# Thread: The Attack on H1 — Carroll and the Philosophical Case Against the Simulation Hypothesis

**Date:** 2026-09-11
**Thread author:** Argus subagent (depth 1)
**Objective:** Assemble the strongest philosophical case AGAINST the simulation hypothesis, in the authors' own words, with exact quotes and URLs. Label every claim: Established / Serious speculation / Anomaly / Anecdote / Inference.
**Sources read directly (primary where possible):** Bostrom 2003 (full text); Weatherson 2003 (full text); Bostrom 2005 Reply to Weatherson (full text); Birch 2013 (full text); Bostrom & Kulczycki 2011 "A Patch" (full text); Bostrom 2009 "Some Explanations" re Brueckner (full text); Carroll 2016 blog post (full text); Carroll arXiv:1702.00850 "Why Boltzmann Brains Are Bad" (sections 5.2–6); Hossenfelder Feb 2021 blog post (full text); Aaronson Oct 2017 blog post (full text); Bostrom FAQ (sections 1–13); Chalmers interview in Philosophy Now (full text); Dainton 2002 "Innocence Lost" (abstract, ch. 1, ch. 20); Ringel & Kovrizhin abstract (via Aaronson's reproduction).

---

## 1. Sean Carroll's argument

### 1a. The "Resolution Conundrum" (2016) — Carroll's own attack on the simulation argument

Carroll's direct assault on the simulation argument is NOT the "cognitive instability" argument — it is the **resolution conundrum**, a bottoming-out objection aimed at the *typicality* step. Full primary source: https://preposterousuniverse.com/blog/2016/08/22/maybe-we-do-not-live-in-a-simulation-the-resolution-conundrum/

He reconstructs the "standard simulation argument" as seven steps, then adds the observation that the hierarchy bottoms out:

> "even simulated civilizations will have the ability to run simulations of their own. But a simulated civilization won't have access to as much computing power as the one that is simulating it, so the lower-level sims will necessarily have lower resolution. No matter how powerful the top-level civilization might be, there will be a bottom level that doesn't actually have the ability to run realistic civilizations at all."

His reconstruction of the pro-simulation chain (verbatim):

> "We can easily imagine creating many simulated civilizations. / Things that are that easy to imagine are likely to happen, at least somewhere in the universe. / Therefore, there are probably many civilizations being simulated within the lifetime of our universe. Enough that there are many more simulated people than people like us. / Likewise, it is easy to imagine that our universe is just one of a large number of universes being simulated by a higher civilization. / Given a meta-universe with many observers (perhaps of some specified type), we should assume we are typical within the set of all such observers. / A typical observer is likely to be in one of the simulations (at some level), rather than a member of the top-level civilization. / Therefore, we probably live in a simulation."

Then the conundrum:

> "Given the above logic, including the idea that civilizations that have the ability to construct simulations usually construct many of them, we inevitably conclude: **We probably live in the lowest-level simulation, the one without an ability to perform effective simulations. That's where the vast majority of observers are to be found.**"

> "Hopefully the conundrum is clear. The argument started with the premise that it wasn't that hard to imagine simulating a civilization — but the conclusion is that we shouldn't be able to do that at all. **This is a contradiction, therefore one of the premises must be false.**"

And his diagnosis of which premises are the weak ones:

> "Personally I think that premise 1. (it's easy to perform simulations) is a bit questionable, and premise 5. (we should assume we are typical observers) is more or less completely without justification. **If we know that we are members of some very homogeneous ensemble, where every member is basically the same, then by all means typicality is a sensible assumption. But when ensembles are highly heterogeneous, and we actually know something about our specific situation, there's no reason to assume we are typical.** As James Hartle and Mark Srednicki have pointed out [arXiv:0704.2630], that's a fake kind of humility — by asserting that 'we are typical' in the multiverse, we're actually claiming that 'typical observers are like us.'"

