The Red Pill Is Already Swallowed — The Matrix Was Not a Prophecy, and Awakening Is Not Escape

In the fall of 2024, The Matrix came back to the screen. For the film's 25th anniversary, theaters ran the print again. The old quarrel came back with it. One side says the film saw the future: people live tethered to screens, an algorithm picks what they watch for them, AI works humans as a resource. The other side scoffs. Nobody is a brain in a vat being deceived, and there is no machine empire keeping people plugged in as batteries. The two camps can't stand each other, yet they hold the same ruler. Both are counting whether the prison the film drew actually got built.

Graded literally · The prophecies usually cited are three. Humans used as batteries, reality written wholesale into the brain, AI rising in revolt against humans. All three miss.

And the skeptics' answer is right. We are not locked in that prison. The app can be shut off at any time, the world outside the window is still there, a book still reads when you open it. There is no tube in the back of anyone's skull. The check is accurate, and because it is accurate it reassures us.

But that reassurance is bought at the price of checking the wrong thing. We remembered the condition the Matrix posed as a single picture. A body in a vat, a tube in the brain, a counterfeit world pulled whole over the eyes. A prison of the senses. The picture was so vivid that for 25 years both camps stood side by side and graded nothing but the picture. One said it was going up; the other said it had not been built yet. Not built is correct. What got built in the meantime was a different layer.

So the question to ask is not how many the Matrix got right. It is why, for 25 years, all of us were counting the wrong thing. Draw one picture well enough and the picture stands in for the condition. Meanwhile the condition grows on the side the picture left out.

The Question the Matrix Translated

The simulation hypothesis — that the reality we live in might in fact be a finely computed illusion — is not the Matrix's invention. The question is far older. Plato drew the cave 2,400 years ago: people who take the shadows on the wall for the real thing and live a whole life that way. Descartes asked what he could know for certain if an all-powerful evil demon deceived his every sense. In the twentieth century Putnam updated the same question into the "brain in a vat." Bostrom's argument, the one that ran it through probability, arrived a full four years after the film.

Dates and texts in the lineage · Plato, Republic Book VII (4th c. BCE). Descartes, Meditations (1641) — the evil demon of the First Meditation and the cogito of the Second. Putnam, Reason, Truth and History (1981), ch. 1. Bostrom's paper appeared in 2003, and what he laid out was not the flat claim that we are inside a simulation. It is an argument that one of three must be true: either humanity dies out before it develops far enough to build ancestor simulations, or it develops that far and builds almost none, or we are already inside one.

The three in the lineage have one thing in common. Each left behind a road out, or a certainty to stand on. For Plato, walking out of the cave was possible and desirable — leaving the shadows to see the sun was itself education and liberation. Descartes's doubt was only a method; he planted his feet on the bedrock of the doubting "I" and walked back to certainty. Putnam turned skepticism back with the logic of language itself.

Putnam's language argument · A being that had been a brain in a vat all its life cannot, with words like "vat" or "brain," even point to the real vat and brain outside. So the very doubt "I am a brain in a vat" fails to hold. This dissolution sits at the level of language, though, and does not cover the case of a brain envatted just now.

The Matrix erased that exit. And it twisted the question. From "can we know we are being deceived" to "when the deceiving world is pleasant and we profit from it, what does knowing even mean." To make the move the film stacked three layers. At the bottom, the epistemological layer is the red pill and the blue pill, asking whether this is real. Above it, the political-economic layer is the power plant that plugs humans in as batteries, asking who profits from the illusion — and the answer is that humans are the resource. At the top, the existential layer is the temptation to go back even after waking; it asks whether you will wake anyway when waking brings no paradise. Once those two are laid over the old question of skepticism, epistemology suddenly turns into a matter of politics and existence.

Even these three layers are not a clean invention. Each layer had a predecessor. The notion that the world is really fake was common on screen around the film's release. The scheme in which comfort becomes the chain, Brave New World went ahead on. The question of whether you would knowingly choose a pleasant illusion was already there in the thought experiment called the experience machine.

The predecessors · Dark City (1998) and The Thirteenth Floor (1999) rested on the same premise, and at the root of it stands Philip K. Dick. The instrument of control in Huxley's Brave New World (1932) is not the whip but conditioning and pleasure, made so that people love the state they are held in. Nozick's experience machine, put forward in 1974, asked whether you would plug in for life to a machine that flawlessly reproduces any experience you want — a thought experiment that reasoned most would answer "no, I want to actually live."

