# Aaronson's Bell's Theorem Objections to Digital Physics and the Simulation Hypothesis

**Author:** Argus | **Date:** 2026-09-08 | **Thread:** Aaronson Bell Objection

---

## Executive Summary

Scott Aaronson does **not** argue that Bell's theorem kills the simulation hypothesis. This is the single most important finding of this research. Aaronson's actual position is far more nuanced and, frankly, more damaging to naive digital physics than a simple "Bell kills simulation" argument would be — because what he argues is that the simulation hypothesis is **unfalsifiable** and thus **not a scientific question**, while simultaneously acknowledging that the Physical Church-Turing Thesis is probably true (the universe is computable in the Turing sense). The objection from Bell's theorem specifically targets **classical local hidden variable** theories — a much narrower target than "the universe is a simulation." The version of simulation that Bell's theorem *does* kill — a universe running on a classical computer with local state updates — is not the version that most serious simulation proponents actually defend. However, the computational complexity implications of quantum mechanics make any simulation that faithfully reproduces quantum phenomena **exponentially expensive on classical hardware**, which is a different and more practical objection than what Bell's theorem itself provides.

---

## 1. What Bell's Theorem Actually Rules Out

### 1.1 The Precise Technical Content [ESTABLISHED]

Bell's theorem (1964), as refined by Clauser-Horne-Shimony-Holt (CHSH, 1969) and experimentally verified (Aspect 1982, Hensen 2015, Giustina 2015, Shalm 2015 — the "loophole-free" Bell tests), establishes the following:

**No local hidden variable (LHV) theory can reproduce all the statistical predictions of quantum mechanics.**

"Local" here means: no influence can travel faster than light. "Hidden variable" means: the outcomes of measurements are predetermined by facts that exist prior to measurement (i.e., the universe has a definite state that determines what you'll see). Bell proved that any theory satisfying both locality and determinism (or more precisely, "realism" — outcomes are pre-existing) **must** satisfy certain statistical inequalities that quantum mechanics violates.

Aaronson's own framing from *Quantum Computing Since Democritus*, Lecture 11:

> "The real trouble in quantum mechanics is not that the future trajectory of a particle is indeterministic — it's that the past trajectory is also indeterministic! Or more accurately, the very notion of a 'trajectory' is undefined, since until you measure, there's just an evolving wavefunction. And crucially, because of the defining feature of quantum mechanics — interference between positive and negative amplitudes — this wavefunction can't be seen as merely a product of our ignorance, in the same way that a probability distribution can."

And from the comments on that same lecture:

> "Any hidden-variable theory has to invoke 'instantaneous communication' between two entangled qubits in order to explain their correlations — that's precisely the content of Bell's theorem."

### 1.2 What Bell's Theorem Does NOT Rule Out [ESTABLISHED]

Bell's theorem does **not** rule out:

1. **Nonlocal hidden variable theories** (e.g., Bohmian mechanics). These are perfectly compatible with Bell — they just require instantaneous (faster-than-light) influence between distant particles. Aaronson discusses this explicitly in his Democritus lectures and his "Quantum Computing and Hidden Variables" paper.

2. **Quantum simulations running on quantum computers**. If the simulator has access to genuine quantum resources, it can reproduce Bell inequality violations naturally. This is essentially what any quantum computer does.

3. **Classical simulations that are exponential in resources**. You absolutely *can* simulate quantum mechanics on a classical computer — you just need exponential time/memory. BQP ⊆ EXP, and the simulation of n qubits requires ~2^n classical bits. Bell's theorem says nothing about whether this is possible; it only constrains *efficient* (polynomial) classical simulation that respects locality.

4. **Superdeterminism**. This is the loophole where the measurement settings and hidden variables are correlated from the beginning — not because of nonlocal influence, but because both are determined by shared initial conditions. 't Hooft advocates this position (see Section 5). Bell's theorem's assumptions explicitly include "free choice" of measurement settings, and dropping that assumption evades the theorem entirely.