**Assessment of the conundrum's logic [Inference]:** the conundrum only bites against the version of the argument that (a) asserts we are *probably* simulated (Bostrom's own conclusion, importantly, is only the *trilemma*, not the assertion of (3)) and (b) requires each simulated civilization to itself spawn many simulations. Bostrom's actual argument hedges the nesting step ("It may be possible for simulated civilizations to become posthuman... may then run their own ancestor-simulations", 2003 p. 253), and the FAQ §6 explicitly addresses the cost of nesting (see §6 below). So Carroll's contradiction is a genuine tension in the *popular* version of the argument, not a formal refutation of Bostrom's trilemma. [Inference]

### 1b. "Cognitive instability" — where Carroll actually uses the notion

The phrase "cognitively unstable" — which the task brief attributes to Carroll's critique of the simulation argument — is actually from Carroll's paper **"Why Boltzmann Brains Are Bad", arXiv:1702.00850** (2017), aimed at Boltzmann-brain-dominated cosmologies, NOT at the simulation argument. Carroll has not, as far as I could locate, applied that term to the simulation hypothesis in a published piece; I searched preposterousuniverse.com, arXiv, and Mindscape episode listings. Report it accurately as follows.

Abstract (verbatim): "A theory in which most observers are of the Boltzmann Brain type is generally thought to be unacceptable, although opinions differ. I argue that such theories are indeed unacceptable: the real problem is with fluctuations into observers who are locally identical to ordinary observers, and their existence cannot be swept under the rug by a choice of probability distributions over observers. **The issue is not that the existence of such observers is ruled out by data, but that the theories that predict them are cognitively unstable: they cannot simultaneously be true and justifiably believed.**"

The core passage, §5.3 "Cognitive instability" (verbatim):

> "A randomly-fluctuating Boltzmann universe puts us, then, in a strange predicament. On the one hand, we use our reasoning skills and knowledge of physics to deduce that in such a cosmos we are probably randomly-fluctuated observers, even after conditioning on our local data. On the other hand, we can also deduce that we then have no reason to trust those reasoning skills or that knowledge of physics."

> "**The randomly-fluctuating universe scenario is therefore self-undermining, or as Albert has characterized similar situations in statistical mechanics, cognitively unstable.** If you reason yourself into believing that you live in such a universe, you have to conclude that you have no justification for accepting your own reasoning. You cannot simultaneously conclude that you live in a randomly-fluctuating universe and believe that you have good reason for concluding that."

And the conclusion (§6, verbatim):

> "we should reject cosmological models that would be dominated by Boltzmann Brains (or at least Boltzmann Observers among those who have data just like ours), **not because we have empirical evidence against them, but because they are cognitively unstable and therefore self-undermining and unworthy of serious consideration.**"

**What the cognitive-instability argument depends on [Established, from the text]:** it depends on the claim that in a BB-dominated cosmology, "everything we think we know about the laws of physics... has randomly fluctuated into our heads" — i.e., the *content of the theory itself* entails that the reasoning by which you would come to believe it is unreliable. This is a *self-referential* epistemic defect: the theory's own truth conditions undermine the justification for believing it. [Inference]

**How the notion could be carried over to the simulation hypothesis [Inference — this is the extrapolation the task asks for, and it must be flagged as an extrapolation]:** the analogous claim would be: if we are simulated, then our empirical evidence (about physics, computation, neuroscience) is itself part of the simulation and may be radically non-veridical, so the empirical premises of the simulation argument — which are *themselves drawn from observations inside the simulated world* — cannot justify its conclusion. Note the crucial difference: in the Boltzmann case the reasoning relies on a theory *about the whole universe one inhabits*; in the simulation case the argument's empirical premises would still be reliable for *prediction inside the simulation* unless the simulators systematically fake them — and if they fake them, the faking is an *external* fact Bostrom's argument can absorb (see FAQ §4 answer, quoted in §6 below). So the transfer is not automatic. [Inference]

Also relevant (Anecdote, from Carroll's own blog): Carroll explicitly does *not* say we're definitely not simulated — the 2016 post ends "I highly doubt this is an original argument... But it illustrates the trickiness of reasoning about who we are in a very big cosmos." His book *Something Deeply Hidden* (2019) discusses the multiverse and observers; I did not locate a sustained simulation-argument critique there — checked via search; no direct quote obtainable. State this as "not located in the time available."

