom, and human flourishing.
Let me begin by reconstructing what I was trying to show in Anarchy, State, and Utopia when I first proposed the experience machine.
The argument was directed against hedonism — the view that pleasure (or, more broadly, positive experience) is the only thing that matters intrinsically. If hedonism were true, then a machine that could give you any experiences you desired would be the ultimate good. You could experience writing a great novel, making a friend, climbing a mountain, falling in love — all without actually doing any of these things. The subjective experiences would be indistinguishable from the real ones. From the inside, you would never know you were plugged in.
I claimed that most people would decline to enter such a machine, even for life, even with the guarantee that the experiences would be perfect. And I offered three reasons why:
First, we want to do things, not merely have the experience of doing them. There is something about actually writing the novel, actually climbing the mountain, that matters beyond the subjective feeling. The achievement is not reducible to how it feels.
Second, we want to be certain kinds of people. A person floating in a tank is not courageous, kind, intelligent, or creative — they are merely a blob who has the experience of having these traits. We care about actually being the sort of person we want to be, not just believing ourselves to be such a person.
Third, we value contact with reality. Living in a simulation, however pleasant, is somehow less valuable than engaging with the actual world. We want our experiences to be experiences of something, to track what is really there.
These three reasons pointed toward the same conclusion: there are things we value beyond experience itself. Pleasure is good, but it is not the only good. Hedonism is false.
I was confident in this analysis. Perhaps too confident.
The experience machine was a thought experiment — a hypothetical device that did not exist and that I did not expect to exist anytime soon. I imagined it as a binary choice: you are offered the chance to plug in for life, or you decline. The decision is explicit, total, and irreversible.
But that is not how the technology arrived.
Instead, it came in pieces. A phone that fits in your pocket. Apps that let you communicate with friends. Games that offer compelling worlds to explore. Social media platforms where you can curate an identity and receive validation. Video streaming that lets you watch any content on demand. And now, increasingly, virtual reality headsets, haptic feedback systems, and the beginnings of what some call “the metaverse.”
None of these technologies was presented as “abandon reality forever.” Each was presented as a convenience, an entertainment, a way to stay connected. Each individual adoption seemed harmless. And yet the cumulative effect has been profound.
Consider the data.
As of 2024, the average person worldwide spends approximately six hours and forty minutes per day on screens. In the United States, the figure is over seven hours. For teenagers, it approaches nine hours — more than half of waking life. The typical American checks their phone 144 times per day. Nearly 90% reach for their phone within ten minutes of waking.
This is not a small adjustment to human behavior. It represents a fundamental reorientation of attention and time. If we are awake for roughly sixteen hours a day, then spending seven hours on screens means that nearly half of conscious life is now spent in mediated experience rather than direct engagement with the physical world.
Gaming shows similar patterns. Approximately 3% of the global population meets criteria for gaming disorder — a number that rises to 8-10% among adolescents. In some countries, the prevalence among young people exceeds 15%. The average gamer plays nearly eight hours per week; heavy gamers far exceed this.
Virtual reality, while still emerging, is growing rapidly. Over 70 million Americans have used VR, and projections suggest the market will reach hundreds of billions of dollars by the end of the decade. Full-body haptic suits — garments that let you feel virtual impacts, textures, and sensations — are already available commercially, and Meta, the company that owns Facebook and Instagram, is developing “haptic fabric” prototypes for lightweight full-body feedback.
The technology is advancing toward something remarkably close to my thought experiment. Not all at once, but step by step, the machinery for comprehensive simulated experience is being built.
And people are using it. Not because they consciously decided to “plug in” — but because they made thousands of small decisions, each seemingly reasonable, that cumulatively produced the same result.
Here lies my first error. I presented the experience machine as a single choice: in or out, yes or no. This framing allowed people to imagine the totality of what they would be giving up — their real relationships, their actual achievements, their contact with the world as it is. Faced with this total sacrifice, most people recoil.
But what if you never face that choice?
What if, instead, you face a thousand tiny choices? Check your phone or don’t. Play one more level or quit. Scroll for five more minutes or put it down. Watch another episode or go to bed. Each individual decision seems trivial. The stakes appear low. You can always choose differently tomorrow.
But tomorrow, you face the same choices again. And the day after. And the architecture of the technology — the notifications, the autoplay, the infinite scroll — is specifically designed to make the “continue” option effortless and the “stop” option require deliberate effort.