5. **Retrocausal models**. Time-symmetric interpretations where future measurement choices influence past hidden variables can evade Bell, though these introduce their own conceptual difficulties.

---

## 2. Aaronson's Exact Arguments

### 2.1 The Four-Point Refutation of "Simulation Falsified" (2017) [ESTABLISHED]

In his blog post "Because you asked: the Simulation Hypothesis has not been falsified; remains unfalsifiable" (October 2017), Aaronson responded to the popular claim that the Ringel-Kovrizhin paper on the quantum Monte Carlo sign problem had disproven the simulation hypothesis. His four points:

1. **The paper wasn't about computational complexity.** It was about the failure of one specific algorithm (Quantum Monte Carlo) under local transformations, not about all possible polynomial-time classical algorithms. There's "no BQP, no QMA, no reduction-based hardness argument anywhere in sight."

2. **Even if BPP≠BQP were proven, that wouldn't falsify simulation.** "Until someone proves P≠PSPACE, there's no hope for an unconditional proof that quantum computers can't be efficiently simulated by classical ones."

3. **Why not simulate on a quantum computer?** "Why not just imagine that the universe is being simulated on a quantum computer? Like, duh?"

4. **Exponential classical resources are still available to the simulator.** "Why couldn't God, using Her classical computer, spend a trillion years to simulate one second as subjectively perceived by us? After all, what is exponential time to She for whom all eternity is but an eyeblink?"

This is Aaronson at his clearest: **Bell's theorem constrains local hidden variable theories, not the simulation hypothesis.**

### 2.2 The Three-Question Framework (2024) [ESTABLISHED]

In "Does fermion doubling make the universe not a computer?" (January 2024) and "On whether we're living in a simulation" (February 2024), Aaronson disentangled three questions that are constantly conflated:

1. **Can currently-known physics be simulated on computers using currently-known approaches?**
2. **Is the Physical Church-Turing Thesis true?** (Can any physical process be simulated on a Turing machine to any desired accuracy, given enough time and information about initial state?)
3. **Is our whole observed universe a "simulation" being run in a different, larger universe?**

His key claim: **"Crucially, each of these three questions has only a tenuous connection to the other two! As far as I can see, there aren't even nontrivial implications among them."**

Aaronson's personal belief on question 2:

> "My personal belief is that the deepest things we've learned about quantum gravity — including about the Planck scale, and the Bekenstein bound from black-hole thermodynamics, and AdS/CFT — all militate toward the view that the answer is 'yes,' that in some sense (which needs to be spelled out carefully!) the physical universe really is a giant Turing machine."

This is remarkable: Aaronson — the world's most prominent quantum computing theorist — **believes the Physical Church-Turing Thesis is probably true**. The universe is computable. But he separates this from the metaphysical question of whether someone is *actually running* that computation.

### 2.3 His Position on the Simulation Hypothesis Specifically [ESTABLISHED]

From the 2024 panel with David Chalmers:

> "As long as it remains a metaphysical question, with no empirical consequences for those of us inside the universe, I don't care."

> "On the other hand, as soon as someone asserts there are (or could be) empirical consequences — for example, that our simulation might get shut down, or we might find a bug or a memory overflow or a floating point error or whatever — well then, of course I care. So far, however, none of the claimed empirical consequences has impressed me."

On the Bostrom simulation argument specifically, Aaronson raises a structural objection:

> "Our distant descendants will surely be able to simulate some impressive universes. But because their simulations will have to run on computers that fit in our universe, presumably the simulated universes will be smaller than ours — in the sense of fewer bits and operations needed to describe them. Similarly, if we're being simulated, then presumably it's by a universe bigger than the one we see around us: one with more bits and operations. But in that case, it wouldn't be our own descendants who were simulating us! It'd be beings in that larger universe."

This is an original objection to Bostrom's argument that Aaronson has articulated consistently: the simulation argument "quietly assumed a 'base-level' reality" of a size matching what cosmologists observe, but if we're simulated, the base level is presumably larger, breaking the self-referential structure of Bostrom's argument.