---

## 2. The Bostrom trilemma's weak points

### 2a. The trilemma itself — Bostrom's own statement (2003)
Primary source: https://simulation-argument.com/simulation.pdf (Philosophical Quarterly 53(211): 243–255)

Abstract (verbatim): "This paper argues that at least one of the following propositions is true: (1) the human species is very likely to go extinct before reaching a 'posthuman' stage; (2) any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof); (3) we are almost certainly living in a computer simulation."

The formal result (§IV, verbatim): "Because of the immense computing power of posthuman civilizations, N<sub>I</sub> is extremely large... we can then see that at least one of the following three propositions must be true: (1) f<sub>P</sub> ≈ 0; (2) f<sub>I</sub> ≈ 0; (3) f<sub>sim</sub> ≈ 1."

Closing line (§VII, verbatim): "**Unless we are now living in a simulation, our descendants will almost certainly never run an ancestor-simulation.**"

### 2b. The indexical / reference-class objection — Weatherson, "Are You a Sim?" (2003)

Primary source: http://brian.weatherson.org/sims.pdf (Philosophical Quarterly 53(212): 425–431)

The whole attack is on what Weatherson calls the underlying indifference principle:

> "(##) ∀Φ: Cr(Φ | fΦ = x) = x"

He runs through five interpretations: (1) unrestricted on current credences — false, since "Rat may well know that the proportion of human-like agents that are like spaghetti westerns is rather low, while rationally being quite confident that she likes spaghetti westerns"; (2) restricted to properties Rat doesn't know she has — still false, via his "suman"/"him" construction: Rat knows a unique fact C about herself, so if she's a suman (human-and-C, or Sim-and-not-C), her credence in being suman is forced below 1/4 while she *knows* she is C; (3) further restricted to maximally specific properties — this "will also rule out the applicability of (##) to properties like being a Sim. For Rat knows that if she is a Sim, then she is a Sim that is also a C"; (4) as a constraint on priors — reduces to whether all evidence is independent of being a Sim (P2); (5) no good reason for P2.

The key negative claim (verbatim, conclusion):

> "Nothing I have said here implies that Rat should have a high credence in her being human. But it does make one argument that she should not have a high credence in this look rather tenuous... **If Rat is very confident that she is human, even while knowing that most human-like beings are Sims, she has not violated any norms of reasoning, and hence is not thereby irrational. In that respect she is a bit like you and me.**"

**Why the reference-class objection "bites" [Inference]:** Weatherson's point is that the indifference principle needs a *principled* reference class — "human-like agents" — but any formal way of picking it is either too broad (you have unique self-knowledge C that shifts the class) or too narrow (it excludes Sim too). This is the same structural problem as the "grue" case: there is no "formal distinction between the 'normal' properties... and the 'deviant' properties." The objection does not refute the trilemma; it refutes the *further step* from f<sub>sim</sub> ≈ 1 to "my credence that I'm a Sim ≈ 1." [Inference]

**Bostrom's reply (2005), which the FAQ and secondary literature largely regard as settling the *formal* point** — https://simulation-argument.com/weathersonreply.pdf (Philosophical Quarterly 55(218): 90–97). Bostrom concedes the general principle ## must be restricted and reformulates P2 as:

> "P2*. All our evidence is probabilistically independent of (SIM), after we conditionalize on fsim ≈ x (for some particular x)."

