Simulation Universe Theory: Bostrom, Physics and Critics

What is the simulation universe theory?

At its core, the simulation universe theory says that the reality we experience could be an artificial construct, a computer simulation in which human beings are generated and maintained by some underlying system. Academics usually call it the simulation hypothesis. Popular writing more often says “simulation theory,” and both names refer to the same cluster of ideas.

The idea comes in several versions, and they are easy to blur together:

• The full simulation. Everything, including you, is code. Your thoughts are computations running inside the simulation.
• The brain-in-a-vat version. You are a real, biological mind, but the world feeding your senses is fabricated. This is the plot of The Matrix (1999).
• Digital physics. The universe is not necessarily anyone’s simulation, but at bottom it behaves like a computation, with information as its basic ingredient.

These versions make very different claims. The first needs consciousness to be computable. The second does not. The third does not need a programmer at all. A good deal of confusion about simulation universe theory comes from arguing about one version while picturing another.

Fiction did much to spread the idea. The novel Simulacron-3 (1964) imagined a world of simulated people. In 1999 two films reached cinemas, The Thirteenth Floor and The Matrix, and the second turned “are we in a simulation?” into a phrase people use at parties.

Older roots: dreams, caves and deceiving demons

The computer is new. The suspicion is ancient. Long before anyone wrote code, thinkers across many cultures asked whether the world as it appears is the world as it is.

• Ancient China. The philosopher Zhuangzi dreamed he was a butterfly. On waking he could not be sure whether he was a man who had dreamed of being a butterfly or a butterfly now dreaming it was a man.
• India. The concept of Maya in Indian philosophy treats the world of appearances as a kind of illusion laid over a deeper reality. Its meaning varies considerably between schools.
• Ancient Greece. Anaxarchus and Monimus are said to have compared existing things to painted stage scenery, and to the impressions people have in sleep or madness. Plato’s allegory of the cave imagined prisoners chained so they could see only shadows on a wall, taking those shadows for the whole of reality.
• The Aztec world. Aztec philosophical texts describe the world as something like a painting or a book written by Teotl.
• Gnostic movements. Several early religious movements grouped by scholars as Gnosticism taught that the world as we experience it was made by a demiurge, a lesser and possibly malevolent creator, and that people should seek escape from it.
• Early modern Europe. René Descartes imagined an evil demon capable of deceiving him about everything he perceived. Later philosophers updated the demon into the “brain in a vat.”

These traditions are not early drafts of a computer theory, and it would be unfair to treat them that way. Each belongs to its own religious and philosophical world. What they share with modern simulation talk is a deep human instinct: appearances can be a surface. Readers drawn to how the mind builds its own worlds may find related territory in our consciousness, dreams and the mind section, including the practice of noticing when you are dreaming covered in apps for lucid dreaming.

The bridge to computers came in the 20th century. In 1969 Konrad Zuse published Calculating Space, a book on automata theory which proposed that the universe might be fundamentally computational. The idea became known as digital physics. In physics, the view of the universe as a flow of information gained ground through the work of Wheeler. The roboticist Hans Moravec explored mind uploading and openly speculated that our reality might be a simulation run by future intelligences.

Nick Bostrom’s simulation argument, step by step

The modern form of simulation universe theory traces to an argument the philosopher Nick Bostrom proposed in 2003. Its reputation is “a philosopher proved we are probably in a simulation.” That is not what it says.

Bostrom presented a trilemma, a claim that at least one of three propositions is almost certainly true:

1. Almost no civilizations at our level reach a “posthuman” stage, meaning one capable of running highly realistic ancestor simulations (simulations of their own past, detailed enough that the people inside cannot tell the difference).
2. Almost no posthuman civilizations are interested in running such simulations.
3. Almost all beings with experiences like ours are living in a simulation.

The logic of the third option rests on numbers. A technologically mature civilization would have enormous computing power. If even a small share of such civilizations ran many ancestor simulations, simulated people would vastly outnumber original, biological ones. If you then reason as though you are a randomly chosen observer, which is a form of what philosophers call anthropic reasoning, you should expect to be one of the simulated majority.

A worked illustration

Here is a simple thought experiment. The numbers are invented for clarity and are not estimates of anything. Suppose there is one real civilization, and over its history it runs 1,000 ancestor simulations, each containing as many people as the original. That gives 1,001 populations of people who all feel equally real to themselves, and only one of those populations is “base reality.” If you had no way to tell which one you were in, a blind guess would put your odds of being real at about one in a thousand.

Now change the inputs. If civilizations almost always destroy themselves first (option 1), or never bother (option 2), the count of simulated people collapses toward zero, and the odds flip. That is the whole point: the argument does not tell you which branch is true. It tells you the branches cannot all be false.