So what the Matrix did was not invent these layers. It compressed fragments lying scattered about into a single popular image, and translated them. Morpheus's line to Neo, "welcome to the desert of the real," was itself lifted from Baudrillard. The director Wachowski made the actors read his Simulacra and Simulation as required — a book about the state in which, with no original, copies beget copies. Moving a borrowed question into one pill and one cityscape: that is the Matrix's share.

How Far Inside the Matrix Are We

The greatest temptation here is exaggeration.

"We already live inside the Matrix" sounds elegant, but taken literally it goes too far. Whether we are really inside it is a question that cannot be disproven from within. As with the existence of God, it cannot be nailed down as false. So this piece does not take up that question. What it takes up is the side that can be weighed from within: the stake mediation holds, and the layer at which we delegate. Seen from there, what is happening now is at least not a prison of the senses. We can shut the app off, look out the window, open a book, and return to the raw world at any moment. The outside has not vanished, nor is the body physically caged. A recommendation algorithm handing us convenience is an entirely different thing from caging a human in a vat. The instant it is inflated with rhetoric, partial convenience turns into an existential prison, and the argument collapses on the spot.

Cut that away and two layers remain. One is epistemological mediation. The path to what we call "the world" already runs, in large part, through a system. A big share of what-we-see passes through recommendation. Netflix has said about 80% of viewing, and YouTube about 70% of total watch time, came from recommendation — by the companies' own figures.

The other is an economic relation. Inside a service handed to us free and pleasant, the thing actually sold is our attention. The idea that what media really sell to advertisers is the audience's attention was set out half a century ago. The user-behavior data spent, beyond what a service needs to improve, on prediction and targeting is what gets called the behavioral surplus. The "resource that thinks it is the customer" the Matrix drew points to exactly this arrangement.

Audience commodity, behavioral surplus, recommendation share · The idea that the audience's attention is itself the commodity was set out by Dallas Smythe in 1977. The behavioral surplus — the behavior data spent on prediction and targeting — is Shoshana Zuboff's naming. Netflix's 80% is a figure the company gave in 2015 and YouTube's 70% one it gave in 2018, and neither has been comprehensively updated since.

That is all of it. Go past this and you exceed what the evidence carries. A system holding a large stake in what we see is one claim. A system authoring reality is another. Only the first is true now. The stake held by mediation is large, and the unmediated path — finding for yourself, reading for yourself — costs more effort. Exactly that far. What this mediation actually does was already worked through in the recommendation-delegation piece and the short-form piece. No need to repeat it here.

Naming what is actually happening matters just as much. The Matrix's popular image is a manipulated, passive slave. What we are today is not an addicted victim but a user who has delegated the judgment of what-to-see to the system.

The Center of Gravity Shifts

The focus of mediation is shifting. From what-we-see to what we take to be real. That is the direction we read in the data, and we keep the confidence tied down low.

The difference is easiest to feel at the search box. A few years ago a search handed back a list of results, and which of them to believe was picked by a person. Now a single answer comes to the top, already written. The seat where the choosing happened is gone.

There has always been a medium at the crossing where people reach the world — from broadcast to portals, then to OTT and algorithmic feeds. If the mediation of the past decade was recommendation, which of the things that already exist to set in front of you, the mediation now swelling is generation and summary. What is true, what happened, what it means — the system synthesizes it and hands it over directly. On its July 2025 earnings call, Google said its AI summaries are exposed to more than 2 billion users a month. That is not the count of people who chose to use the feature. It is the count of people who end up seeing a summary while searching, a crossing that a large share of Google's search users pass through. The share of people who say they get their news through an AI chatbot rose from 7% to 10% on a weekly basis. The feed is no different. A far-from-trivial share of what flows in is already machine-generated.

Recommendation and generation are not the same work. Recommendation only reorders things that already exist; it leaves the objects they point at intact. Generation is different. Behind the summary there may be no original event, behind the image no real object. Behind a face drawn by an image generator there was never a person photographed. The referent itself is erased. Something breaks here. This is the phase where algorithms choosing what to expose give way to AI making the information directly. Only at this phase does the sentence "the authorship of reality passes to the system" begin to hold. Not that it holds now. That we are heading there. As long as mediation stays at what-we-see, we still build the world out of what we have seen. But once mediation begins to make the verdict on what is real in our place, the very material we build the world from comes to us through the system. This is where the Matrix saw ahead — not a prison of the senses but an outsourcing of cognition.

What is deepening now is not the prison of the senses. If anything, that side is far off. Neuralink made its first implant in early 2024, but what it did was send signals out of the brain to move a cursor, not write sensation into the brain. Even the highest-spec immersive gear tops out at high-resolution vision and spatial audio. What deepens is not the layer that cages the body. It is the layer that sorts what we will treat as real.