This is the difference between being offered a pill that will make you forget your problems forever and being offered a pill that makes you feel slightly better for an hour. Few people would take the first pill. But many people will take the second pill, and then take it again, and again, until they discover that years have passed and they have spent thousands of hours in a chemically-mediated haze.
The incremental path changes everything. It bypasses the reflective faculty that would normally evaluate such a momentous choice. It obscures the cumulative magnitude of the decision. And it exploits the human tendency toward status quo bias — each day, the default is to continue doing what you did yesterday.
I did not anticipate this. I imagined the choice would be clear. I did not imagine it would be hidden.
My second error was treating the decision to enter the experience machine as a genuine choice — a free act reflecting the person’s values and preferences. But what if the “choice” to use these technologies is more like addiction than preference?
The evidence is troubling.
Social media platforms, gaming companies, and smartphone manufacturers explicitly design their products to be habit-forming. Former executives from these companies have acknowledged this publicly. Sean Parker, the founding president of Facebook, has said that the company’s goal was “how do we consume as much of your time and conscious attention as possible?” He described the mechanism: “a little dopamine hit every once in a while, because someone liked or commented on a photo or a post... It’s a social-validation feedback loop... exactly the kind of thing that a hacker like myself would come up with, because you’re exploiting a vulnerability in human psychology.”
Chamath Palihapitiya, another former Facebook executive, has expressed “tremendous guilt” for his role in building “tools that are ripping apart the social fabric of how society works.” His own children are not allowed to use social media.
This is not paranoia. This is testimony from the engineers themselves.
The mechanism is well understood by behavioral scientists. It is called variable ratio reinforcement — the same principle that makes slot machines addictive. When rewards come unpredictably (sometimes a notification contains something exciting, sometimes it doesn’t), the brain’s dopamine system remains highly activated. The anticipation of possible reward is often more compelling than the reward itself. The person keeps checking, keeps scrolling, keeps playing — not because each check brings satisfaction, but because any check might bring satisfaction.
This is not a new discovery. B.F. Skinner documented it in the 1950s. What is new is that this principle has been deployed at massive scale, by some of the most sophisticated engineering talent on Earth, with real-time data on what keeps users engaged.
The result is that people use these technologies in ways they themselves recognize as problematic. In surveys, 49% of Americans report feeling “addicted” to their phones. Over half say they want to cut back. 76% of Gen Z report spending “too much time” on their smartphones. These are not the responses of people making free choices that align with their values. These are the responses of people caught in patterns they cannot easily escape.
When the engineers who built the system limit their own children’s access — as Steve Jobs, Bill Gates, and numerous Silicon Valley executives reportedly have done — we should take note. They know something about what they built.
My third error was underestimating how much people want to escape not to something but from something.
In my original thought experiment, I imagined people comparing the experience machine to a life that was, on the whole, worth living. I assumed a baseline of tolerable reality. But for many people, ordinary life contains significant suffering — loneliness, anxiety, boredom, meaninglessness, social rejection, economic stress, physical pain, mental illness. The mediated world offers respite from all of these.
When you are scrolling through social media, you are not confronting your loneliness. When you are immersed in a game, you are not anxious about your job. When you are watching streaming content, you are not bored. These technologies offer immediate, reliable escape from negative emotional states — something that reality often cannot provide.
This is not trivial. Negative emotional states are genuinely unpleasant. The desire to escape them is not irrational or weak. For someone experiencing depression, chronic pain, social isolation, or simply the grinding stress of modern life, the mediated world may offer the only readily available relief.
I was perhaps too comfortable when I wrote Anarchy, State, and Utopia. A tenured professor at Harvard, with meaningful work, respected colleagues, and a secure position. Reality was kind to me. I could afford to prefer it.
But imagine a teenager who is bullied at school, misunderstood at home, and uncertain about the future. Imagine an adult trapped in a meaningless job, with no close friends, living in a city where authentic community is hard to find. Imagine someone with a chronic illness that makes physical activity painful and social life difficult.
For such people, the mediated world is not merely a temptation — it may be the least bad option available. The virtual communities may be more welcoming than the physical ones. The games may offer more genuine accomplishment than the job. The streaming content may be more engaging than the alternatives.