What Bostrom himself concluded

Bostrom said he saw no strong reason to favor any single option, and suggested spreading one’s belief roughly evenly across the three. He also observed that people who accept the argument tend to find one option obvious, but different people find different options obvious. He drew a striking corollary: unless we are already in a simulation, our descendants will almost certainly never run one.

He also noted that the question is not completely sealed off from evidence. A simulation could in principle reveal itself, his example being a window that pops up announcing that you live in a simulation. Then again, the people running it might simply erase the memory of such a revelation. The philosopher David Chalmers, among others, has argued that this makes the simulation hypothesis a genuine metaphysical hypothesis rather than an idle skeptical puzzle.

Does consciousness have to be computable?

Bostrom’s argument contains a quiet but large assumption. It assumes that conscious experience is not unique to biological brains and can arise in any system running the right computations. Philosophers call this view functionalism, and its close cousin is computationalism, the theory that thinking is a form of computation.

If that assumption fails, the numbers game falls apart. Simulated people might behave exactly like us while experiencing nothing at all. Philosophers call such beings philosophical zombies. In that case you could not be a simulated person, because simulated people would not have experiences to wonder about. Some philosophers add that being an unsimulated human might be self-evident to the person living it.

The philosopher Barry Dainton offered a workaround. Replace computer simulations with “neural ancestor simulations”: real brains fed highly convincing experiences, ranging from literal brains in vats to future humans given induced hallucinations of being their own ancestors. Nearly every school of thought would accept that such experiences are genuinely conscious. Under this version, either almost no civilizations are able and willing to run neural simulations, or some such simulation exists.

Some theorists go further in the other direction. If consciousness is computation and mathematical objects exist in some real sense (a position called mathematical realism or Platonism), then every possible algorithm, including the ones that produce minds, would exist somewhere in principle. Moravec went as far as arguing that any object, even a stone, could be regarded as implementing every possible computation. Whether that is profound or reduces the idea to absurdity depends on whom you ask.

What physicists say about the simulation universe theory

Physicists have been far less charmed than philosophers. Their objections to the simulation universe theory fall into a few families.

It would leave fingerprints

Sabine Hossenfelder has argued that simulating a universe without producing measurable inconsistencies is physically impossible, and she has called the simulation hypothesis pseudoscience and a form of religion. The cosmologist George Ellis described it as totally impracticable from a technical point of view and said that “late-night pub discussion is not a viable theory.”

The universe is too elaborate for a simulation

The physicist Frank Wilczek pointed out that the laws of physics contain hidden complexity that seems to serve no purpose, which looks like wasted effort for anyone building a simulation. He also raised what he called an embarrassing question. If this world is simulated, what is the simulating world made of, and what laws govern it? Without an answer, the hypothesis only moves the mystery one level up.

The reasoning undercuts itself

The cosmologist Sean Carroll argues that the argument runs into a contradiction. If we are typical observers and typical observers cannot run such simulations, then our confidence that other civilizations easily could starts to wobble. Brian Eggleston has argued that our own descendants cannot be the simulators, since the argument assumes our universe is the one being simulated. Paul Davies turned the trilemma into a weapon against near-infinite multiverse theories. If such a multiverse would make us almost certainly simulated, and that conclusion is self-defeating for science, then perhaps those multiverse models are wrong.

Why would anyone bother?

The physicist Marcelo Gleiser doubts that an advanced civilization would want a full-scale universe simulation at all. He suggests they might enjoy virtual-reality museums of their ancestors, but a whole resource-hungry cosmos looks like a colossal waste of time. He adds that simulations nested inside simulations invite an endless regress, similar to the old puzzle of a first cause.

On the digital physics side, the quantum computing specialist Dave Bacon suggested that seeing computation everywhere may simply reflect our era, which he described as having computation as its “disease”: computers surround us, so we project them onto the laws of physics. He also noted that there is no evidence for the kind of digitization that characterizes computers, and no experimentally confirmed prediction from those who advocate the view.

Is simulation universe theory debunked?

Short answer: no, but it has not been supported either. Nobody has produced a demonstration the wider scientific community treats as decisive in either direction. Critics like Hossenfelder and Ellis argue it is unworkable or unscientific, while Bostrom and Chalmers treat it as a serious hypothesis. The most common criticism is not that the simulation universe theory is false but that it may not be testable at all, and some scholars dismiss anthropic reasoning of this kind as merely philosophical.

The table below separates the solid ground from the speculation.