He argues the *whole point of the simulation argument* is that we have evidence relevant to SIM (via f<sub>sim</sub>), and that after conditioning on f<sub>sim</sub> ≈ 1, remaining specific evidence is "at best weak and speculative." On the suman worry he responds that no "deep philosophical difference between predicates like 'suman' and 'non-simulated human'" is needed — "the difference is merely one of practicality." [Established: these are his actual words in the reply.]

### 2c. Birch — selective scepticism (2013)

Primary source: https://simulation-argument.com/pdf/birch-on-the-simulation-argument-and-selective-skepticism.pdf (Erkenntnis 78(1): 95–107)

Birch's framing of the trilemma (H1/H2/H3 notation) and the "bold further step" — CLAIM: Cr(SIM | H3) ≈ 1 — justified by the bland indifference principle BIP: Cr(SIM | f<sub>sim</sub> = x) = x.

The central charge (verbatim, abstract):

> "I first show that the Simulation Argument requires a curious form of selective scepticism, for **it presupposes that we possess good evidence for claims about the physical limits of computation and yet lack good evidence for claims about our own physical constitution.**"

The incompatibility argument (verbatim, §3):

> "There is, however, a fairly obvious tension between Good Evidence and Impoverished Evidence. If my evidence is unable to support the mundane claim that I possess two real human hands, how can I nevertheless have good evidence for exotic claims regarding the fundamental physical limits of computation? We can turn this apparent tension into an outright contradiction by introducing a third assumption: **(Parity of Evidence) My epistemic access to the facts about my own constitution is at least as good as my epistemic access to the facts about the physical limits of computation.** Good Evidence, Impoverished Evidence and Parity of Evidence are jointly incompatible."

Birch shows two attempted escapes fail: (i) rejecting Good Evidence (retreating to a quadripartite disjunction SIM∨H1∨H2∨H3) fails because the analogous BIP* "seems far from intuitively compelling" (the modified-DNA-scenario dilemma); (ii) rejecting Parity of Evidence via "the substrate-independence of computing power" fails because "observations of simulated computers will not reliably indicate the true physical limits of the real computers on which the simulation is running," and the "lower bound" version fails because "the mere appearance of hugely powerful machines could be simulated far more straightforwardly by simulating the experiences of a single observer: no virtual machines would need to be constructed at all." Closing: "There is no good reason to uphold the selective scepticism the Simulation Argument presupposes. There is thus no good reason to believe its conclusion."

### 2d. A genuine formal bug the critics found — Bostrom & Kulczycki, "A Patch for the Simulation Argument" (2011)

Primary source: https://simulation-argument.com/patch.pdf (Analysis 71(1): 54–61)

Bostrom's own side concedes a real mathematical non sequitur in the 2003 formula (verbatim):

> "This article reports on a newly discovered bug in the original simulation argument... **What has so far passed unnoticed is a mathematical non sequitur in the original paper.**" The vulnerability: "if those civilizations that eventually reach a posthuman phase have unusually brief pre-posthuman phases compared to other civilizations, then—since the ancestor simulations only cover the pre-posthuman phase—**it could happen that most pre-posthuman observers live outside simulations even if most pre-posthuman civilizations eventually become posthuman, and even if each posthuman civilization runs several ancestor simulations.**"

They patch it two independent ways (a bounded pre-posthuman-population assumption; an indexical "computer age birth rank" argument). **[Established — this is a documented, author-conceded defect in the original paper's derivation, repaired in 2011.]**

### 2e. Brueckner (2008) — Sims can't really build computers

Original: A. Brueckner, "The Simulation Argument Again", Analysis 68(3): 224–226 (paywalled; I read Bostrom's reply reproducing his argument). Bostrom's account (https://simulation-argument.com/brueckner.pdf, 2009): Brueckner argued, per Bostrom's quotation: "**since a Sim cannot really build a computer—he merely seems to build one within his simulation—it follows that he cannot really create another human-like Sim mind that is instantiated in the programming of a real computer that he really builds**" and "Just as a brain in a vat is incapable of really building another brain in a vat, a Sim is incapable of really building another computer which instantiates another human-like conscious Sim mind."