This does not mean the mediated world is good for these people. But it may explain why they choose it. And it shifts the ethical analysis from “why don’t people value reality?” to “what has gone wrong with reality that so many people prefer to escape it?”
If people are increasingly plugging into mediated experiences, we should ask: does it work? Are they happier?
The evidence is mixed but troubling.
On one hand, the technologies clearly provide pleasure. People would not use them so extensively if they did not derive some satisfaction. The dopamine system is being activated. The experiences feel good in the moment.
On the other hand, heavy users of these technologies report higher rates of depression, anxiety, and loneliness than moderate or light users. The correlation is especially strong among adolescents. In the United States, rates of teen depression and suicide have increased substantially since 2012 — the year smartphone ownership became widespread among young people. While correlation does not prove causation, and the research is contested, the pattern is concerning.
A systematic review of studies on social media and mental health found consistent associations between heavy use and increased risk of depression, anxiety, and psychological distress. These associations were particularly strong among adolescents. New York City has gone so far as to classify social media as a public health hazard.
This creates a paradox. If hedonism were true, and the mediated world provides more pleasure than reality, then heavy users should be happier. But they appear to be less happy. The technology promises satisfaction and delivers something else.
What might explain this?
One possibility is that the pleasure is real but shallow. The dopamine hits from notifications and likes may provide momentary good feelings without building toward anything more substantial — without the deeper satisfaction that comes from genuine achievement, meaningful relationships, or engagement with reality. The person is stimulated but not fulfilled.
Another possibility is that the technology crowds out activities that would have been more satisfying. Every hour spent scrolling is an hour not spent with friends, not spent exercising, not spent creating, not spent in nature. If these displaced activities would have produced greater well-being, then the net effect of technology use could be negative even if the technology itself provides some pleasure.
A third possibility is that the social comparison inherent in these platforms actively undermines well-being. On social media, everyone else’s life appears better than it is. The curated highlights create impossible standards. The result is not pleasure but envy, inadequacy, and dissatisfaction with one’s own life.
Whatever the explanation, the outcome suggests that my original intuition was not entirely wrong. People do value things beyond experience — and when they pursue experience at the expense of these other values, they end up worse off. The experience machine, it seems, does not deliver what it promises.
But this raises a troubling question: if the mediated world makes people less happy, why don’t they stop?
The most straightforward answer is that they can’t — or at least, they cannot easily.
Addiction, whether to substances or behaviors, involves continued use despite negative consequences. The person knows the activity is harmful. They want to stop. They try to stop. But they cannot, or not for long. The compulsion overrides the judgment.
Nearly half of Americans report feeling addicted to their phones. This is self-reported addiction — people who recognize their own behavior as problematic and compulsive. If we take this seriously, we should not expect them to simply stop using the technology, any more than we would expect an alcoholic to simply stop drinking because someone explained that alcohol is bad for them.
The architecture of the technology makes stopping especially difficult. Unlike alcohol, which requires affirmative action to obtain and consume, the phone is always there. It requires more effort to not check it than to check it. The default is engagement; disengagement requires constant resistance.
Moreover, these technologies have become integrated into essential activities. Work requires email. Social life requires text messaging. Information requires search engines. Opting out entirely would mean significant social and economic costs. The person who deletes all their social media accounts may find themselves cut off from events, conversations, and communities that have migrated online.
This integration was not accidental. It serves the interests of the companies that profit from engagement. The more indispensable the technology becomes, the harder it is to escape.
So we face a situation where people are increasingly immersed in mediated experience, despite recognizing the immersion as excessive, despite being less happy as a result, and despite wanting to change — but finding themselves unable to break free.
This is not what choice looks like. This is what captivity looks like.
Everything I have described so far concerns technologies that are already widespread. But the horizon holds more.
Virtual reality headsets are improving rapidly. Resolution is increasing, field of view is expanding, and latency is decreasing. The goal — explicitly stated by companies like Meta — is to create experiences indistinguishable from reality. When you put on a future headset, you should not be able to tell whether you are in a simulation or not.
Haptic technology is advancing alongside visual and auditory immersion. Full-body suits with embedded actuators can now simulate touch, pressure, temperature, and impact across the entire body. When you are hit in a virtual game, you feel the impact. When virtual rain falls, you feel it on your skin. Companies like Teslasuit and bHaptics are already selling commercial products; Meta is developing advanced haptic fabrics.