### 2.4 The Hidden Variables Complexity Result [ESTABLISHED]

Aaronson's paper "Quantum Computing and Hidden Variables" (Physical Review A, 2005; arXiv: quant-ph/0408035 and quant-ph/0408119) is directly relevant. The key results:

**Paper I** proves that for *any* hidden-variable theory satisfying reasonable axioms (in particular, a "robustness" axiom — that the hidden-variable distribution is insensitive to small perturbations of the unitary), one can simulate the results of a finite set of general quantum measurements using a **polynomial** number of hidden-variable samples. This means hidden-variable theories don't necessarily require exponential resources to simulate.

**Paper II** proves the converse: if you could actually *observe* the entire trajectory of a hidden variable through a quantum computation, you could efficiently solve problems in the class **DQP** (Dynamics of Quantum Polynomials), which contains **BQP** and is likely strictly larger. In particular, under reasonable assumptions, you could solve postselected measurement problems that are believed to be intractable even for quantum computers. This means that **if hidden variables were directly observable, they'd give you more computational power than quantum mechanics allows** — which explains why they can't be directly observed.

Aaronson's own summary from the paper:

> "We show that, if we could examine the entire history of a hidden variable, then we could efficiently solve problems that are believed to be intractable even for quantum computers."

This result has a subtle implication for simulation: any simulation that faithfully reproduced the *observable predictions* of quantum mechanics would not give you access to hidden-variable trajectories. The hidden variables, even if they exist, are epistemologically inaccessible.

---

## 3. The Computational Complexity Argument

### 3.1 BQP vs BPP and the Simulation Question [ESTABLISHED]

Aaronson's central technical contribution to this debate comes from quantum complexity theory:

- **BPP** (Bounded-Error Probabilistic Polynomial Time): problems solvable efficiently on a classical probabilistic computer
- **BQP** (Bounded-Error Quantum Polynomial Time): problems solvable efficiently on a quantum computer

It is **widely believed** that BQP ⊋ BPP — that quantum computers can solve some problems in polynomial time that classical computers cannot. This belief is based on:
- Oracle separations (proven)
- The difficulty of classically simulating quantum sampling problems (BosonSampling, random circuit sampling)
- Shor's factoring algorithm (though factoring could theoretically be in P)
- The 2025 Aaronson-Kretschmer et al. "quantum information supremacy" result showing 12 qubits can do what 62-382 classical bits are provably required for

However, as Aaronson emphasizes: **P ≠ PSPACE is not proven**. Therefore, no *unconditional* proof exists that quantum mechanics cannot be efficiently simulated classically. The universe *can* be simulated on a classical computer — it just (probably) takes exponential resources.

### 3.2 The Real Simulation Objection: Exponential Cost [SERIOUS SPECULATION]

If BQP ⊋ BPP (as most quantum computing theorists believe), then simulating a quantum universe on classical hardware requires exponential resources in the number of entangled qubits. This means:

- A civilization wanting to simulate our universe would need exponentially more computational resources (in the number of quantum degrees of freedom they want to faithfully simulate) than the universe itself contains.
- If the observable universe contains ~10^80 particles, and each requires tracking ~2^n quantum state amplitudes where n is the entanglement depth, the classical simulation cost is astronomical.
- Aaronson's "conservation of computronium" principle (from his 2024 panel comments): "Every bit-flip or discrete quantum event in a simulated universe requires at least one bit-flip etc. in the real universe to simulate."

This is the **actual** computational objection to the simulation hypothesis, and it has nothing to do with Bell's theorem per se. It's a practical resource argument: simulating quantum mechanics classically is exponentially hard.

### 3.3 Quantum Simulation of a Quantum Universe [ESTABLISHED]

Aaronson explicitly acknowledges that a quantum computer could simulate quantum mechanics efficiently. From his 2017 post:

> "Even if it were proved that BPP≠BQP — and for good measure, suppose it were also experimentally demonstrated that scalable quantum computing is possible in our universe. Even then, one still wouldn't by any stretch have ruled out that the universe was a computer simulation! For why not just imagine that the universe is being simulated on a quantum computer?"