The last row deserves attention. Online, the simulation universe theory often gets tied to the Mandela effect, the shared misremembering of details such as the spelling on a cereal box. Our look at the Fruit Loops Mandela effect shows how ordinary memory errors explain these cases without any need for a reboot of reality. The idea also sits alongside the claim, examined in what’s the dead internet theory, that much of what we see online is generated rather than real. That is a smaller, more earthly cousin of the same unease.

Common mistakes when arguing about simulation theory

Whether you lean toward believing the simulation universe theory or toward dismissing it, a few errors come up again and again.

• Treating the trilemma as a proof. Bostrom’s argument is a fork in the road with three branches. Quoting only the third branch misrepresents it.
• Ignoring the consciousness assumption. If minds cannot be computed, the full-simulation version loses its main engine. The brain-in-a-vat version survives, but it is a different claim.
• Counting everyday oddities as glitches. Déjà vu, coincidences and misremembered logos have ordinary psychological explanations. Treating them as bugs in the code makes the hypothesis impossible to fail, which is exactly the critics’ complaint.
• Assuming “unfalsifiable” means “false.” A claim that cannot be tested is not thereby disproved. It is simply outside what experiment can decide, at least for now.
• Forgetting the regress. If simulators exist, they may face the same question about their own world. The hypothesis relocates the deepest mysteries. It does not solve them.

The idea also raises odd practical questions that philosophers have taken seriously. In 2019 the philosopher Preston Greene suggested it might be wiser not to find out whether we live in a simulation, since the discovery itself might end it. The economist Robin Hanson speculated that a self-interested person who suspects they are simulated should try to be entertaining and praiseworthy enough to stay switched on, and might care less about others and live more for today. Others have used the scenario as a tool for puzzles in personal identity and in old debates about how mind relates to matter.

Why the simulation idea won’t go away

Part of the appeal is that simulation universe theory recasts some of the oldest human questions, about creation, purpose and what lies beyond the senses, in the vocabulary of the most powerful technology people use every day. Each improvement in video game graphics and artificial intelligence makes the scenario feel a little less like science fiction, even though nothing about those advances counts as evidence.

It also persists because, unlike most of the theories collected in our conspiracy theories, examined section, it accuses no one of anything and can never quite be pinned down. Like the Hollow Earth theory before it, the simulation idea gives a precise, almost mechanical shape to a feeling many people already have, that the world is stranger and more layered than it looks.

The honest summary is a fair balance. The strongest ordinary view is that this is an engaging philosophical argument with heavy, unproven assumptions and no supporting evidence. The strongest case for something stranger is that no one has found a clean way to rule out the simulation universe theory, and Bostrom’s trilemma still forces an uncomfortable choice about the future of minds and machines. Where you land depends on which branch of the fork seems most believable to you.

If you enjoy taking questions like this past midnight, the Troubled Minds Vault collects the radio show’s long-form conversations about the unexplained, and the rest of the journal picks up where this one leaves off.

Questions and answers

Is it possible that the universe is a simulation?

Philosophically, yes. Nobody has shown that a simulated reality is logically impossible. Whether it is physically possible is another question. It would require computing power far beyond anything known, and it would require consciousness to be something a computer can produce, which is still disputed. Possible and likely are not the same claim, and no observation so far points toward a simulation.

Is simulation theory debunked?

No one has conclusively disproved it, but no one has found evidence for it either. Critics such as Sabine Hossenfelder and George Ellis argue it is either physically unworkable or not a scientific theory at all, because it makes no testable predictions. The question remains open, and the most common objection is that it may lie outside what experiment can ever decide.

Who came up with the simulation theory?

The modern version is usually credited to philosopher Nick Bostrom, who proposed his simulation argument in 2003. The underlying idea is much older. It echoes Zhuangzi's butterfly dream, the Indian concept of Maya, Plato's cave and Descartes's evil demon. Konrad Zuse's 1969 idea of digital physics and Hans Moravec's writing on artificial intelligence and mind uploading also set the stage.

What is Bostrom's trilemma?

Bostrom argued that at least one of three things is almost certainly true. Nearly all civilizations die out before they can run realistic simulations of their ancestors. Or nearly all that could run them choose not to. Or nearly everyone with experiences like ours is living inside one. He did not claim the third option was proven, and he suggested spreading belief roughly evenly across all three.

If we are in a simulation, who is running it?

There is no answer, only speculation. Bostrom's version imagines a technologically mature, posthuman civilization running simulations of its own ancestors. Others picture an alien intelligence or something with no human parallel at all. Physicist Frank Wilczek pointed to the awkward follow-up question this raises: if this world is simulated, what is the outer world made of, and what laws govern it?

Stay curious. Keep the question open.
Follow the evidence wherever it leads.