Bostrom's reply (verbatim): "the idea of stacked virtual machines is actually quite unproblematic... **each computation that any of these virtual machines implements is really being implemented... all of these computations being instantiated by the electrons shuffling through its circuitry.**" Note that in "Some Explanations" Bostrom also stresses Brueckner misquoted "will run" for "may run" (2003 p. 253 actually says "may then run their own ancestor-simulations"). [Established.]

### 2f. The indexical / reference-class objection and the "bland indifference principle" — summary of the state of play [Inference]

- The BIP is rejected by Weatherson (as underdetermined: five readings, none support the conclusion) and by Birch (it requires Impoverished Evidence, which conflicts with Good Evidence + Parity of Evidence). Bostrom's own final position (2005) narrows the claim to P2* — a *weak* posterior-independence claim — which is considerably weaker than the 2003 "bland indifference principle" headline. So the critics did change the argument's shape: what survives is not "indifference", it's "no *specific* evidence discriminates after conditioning on f<sub>sim</sub>≈1, and such evidence as exists is weak and speculative."
- The indexical/reference-class problem bites on the *further step* (from trilemma to "I am probably simulated"), not on the trilemma itself. Weatherson himself accepts the trilemma's earlier steps ("prepared to accept the correctness of the simulation argument up to but not including this last step").

---

## 3. The substrate-independence assumption

### 3a. Bostrom's own hedge (2003, §II) — the assumption is load-bearing and he knows it

Verbatim: "Provided a system implements the right sort of computational structures and processes, it can be associated with conscious experiences... Arguments for this thesis have been given in the literature, and **although it is not entirely uncontroversial, we shall here take it as a given.**" And the attenuated version he actually needs: "We need only the weaker assumption that it would suffice for the generation of subjective experiences that the computational processes of a human brain are structurally replicated in suitably fine-grained detail, such as on the level of individual synapses. **This attenuated version of substrate-independence is quite widely accepted.**" [Established.]

### 3b. Tononi & Koch (2015) — IIT implies simulated minds are not conscious

Primary source: G. Tononi & C. Koch, "Consciousness: here, there and everywhere?", Phil. Trans. R. Soc. B 370(1668): 20140167, https://royalsocietypublishing.org/doi/10.1098/rstb.2014.0167 (journal site Cloudflare-blocked my direct fetch; the abstract is quoted identically across PhilPapers, the publisher's own page, and multiple citations — verified verbatim):

> "Also, in sharp contrast to widespread functionalist beliefs, **IIT implies that digital computers, even if their behaviour were to be functionally equivalent to ours, and even if they were to run faithful simulations of the human brain, would experience next to nothing.**"

[Established as the paper's abstract wording; the paper itself is behind the publisher's block for me, so I could not verify the full context of the claim in the body.]

Related (Anecdote, second-hand but on-record): Tononi in conversation (Closer to Truth): "to be truly conscious, an entity must be 'of a certain kind that can constrain its past and future—and certainly a simulation is not of that kind'." https://loc.closertotruth.com/theory/tononi-s-integrated-information-theory

**Why this matters for Bostrom [Inference]:** Bostrom's trilemma requires N<sub>I</sub> to be "extremely large," which requires that *faithful simulations of brains are conscious* — otherwise simulations don't add observers, f<sub>sim</sub> stays ≈ 0, and disjunct (3) collapses. IIT is one influential theory of consciousness that directly denies the premise. **But note the dialectical situation [Inference]: IIT is itself contested (the 2023 "toy problem" debate — see below), and Bostrom's premise only needs *some* account of substrate-independence to be right; IIT's denial is a rebuttal only if IIT is correct.** Bostrom's explicit escape: he only needs "in fact, a computer running a suitable program would be conscious" to be true — a claim IIT denies but which remains, at present, a live philosophical disagreement, not settled science. [Established: it is a live disagreement; the IIT-vs-functionalist dispute is ongoing.]