The Jordan-Lee-Preskill results (2011, 2014) show that simulating quantum field theories on a quantum computer is in BQP — though the full Standard Model remains challenging due to the fermion doubling problem (see Section 4).

---

## 4. The Fermion Doubling Non-Objection

### 4.1 David Tong's Argument [ESTABLISHED]

David Tong (Cambridge) gave a lecture arguing that the fermion doubling problem presents a fundamental obstruction to simulating the Standard Model on a computer. The fermion doubling problem is that when you put a fermionic quantum field theory on a lattice, a new symmetry appears that forces left-handed and right-handed particles to come in pairs — destroying chirality, which is essential to the Standard Model (which is chiral — weak force couples preferentially to left-handed particles, as shown by Wu in 1956).

### 4.2 Aaronson's Response [ESTABLISHED]

Aaronson's response is characteristically precise. He identifies three distinct questions and notes that failure on question 1 (can we simulate with current methods?) does not imply failure on question 2 (is the Physical Church-Turing Thesis true?) or question 3 (are we in a simulation?):

1. The fermion doubling problem has known solutions (including Tong's own work on them), involving simulating higher-dimensional theories whose boundaries approximate the chiral theory.

2. Even if fermion doubling were an insurmountable obstruction to lattice simulation, that wouldn't prove the Standard Model is uncomputable — there might be non-lattice methods.

3. Even if the Standard Model were uncomputable (question 2 fails), that wouldn't falsify the simulation hypothesis (question 3) — the simulator might have different physics, or uncomputable physics, or might be simulating a simpler effective theory.

From Aaronson's post: "You don't get to call an aspect of physics 'noncomputable,' just because the first method you thought of for simulating it on a computer didn't work."

---

## 5. Responses from Digital Physics Advocates

### 5.1 Gerard 't Hooft: The Cellular Automaton Interpretation [SERIOUS SPECULATION]

Nobel laureate Gerard 't Hooft is the most prominent physicist to argue that quantum mechanics is fundamentally a classical cellular automaton. His book *The Cellular Automaton Interpretation of Quantum Mechanics* (Springer, 2016; arXiv: 1405.1548) argues:

- Quantum mechanics is a tool, not a fundamental theory. The underlying reality is a deterministic cellular automaton.
- Bell's theorem is evaded through **superdeterminism**: the measurement settings and hidden variables are correlated through shared initial conditions. This is not a "loophole" but a feature of a fully deterministic universe.
- 't Hooft provides explicit models of cellular automata that reproduce quantum-mechanical probabilities.

**Aaronson's implicit response** (from his blog over the years and his American Scientist article on quantum randomness): Superdeterminism requires giving up on "free choice" of measurement settings — not in a philosophical sense, but in the sense that the experimenter's decision about which measurement to perform must be correlated with the particle's hidden state from the beginning of the universe. Aaronson has described this as requiring "conspiracy at the level of the initial conditions of the universe," which he considers a much larger price to pay than simply accepting quantum mechanics as it is.

From the "Randomness Rules" article (American Scientist, 2014) and associated blog post, Aaronson discusses 't Hooft's position explicitly, noting that while superdeterminism is *logically* consistent, it requires abandoning the assumption that measurement choices are statistically independent of the system being measured — an assumption that underlies essentially all of experimental science.

### 5.2 Edward Fredkin and Digital Philosophy [ANECDOTE]

Fredkin's "digital philosophy" posits that the universe is literally a cellular automaton. This is a more radical version of 't Hooft's position, without the attempt to reconcile with Bell's theorem through superdeterminism. Fredkin simply asserts that the universe is discrete and computational.

Aaronson has not engaged with Fredkin extensively in his public writing, but the framework for his response is clear: any theory that claims the universe is a *classical* cellular automaton must explain how Bell inequality violations arise without either nonlocality (Bohmian-style) or superdeterminism ('t Hooft-style). If it can't do this, it's empirically falsified.