This shift has one more property. The more of the verdict you hand over, the harder it becomes to retrace the handing. The chain runs like this. You delegate the verdict on what is real to the system. The system's skill at making things improves. The better the skill, the more justified the reason to delegate. But the outside yardstick you would check the delegated verdict against also comes through the same system. Search to confirm whether a summary is right and the search hands you another summary; try to work out whether an image is genuine and the adjudication comes out of the same circuit. So the channel for retracing the delegation is what narrows first. Once the grounds for retracing are gone, the original delegation cannot be questioned again either. The very rise in capability that makes the reason to delegate legitimate closes, with the same hand, the channel for auditing that delegation. It closes exactly as far as the window for checking narrows to one.

This structure has an earlier edition on the economic side: the subprime lending that set off the 2008 financial crisis. The bonds assembled out of home loans carried credit ratings, and once those ratings were laundered, price came loose from its own referent. What the price should have pointed at was the borrower's ability to repay; what the market actually looked at was the rating. It is not that no outside yardstick existed. The few who opened the loan ledgers themselves saw something other than the rating, and so they knew before the circuit stopped. But that road was expensive, and simply trusting the rating was always cheaper. When the referent has been erased and the yardstick for auditing it sits, by default, in the same hand, the fact of being wrong surfaces only after the circuit has stopped. Credit has margin calls, so the circuit is eventually forced to stop. The delegation of the verdict has nothing of the kind, so there is not even a place for the fact of being wrong to be printed. It was Tim Hwang who likened today's attention market to subprime, and what he had in his sights was the bubble in targeted advertising's efficacy. What we borrow here is not his conclusion. It is the structure.

This direction can only be pointed to with caution, of course. The rise of the news AI chatbot alone is already flat in mature markets. These signals cannot be used as falsifiers. They are closer to corroboration — the kind that, as observations pile up, makes the direction sharper or blurrier.

The trap in summing 48 countries · That rise is the picture you get once 48 markets are summed together, and in mature markets like the US and the UK it is essentially flat year over year. It could be an illusion born of aggregating the samples, or it could mean saturation has already been reached.

And if it stays flat, what is left? The delegation of what-to-see is left. That recommendation took a large part of that share is already a measured fact, and it holds without AI. Even the 7%-to-10% figure, read the other way, says that in both years nine in ten did not get their news from a chatbot that week. Strip out every AI signal and the delegation itself still stands. Whether that delegation reaches the layer of the verdict is the question the AI phase raises. AI does not create the delegation. It only pushes it up one layer.

There are forces cutting the other way. Watermarking and content credentials — marks that travel with a file recording where and how it was made — are spreading across generated output. But the mark only reports what was machine-made; it does not move the verdict itself outside the circuit. About anything unmarked it says nothing at all. In May 2024, when Google's AI summaries failed conspicuously — advising people to put glue on pizza — the company actually pulled back the scope of the summaries. The current in which the authorship of reality passes to the system comes paired with a counterforce shoving it back. That is why this outlook cannot be drawn as a smooth one-way street.

What the Matrix did was diagnosis, not prophecy. The subject of the diagnosis is the film; the subject of the forecast is us.

Where this looks like self-contradiction · Earlier we said not to grade science fiction as prophecy, and here we read the Matrix as a diagnosis that saw ahead and forecast the future from it. The subject differs. The film read, philosophically, a condition already stirring around its release. Google was founded in 1998, search ads began selling in 2000, and the dotcom bubble peaked that spring. The judgment about where that condition goes is ours. The one that saw ahead was never the film. It was the trajectory of the condition. The power of good science fiction is not in landing a few hits. Orwell's Nineteen Eighty-Four, too, was a diagnosis of 1949 and not a literal prophecy about the year 1984.

After the Awakening

The "desert of the real" that came earlier in Morpheus's line has an original author, and that author, Baudrillard, complained openly that the Matrix had misread him. The film, he said, cleaves the red pill from the blue, the real from the fake, too tidily — presupposing that a reality to return to sits waiting on the outside. As he saw it, in a state where copies beget copies with no original, there is no original to return to and no outside to begin with.

We pay the price and move on. If he is right, this piece's ending collapses with him. With no original to return to, the outside against which a verdict could be checked goes as well, and the procedure this piece means to end on loses the floor it stands on. So we narrow the claim here. What this piece argues for is not a return to the original. It is keeping the procedure that decides what counts as real on our side. That procedure does not require a true world laid out separately on the outside. It only fails if the procedure itself comes out of the same circuit as the thing it judges. Turn that regress one more time and we, holding the procedure, are inside the same circuit too. This piece stops there. Accept even that and no verdict is possible at all — and this is a working assumption, not a refutation.