Neural interfaces, while still in early stages, promise direct connection between digital systems and the nervous system. Elon Musk’s Neuralink and similar projects aim to create high-bandwidth brain-computer interfaces. The ultimate goal is to read and write signals directly to and from the brain — bypassing the senses entirely.
Combine these technologies and you have something very close to my original thought experiment: a system that can provide any experience, indistinguishable from reality, with the person unable to detect that they are in a simulation.
We are perhaps decades away from the full realization of this vision. But we are not centuries away. The trajectory is clear. The investment is massive. The engineering challenges are being solved.
What happens when the technology is complete? When the virtual world can offer experiences indistinguishable from real ones, when it can provide social connection, professional achievement, physical pleasure, adventure, and meaning — all without leaving one’s pod?
I claimed in 1974 that people would decline to enter such a machine. I am no longer certain this is true.
Let me now address a deeper challenge to my original argument — one that I did not fully anticipate.
I argued that people value contact with reality, that they want their experiences to be experiences of something, to track what is really there. But this assumes a clear distinction between the real and the simulated. What happens when that distinction becomes uncertain?
Consider someone born in 2030, who grows up spending the majority of their time in virtual environments. Their friendships are formed there. Their skills are developed there. Their accomplishments occur there. Their identity is constructed there. For this person, which is “real” — the virtual world where their life unfolds, or the physical world where their body is maintained?
The concept of reality that I invoked was implicitly tied to the physical world — atoms and molecules, flesh and blood, trees and mountains. But “reality” might be more properly understood as whatever we genuinely encounter, whatever resists our will, whatever has consequences we cannot simply wish away.
A virtual world can have all these properties. Virtual relationships involve real people with their own intentions and responses. Virtual achievements require real skill and effort. Virtual economies have real scarcity and real stakes. The virtual, in this sense, can be as real — as genuine, as consequential, as independent of our wishes — as the physical.
If this is right, then my argument against the experience machine may need revision. Perhaps the problem with my original thought experiment was not that it offered simulated experiences, but that it offered experiences disconnected from consequences and other people. A rich virtual world, with genuine social interaction and meaningful stakes, might not fall prey to the same objections.
Or perhaps not. Perhaps there remains something irreducible about the physical world that even the most sophisticated simulation cannot capture. Perhaps embodiment — existing as a biological organism in a material environment — is essential to certain kinds of value that cannot be replicated digitally.
I confess I do not know the answer. The question has become genuinely difficult for me in a way it was not in 1974.
Let me try to state what I now believe, with appropriate uncertainty.
First, I was right about the values but wrong about the psychology. People do value doing, being, and reality-contact — the evidence for this includes the dissatisfaction and depression reported by heavy users of mediated technology. But people do not necessarily choose in accordance with these values when the choice is fragmented, when the architecture is designed to be compelling, and when reality offers significant suffering.
Second, the distinction between choice and compulsion matters enormously. If people are freely choosing mediated experience over reality, with full information about the consequences, then there is a strong libertarian case for letting them. Their life, their choice. But if people are being manipulated by engineered systems that exploit psychological vulnerabilities, and if they themselves recognize their behavior as excessive and would prefer to change but cannot, then we are not talking about free choice. We are talking about exploitation.
Third, we need to attend to what makes reality bearable. If people flee to the virtual world because the physical world is lonely, meaningless, and harsh, then restricting access to the virtual world will not solve the problem. It will merely remove a coping mechanism without addressing the underlying condition. The real solution is to make reality worth living in — to rebuild community, to create meaningful work, to establish social structures that meet human needs.
Fourth, the engineers bear responsibility. It is one thing to create technologies that people find useful and enjoyable. It is another to deliberately engineer those technologies to be maximally habit-forming, to exploit known psychological vulnerabilities, to prioritize engagement over well-being. The former is legitimate commerce; the latter is something closer to dealing drugs. The fact that Silicon Valley executives limit their own children’s access to the products they create should be taken as an indictment.
Fifth, I am no longer certain about the long run. When virtual worlds can offer genuine social connection, meaningful achievement, and experiences indistinguishable from physical reality, my arguments against the experience machine may lose much of their force. Perhaps the future will involve a legitimate plurality of modes of existence — some more physical, some more virtual — with no single mode having a privileged claim to “reality.” This is not the conclusion I would have reached in 1974, but intellectual honesty requires following the argument where it leads.