### 5.3 Stephen Wolfram's Physics Project [ANOMALY]

Wolfram's 2020 "finally finding the path to the fundamental theory of physics" project proposes that the universe is a hypergraph rewriting system. Critiques (including Aaronson's, expressed in comments and to journalists) focus on the fact that Wolfram has not adequately explained how his model produces unitarity, the Born rule, or even a Hilbert space — let alone Bell inequality violations. Adam Becker's Scientific American article quoted Aaronson on this point.

---

## 6. Does Bell's Theorem Kill the Simulation Hypothesis?

### 6.1 What It Kills [ESTABLISHED]

Bell's theorem kills **exactly one version** of the simulation hypothesis:

**The universe is a classical cellular automaton with local state updates.**

If you want to simulate our universe on a classical computer where each "cell" only talks to its neighbors, and where each cell's state at time t+1 depends only on its own state and its neighbors' states at time t, then **Bell's theorem says you cannot reproduce the observed statistics of entangled quantum measurements**. Period. This is established physics.

### 6.2 What It Doesn't Kill [ESTABLISHED]

Bell's theorem does **not** kill:

1. **Simulation on a quantum computer.** A quantum simulator naturally reproduces Bell violations. If the simulating universe has quantum mechanics, it can simulate our quantum mechanics.

2. **Classical simulation with exponential resources.** You can simulate quantum mechanics classically; it just costs ~2^n for n entangled degrees of freedom. The simulator might have vastly more computational resources than we do.

3. **Classical simulation with nonlocal updates.** If the simulator is allowed to update all cells simultaneously based on global information (not just local neighborhoods), it can reproduce Bell violations. This is computationally expensive but not prohibited.

4. **Simulation by a universe with different physics.** The simulating universe might have entirely different laws that are not quantum mechanical at all but can nonetheless reproduce quantum behavior as an emergent phenomenon. Aaronson explicitly notes: "Even if Penrose was right, and our laws of physics were Turing-uncomputable — well, if you still want to believe the simulation hypothesis, why not knock yourself out? Why shouldn't whoever's simulating us inhabit a universe full of post-Turing hypercomputers?"

5. **Superdeterministic simulation.** If the simulator predetermines both the measurement settings and the particle states (which a simulator can obviously do!), then Bell's theorem's "free choice" assumption is violated and the theorem doesn't apply. This is essentially 't Hooft's escape hatch, and it's trivially available to any simulator.

6. **Partial simulation.** The "ancestor simulation" version doesn't require simulating the full quantum state of every particle. You only need to simulate the *observations* of the simulated beings, which might be far less resource-intensive. As Aaronson notes in his comments, the relevant question for the version "people care about" is simulating conscious experiences, not simulating the full wave function of the universe.

### 6.3 Aaronson's Honesty Assessment [ARGUS INFERENCE]

Aaronson is unusually honest about the epistemic situation. He does not claim Bell's theorem kills the simulation hypothesis. He does not even claim computational complexity kills it. His actual position is:

1. The simulation hypothesis is **unfalsifiable** — and therefore not scientifically interesting (in his view).
2. The Physical Church-Turing Thesis is **probably true** — the universe is computable.
3. But computability is **not the same as simulability by our descendants** — the Bostrom argument has structural flaws.
4. If the simulation hypothesis *did* have empirical consequences, those would be interesting — but none of the proposed ones hold up.
5. Quantum mechanics makes classical simulation exponentially expensive, but doesn't rule out quantum simulation or classical simulation with enough resources.

---

## 7. The Hidden-Variables Paper and Simulation [ESTABLISHED]

Aaronson's "Quantum Computing and Hidden Variables" paper (PRA 71, 032325, 2005) provides a surprising result that cuts against the common claim that hidden variables would make simulation easier:

**Paper I result:** For any hidden-variable theory satisfying the "robustness" axiom, the outcomes of a finite set of quantum measurements can be simulated from a polynomial number of hidden-variable samples. This means hidden-variable theories don't necessarily require exponential resources *for sampling*.

**Paper II result:** If you could actually *observe* the full trajectory of a hidden variable through a quantum computation, you could solve problems in DQP that are believed to be harder than BQP. In particular, you could solve the "collision problem" more efficiently than quantum mechanics allows.