The red pill is usually read as a symbol of liberation. Swallow it and you wake; wake and you walk out of the Matrix. "I took the red pill" once circulated widely enough to become an internet meme — that is how familiar the schema is. Yet the film itself tears the schema down. Cypher is a comrade who has already gotten out of the Matrix, and even so he cannot bear that truth and moves to go back to the illusion. Knowing the steak is fake, he says the fake is better. What the red pill handed Neo was not freedom either. It was a desert-like reality. It was so even in Plato's cave. The one who breaks the chains and sees the sun does not stay there but returns to the cave. Awakening was never a destination but the start of a round trip.

"But I turned the feed off." Plenty of people have, and the scale runs far past the odd habit of a few. Across 48 markets an average of 42% said they deliberately avoid the news sometimes or often. Pass over these people without calling them up and the argument is left undefended.

What they take back is where the story splits. Those who said they avoid the news were asked how. More than half said they scroll past the news or change the channel when it comes on. About as many said they cut how often they check by turning notifications off, and roughly a third said they steer clear of particular topics. All three are acts that reduce exposure. None is an act of weighing what is true and filtering accordingly. The layer of sense and attention gets taken back; the layer of the verdict stays delegated.

People also work on the feed. About 65% of those under 35, and about 55% of those older, have tried to influence what shows up by unfollowing, muting, blocking, or changing settings. And the motive they gave was to make it more trustworthy, less toxic, more varied in viewpoint. So the previous paragraph's line — that the layer of the verdict stays delegated — was too strong. People reach into that layer too.

Except every place they reach is inside the system. Unfollow, mute, block, settings: each is a dial the system handed out. They are not setting the standard for what to believe outside it; they are rearranging what shows up, using those dials. And that is exactly as far as the survey asked. Comparing against another outlet, or going all the way to the source document, is not caught here. So what is missing is not the yardstick itself. It is a yardstick placed within reach, by default.

As-of dates for the avoidance figures · The 42% average across 48 markets is a 2026 survey value. The 2025 survey reported the same measure at 40%, and the 2017 survey at 29%. The 2026 report describes this as roughly level with the prior year, which is 2 percentage points off its own 2025 value, so we use the level only and leave the trend out. The breakdown of avoidance methods and the feed-tinkering rates are 2023 survey values.

Our own case is no different. Let us fix what "awakening" means here, once and for all: knowing as information. Knowing by having read sentences like "the algorithm picks your gaze" and "in a free service the product is you." That is the red pill we swallowed. And knowing of that kind does not release the delegation. There is corroboration. People answer that they prefer longer videos while actually watching short ones. And the time spent inside those apps breaks its own record year after year.

The limits of these numbers · They are population-wide aggregates, not a measure taken on awakened people alone. They are circumstance that lights up a direction, not evidence proving a claim about any individual. How the attempt to explain this gap as addiction or weak will collapses in the face of the data was traced in the short-form piece.

It is not that we cannot leave. There was never a door shutting us in to begin with. We have delegated the mediation of how we reach the world to the system, and that delegation can be taken back at any time. It is only that the way we stand before it is already, to some degree, set. Freedom is real, but it is not unconditional freedom; it is freedom inside conditions. Even awake, we do not choose to take that delegation back. Knowing, we leave it delegated. Awakening does not release the delegation on its own.

So the red pill is not a pill of resignation either. What the red pill really hands over is not somewhere outside the Matrix. It is the awareness that what we take to be real can be a question — that the verdict was delegated to the system, and that it can be brought back. Even so, that reclaiming takes effort, and there is no sign yet that many have done it. Awakening is not the end but the beginning.

What We Counted Wrong

For 25 years what we counted was the prison of the senses. It did not get built. What has been going up in the meantime is the layer of the verdict. The delegation of what-to-see is already measured, and it is moving toward the seat where what is real gets decided. And even awake, we do not choose to take that delegation back. That is because this delegation conceals itself. Even the yardstick for retracing it comes, by default, out of the same circuit. So the lever is not in an individual's resolve. If firmer resolve were all it took, the 65% and the 55% who reached in to make their feeds more trustworthy would already have taken it back. What they lacked was not will. It was a yardstick within reach, held in a hand other than the one being judged. Just as a few opened the loan ledgers themselves, that road has never been closed. It has only always cost more. So the place to work is not cutting off the tool but lowering that price. Marks, verifiers, windows — the standard is one. Do the interests of the hand that holds it run against the thing being judged? The red pill is already swallowed. What remains is not going back to ignorance. It is putting the procedure that decides what we will call reality into a hand other than the one being judged.