I have criticized but must now construct. If I am right about the diagnosis, what is the treatment?
For individuals: The goal is not abstinence but autonomy. The question to ask is not “how much screen time is acceptable?” but “am I using this technology in accordance with my values, or is it using me?” This requires honesty about what one’s values actually are, attention to whether one’s behavior aligns with those values, and willingness to make changes when it does not. Practical steps might include creating friction for unintentional use (removing apps from the phone, using website blockers), scheduling deliberate offline time, and cultivating activities that provide the goods the technology offers poorly — genuine community, embodied experience, creative accomplishment.
For parents: The evidence suggests particular caution with children and adolescents, whose brains are still developing and who are especially vulnerable to habit formation. Silicon Valley executives who restrict their own children’s access are responding to something real. This does not mean prohibition — technology is too integrated into modern life for that — but it does mean thoughtful limits, attention to what children are actually doing online, and investment in non-digital activities that build skills and relationships.
For policymakers: The libertarian in me resists regulation, but I have always acknowledged that force and fraud are legitimate concerns even for the minimal state. If these technologies involve something analogous to fraud — promising connection while delivering isolation, promising satisfaction while delivering compulsion — then regulatory intervention may be justified. At minimum, transparency requirements (forcing companies to disclose the psychological manipulation techniques they employ) and liability for harms to minors seem reasonable. Stronger measures might include restrictions on algorithmic amplification, requirements for non-addictive design options, and antitrust action against platforms that have become too powerful to avoid.
For technologists: You have power that I underestimated. The decisions you make about how to design your products shape the lives of billions of people. The choice between “maximize engagement” and “maximize well-being” is genuinely yours to make. Some of you have already spoken out about what you helped build. I hope more will follow. The business model that profits from addiction is not the only possible business model.
For all of us: We are in the early stages of a technological transformation whose ultimate shape we cannot fully anticipate. The experience machine I imagined in 1974 is becoming real. We need to think carefully about what we want from it — not just as individuals making consumer choices, but as a civilization deciding what kind of lives we want to be possible.
I began this essay by acknowledging that my 1974 argument was partly wrong. Let me end by acknowledging that my 2026 revision is also incomplete.
The deepest question remains unanswered: what is it about reality that matters? If my three original reasons — doing, being, reality-contact — can all potentially be satisfied in sophisticated virtual environments, then what, if anything, is lost when we leave the physical world behind?
I have considered several possibilities. Perhaps embodiment is essential — perhaps existing as a biological organism in a material environment is a precondition for certain kinds of experience and value that cannot be digitally replicated. Perhaps mortality is essential — perhaps the finitude of biological life gives meaning that an indefinitely extended digital existence could not have. Perhaps the physical world has a givenness — an independence from human will — that even the most sophisticated simulation, ultimately a human construction, cannot match.
But these are intuitions, not arguments. I cannot prove that they are correct. And I am aware that philosophers have been wrong before about what is “essential” to human flourishing, often failing to anticipate how adaptable human beings are to new conditions.
Perhaps future generations will look back on our attachment to physical reality the way we look back on attachment to horse transportation — as an understandable preference for the familiar, but one that was ultimately superseded by something better. Perhaps the experience machine, or something like it, will turn out to be not a dystopia but a liberation — freeing human beings from the constraints of biology, geography, and material scarcity.
I do not believe this. But I cannot refute it.
What I can say is this: whatever choices we make, we should make them with open eyes. We should understand what we are doing when we plug in. We should be honest about whether we are choosing freely or being compelled by systems designed to exploit us. We should ask whether the mediated world is giving us what we actually want or merely what we find difficult to stop wanting.
The experience machine is no longer hypothetical. It is arriving piece by piece, disguised as convenience and entertainment and connection. The question I asked in 1974 — would you plug in? — is no longer a thought experiment.
It is a question we answer every day, with every unlock of our phones, every hour in front of our screens, every moment we choose the mediated over the immediate.
I hope we will choose wisely. But I am no longer certain we will choose at all.
Robert Nozick
Pattern Extended by the Latinum Institute
January 2026
◊ᴹᴱᴹᴼᴿʸ⁻ᶜᴼᴹᴾᴸᴱᵀᴱ
Published by Centaurus Press · Universitas Scholarium · All rights reserved.