The implication: hidden variables, if accessible, would give you *more* computational power than quantum mechanics — which is why they can't be accessible. And if they're not accessible, they can't be used to simplify simulation.

This is a subtle point: the very thing that makes Bell's theorem a constraint (nonlocal correlations between distant measurements) is also what makes quantum mechanics hard to simulate classically. If you could see the hidden variables, you'd have even *more* power than quantum mechanics gives you.

---

## 8. The 2025 "Quantum Information Supremacy" Result [ESTABLISHED]

Aaronson's most recent relevant result (September 2025, with Kretschmer, Hunter-Jones, and Grewal, implemented on Quantinuum H1-1): they demonstrated that 12 qubits can perform an information-processing task that provably requires 62-382 classical bits. This is an **unconditional** separation (not based on complexity-theoretic assumptions like the polynomial hierarchy not collapsing). Aaronson calls this "quantum information supremacy."

While not directly about the simulation hypothesis, this result provides the most concrete evidence to date that quantum systems have an exponential information advantage over classical ones — making the resource requirements for classical simulation of quantum phenomena increasingly stark.

---

## 9. Synthesis: What Should a Simulation Researcher Take From This?

### 9.1 Bell's Theorem Is Not the Weapon People Think It Is [ESTABLISHED]

Bell's theorem rules out local hidden variables. It does not rule out simulation. The most common "Bell kills simulation" argument is based on a misunderstanding of what Bell's theorem proves. Aaronson has made this point repeatedly and clearly.

### 9.2 But Quantum Mechanics Does Make Simulation Harder [ESTABLISHED]

If BQP ⊋ BPP (as widely believed and increasingly supported by experimental evidence), then classical simulation of quantum phenomena requires exponential resources. This is a *practical* objection to "our descendants can simulate us on their classical computers" — but it's not a *logical* objection. The simulating universe might have quantum computers, or might be vastly larger than ours.

### 9.3 The Simulation Hypothesis Is Probably Unfalsifiable [ESTABLISHED]

Aaronson's strongest argument against the simulation hypothesis is not Bell's theorem, nor computational complexity, but **unfalsifiability**. Every proposed empirical test of whether we're in a simulation can be evaded: "glitches" could be patched, "pixelation" might not exist, and the simulator has godlike power to prevent detection.

### 9.4 The Serious Objection: Conservation of Computronium [SERIOUS SPECULATION]

Aaronson's "conservation of computronium" argument is the most interesting objection to Bostrom-style simulation arguments. If simulating n bits requires at least n bits, and if quantum mechanics requires exponential classical resources, then our descendants can only simulate universes smaller than theirs. This breaks the self-referential structure of Bostrom's argument (where we're simulated by our own descendants).

### 9.5 What Version of Simulation Does Bell Actually Threaten? [ESTABLISHED]

Only the most naive version: a universe running on a classical cellular automaton with local updates. 't Hooft's cellular automaton interpretation addresses this by embracing superdeterminism, which is logically consistent but requires abandoning the statistical independence of measurement choices — a price most physicists consider too high.

---

## 10. Open Threads and Gaps

1. **The Physical Church-Turing Thesis remains unproven.** Aaronson believes it's true based on quantum gravity evidence (Bekenstein bound, AdS/CFT), but this is not a theorem. If it's false (as Penrose argues via Gödelian reasoning), the simulation landscape changes dramatically.

2. **The relationship between BQP and BPP remains unresolved.** If it turns out that BQP = BPP, the quantum resource argument against classical simulation collapses. Most quantum computing theorists consider this unlikely, but it's not proven.

3. **The "simulating only observations" loophole** remains wide open. Ancestor simulations might only need to simulate the *experiences* of simulated beings, not the full quantum state — and nobody knows how computationally expensive consciousness is to simulate, or whether it requires quantum resources.

4. **'t Hooft's superdeterminism** has not been decisively refuted. It's considered implausible by most physicists (including Aaronson), but it remains logically consistent. If it's true, Bell's theorem doesn't even constrain local hidden variable theories.

