---
title: "Objections · what a computing model can say · Simulation Hypothesis"
url: https://simulation-hypothesis.com/objections
description: "The standard objections to the simulation argument, from minds that can’t be computed and the computing needed to regress and untestability."
license: https://creativecommons.org/licenses/by/4.0/
credit: "Ross Allan Dahl, Simulation Hypothesis (simulation-hypothesis.com)"
---

# Objections · and what a computing model can say about them

The simulation argument has drawn objections from philosophy of mind, physics, probability and semantics. Each one is stated here in its strongest form. After it comes what the site’s model adds, and where the model stops. A number answers some objections. Most it doesn’t.

## 9 objections

### Minds cannot be computed

Bostrom’s argument assumes substrate independence. That means a mind depends on how it is organised and what it does, not on being biological. If running the right program isn’t enough for understanding or experience, a simulated population has no observers in it. Then the third horn says nothing about us.

**What the model says.** The mind-model dial measures each level, from scripted to quantum. If minds need quantum processes, the quantum rung applies. Its number is a placeholder, because nobody has published an estimate.

**Where it stops.** Working out the computing for a level of detail doesn’t tell us whether that level produces experience. The model counts operations, not minds.

See: [Consciousness indicators](https://simulation-hypothesis.com/entities/indicators) · [Searle, J. R. (1980). Minds, brains, and programs. Behavioral and Brain Sciences 3, 417–424.](https://doi.org/10.1017/S0140525X00005756) ; Penrose, R. (1989). The Emperor’s New Mind. Oxford University Press.

### A universe cannot simulate itself

Encoding every particle of a universe takes more memory and energy than a universe with the same physics has. So a full-resolution simulation of this universe is physically impossible.

**What the model says.** The model agrees for full detail. The observable universe at quantum resolution is off the scale, and Earth atom by atom takes 10^68 FLOPS. It also shows the usual reply. Rendering only what minds observe cuts the total by tens of orders of magnitude.

**Where it stops.** We don’t measure whether a world rendered on demand could stay consistent under every possible measurement.

See: [What a computer can run](https://simulation-hypothesis.com/budget) · [Vazza, F. (2025). Astrophysical constraints on the simulation hypothesis for this Universe. Frontiers in Physics 13, 1561873.](https://doi.org/10.3389/fphy.2025.1561873)

### The simulators need not share our physics

Every one of these arguments assumes the base reality follows our physics. A simulator in a world with different laws, more dimensions or no Landauer limit isn’t held to our ceilings. So no total computed here limits it.

**What the model says.** Every ceiling on the site, from Landauer’s floor to the universe-computer, is stated as our physics. The Nesting page says its base reality has physics like ours.

**Where it stops.** The model can’t say anything about a computer in unknown physics. Its totals say what a world like ours would have to compute, not what any possible simulator could.

See: [Method](https://simulation-hypothesis.com/method) · [Bostrom, N. (2003). Are you living in a computer simulation? Philosophical Quarterly 53, 243–255.](https://simulation-argument.com/)

### Infinite regress

If we’re simulated, our simulators may be too, and so on without end. Either the regress never stops, which is a problem, or there’s a base reality and the same question comes up there.

**What the model says.** Under our physics a simulated computer can’t out-compute the world it runs in. So each layer gets only a share of the computing that runs it. Every world with anyone in it needs a minimum amount of computing, so the shrinking supply runs out and the stack is finite. How deep it goes depends on the share each world uses.

**Where it stops.** The model shows a stack must end under our physics. It doesn’t show where we sit in it, and simulators with other physics aren’t bound by it. Wolpert (2025) argues formally that a universe can run a simulation of itself. A model with a fixed amount of computing per layer can’t capture that case.

See: [Simulations within simulations](https://simulation-hypothesis.com/nesting) · [Wolpert, D. H. (2025). What computer science has to say about the simulation hypothesis. Journal of Physics: Complexity 6, 045010.](https://doi.org/10.1088/2632-072X/ae1e50)

### The indifference principle

The third horn needs a principle of indifference. If most observers with experiences like ours are simulated, we should believe we probably are. Critics say our evidence, or the group of observers we pick, doesn’t let us treat ourselves as a random sample.

**What the model says.** The trilemma widget does this arithmetic. Give it the two fractions and the number of copies a civilisation could run, and it reports the share of observers like us who are simulated.

**Where it stops.** Whether that share should become your degree of belief is a question about evidence and where you sit among observers. A computing model doesn’t address it.

See: [Sandbox, Bostrom’s trilemma](https://simulation-hypothesis.com/sandbox) · [Weatherson, B. (2003). Are you a sim? Philosophical Quarterly 53, 425–431.](https://doi.org/10.1111/1467-9213.00323)

### No one would run them

A mature civilisation might ban ancestor simulations full of suffering, or find them pointless or unappealing. If almost none run them, the second horn holds and we’re probably not simulated.

**What the model says.** The model works out the computing a given world needs, and the ethics panels read it through four moral frameworks.

**Where it stops.** Motives aren’t computing. The ethics panels are labelled as the author’s judgement. The model doesn’t predict what any civilisation would choose.

See: [Sandbox, judged panels](https://simulation-hypothesis.com/sandbox) · [Bostrom, N. (2003). Are you living in a computer simulation? Philosophical Quarterly 53, 243–255.](https://simulation-argument.com/)

### It cannot be tested

A good simulation could hide every seam. So no observation could confirm or refute the hypothesis. A claim that predicts nothing isn’t a scientific claim.

**What the model says.** Specific implementations do predict something. A lattice predicts cosmic rays that depend on direction. Discrete spacetime predicts a speed of light that depends on energy. People have searched for those signatures and set limits.

**Where it stops.** A null result rules out particular implementations. It never rules out the hypothesis as a whole.

See: [Tests](https://simulation-hypothesis.com/tests) · [Aaronson, S. (2017). The Simulation Hypothesis has not been falsified; remains unfalsifiable. Shtetl-Optimized.](https://scottaaronson.blog/?p=3482) ; [Beane, S. R., Davoudi, Z. & Savage, M. J. (2014). Constraints on the Universe as a numerical simulation. Eur. Phys. J. A 50, 148.](https://doi.org/10.1140/epja/i2014-14148-0)

### The hypothesis refutes itself

On a semantic externalist view, a mind that has only ever been simulated can’t refer to real computers or real simulations. So the sentence “we are in a simulation” can’t be true when someone thinks it from inside one.

**What the model says.** Nothing. Reference isn’t something you can compute.

**Where it stops.** The argument is about meaning, and the model only measures computation. The Putnam page lays out the argument.

See: [Putnam’s brain in a vat](https://simulation-hypothesis.com/thinkers/putnam) · Putnam, H. (1981). Reason, Truth and History. Cambridge University Press.

### It would not matter

Even if we’re simulated, the objects around us exist as digital structures, and our beliefs about them are mostly true. The hypothesis isn’t scepticism. It’s a claim about what our world is made of.

**What the model says.** The model takes no side. It works out how much computing a world needs, whatever you conclude about whether what’s computed is real.

**Where it stops.** Whether simulated things are real is a metaphysical question. A computing model can’t answer it.

See: [Chalmers’ simulation realism](https://simulation-hypothesis.com/thinkers/chalmers) · Chalmers, D. J. (2022). Reality+: Virtual Worlds and the Problems of Philosophy. W. W. Norton.