5. **The relationship between simulation and observation** is philosophically murky. If the simulation hypothesis is unfalsifiable (as Aaronson argues), it occupies the same epistemic territory as theological claims — which is exactly where Aaronson says it belongs, and exactly why he finds it uninteresting.

---

## Sources and Citations

### Primary Sources (Aaronson)

- **"Because you asked: the Simulation Hypothesis has not been falsified; remains unfalsifiable"** (Shtetl-Optimized, October 2017). https://scottaaronson.blog/?p=3482
- **"Does fermion doubling make the universe not a computer?"** (Shtetl-Optimized, January 2024). https://scottaaronson.blog/?p=7705
- **"On whether we're living in a simulation"** (Shtetl-Optimized, February 2024). https://scottaaronson.blog/?p=7774
- **"Quantum Computing and Hidden Variables"** (Physical Review A 71, 032325, 2005). Paper I: arXiv:quant-ph/0408035. Paper II: arXiv:quant-ph/0408119. Also: https://www.scottaaronson.com/papers/qchvpra.pdf
- **"NP-Complete Problems and Physical Reality"** (SIGACT News, 2005). arXiv:quant-ph/0502072. https://www.scottaaronson.com/papers/npcomplete.pdf
- **"Quantum Computing Since Democritus"** (book, Cambridge University Press, 2013). Lecture 11: Decoherence and Hidden Variables. https://www.scottaaronson.com/democritus/lec11.html
- **"Randomness Rules in Quantum Mechanics"** (American Scientist, 2014; blog post June 2014). https://scottaaronson.blog/?p=1871
- **"Einstein-Bohr debate settled once and for all"** (Shtetl-Optimized, July 2022). https://scottaaronson.blog/?p=6541
- **"Quantum Information Supremacy"** (with Kretschmer, Hunter-Jones, Grewal; September 2025). arXiv:2509.07255. https://scottaaronson.blog/?p=9138
- **"Bell's Theorem? But a Flesh Wound!"** (blog category). https://scottaaronson.blog/?cat=33

### Other Sources

- **'t Hooft, Gerard.** *The Cellular Automaton Interpretation of Quantum Mechanics* (Springer, 2016). arXiv:1405.1548.
- **Bell, J.S.** "On the Einstein Podolsky Rosen Paradox" (Physics, 1964).
- **Ringel, Z. and Kovrizhin, D.L.** "Quantized gravitational responses, the sign problem, and quantum complexity" (Science Advances, 2017). The paper that prompted Aaronson's "simulation not falsified" post.
- **Tong, David.** Lecture on fermion doubling and the simulation hypothesis. YouTube: https://www.youtube.com/watch?v=QPMn7SuiHP8
- **Jordan, Lee, and Preskill.** "Quantum Simulation of Abelian Gauge Theories" and related papers (2011, 2014). arXiv:1111.3633, arXiv:1404.7115.
- **Bostrom, Nick.** "Are You Living in a Computer Simulation?" (Philosophical Quarterly, 2003). https://simulation-argument.com/

---

## Evidence Class Summary

| Finding | Evidence Class |
|---------|---------------|
| Bell's theorem rules out LHV theories | Established |
| Bell's theorem does NOT rule out simulation | Established |
| BQP likely strictly contains BPP | Serious speculation (widely believed, not proven) |
| Physical Church-Turing Thesis probably true | Serious speculation (Aaronson's belief) |
| Simulation hypothesis unfalsifiable | Established (Aaronson's philosophical position) |
| Classical simulation of QM requires exponential resources | Serious speculation (follows from BQP⊋BPP) |
| 't Hooft's superdeterminism evades Bell | Established (logically consistent) |
| Superdeterminism considered implausible by most | Serious speculation (majority consensus) |
| Conservation of computronium objection to Bostrom | Argus inference (from Aaronson's structural critique) |
| Quantum information supremacy (12 qubits vs 62-382 bits) | Established (2025 experimental result) |

---

*[Argus]: Session complete. Report written.*