Recent IIT-vs-simulation literature (Serious speculation / ongoing research): Findlay, Marshall, Albantakis, David, Mayner, Koch, Tononi (2024 preprint) apply IIT's postulates to "a basic stored-program digital computer programmed to simulate" a target system "with full fidelity" and argue the simulation lacks integrated causation — i.e., IIT's formal machinery *predicts* the simulated system is not conscious (summarized in the IIT Wikipedia article, https://en.wikipedia.org/wiki/Integrated_information_theory). I did not read this preprint directly; flag as secondary.

### 3c. Searle's Chinese Room as applied here

I could NOT locate a published primary-source argument that applies the Chinese Room specifically to Bostrom's simulation argument (searched for "Searle simulation argument Chinese room Bostrom"; the literature cites Searle's 1980 paper as the canonical statement of the underlying point — syntax ≠ semantics — but the application to *Bostrom* is mostly made informally/on blogs). Report honestly: the mapping (Searle's room = the simulation running a brain-model; neither understands, therefore no mind) is a *transposition* [Inference], and its force against Bostrom is exactly the same as its force against functionalism generally — Bostrom's attenuated substrate-independence (synaptic-level structural replication) was explicitly designed to sidestep Turing-test-style behavioral objections, but the Chinese Room attacks the stronger claim that *any* symbol-manipulation could generate consciousness. Since Bostrom only needs the "in fact" reading, the Chinese Room (which is a thought experiment against *sufficiency of syntax*, not an empirical disproof) does not decisively refute him — it relocates the burden onto the empirical question of whether structural replication *in fact* yields experience. [Inference, clearly labeled.]

---

## 4. The computational-cost / physics objections (Ringel & Kovrizhin 2017)

### 4a. What the paper actually claims

Primary source: Z. Ringel & D. L. Kovrizhin, "Quantized gravitational responses, the sign problem, and quantum complexity," Science Advances 3(9): e1701758 (2017), https://www.science.org/doi/10.1126/sciadv.1701758; arXiv:1708.06649. Abstract reproduced verbatim (via the publisher and via Aaronson's full reproduction):

> "It is believed that not all quantum systems can be simulated efficiently using classical computational resources. This notion is supported by the fact that it is not known how to express the partition function in a sign-free manner in quantum Monte Carlo (QMC) simulations for a large number of important problems. The answer to the question—whether there is a fundamental obstruction to such a sign-free representation in generic quantum systems—remains unclear. **Focusing on systems with bosonic degrees of freedom, we show that quantized gravitational responses appear as obstructions to local sign-free QMC.** In condensed matter physics settings, these responses, such as thermal Hall conductance, are associated with fractional quantum Hall effects... The connection between quantized gravitational responses and the sign problem is also manifested in certain vertex models, where TR symmetry is preserved."

**What they claim [Established]:** for certain *bosonic* quantum systems, quantized gravitational responses (e.g. thermal Hall conductance in fractional quantum Hall states) obstruct *local sign-free* quantum Monte Carlo — i.e., one specific, widely used classical simulation algorithm cannot be made sign-free by local basis changes for those systems. That's it.

**What they do NOT claim [Established by inspection of the abstract + careful readings]:** nothing about the universe being unsimulatable, nothing about us not living in a simulation, no claim about all classical algorithms (only "local" transformations within QMC), nothing about quantum computers.

### 4b. The careful readings

Scott Aaronson, "Because you asked: the Simulation Hypothesis has not been falsified; remains unfalsifiable" (Oct 3, 2017), https://www.scottaaronson.com/blog/?p=3482. He quotes the abstract and then lists "four huge difficulties, any one of which would be fatal by itself":

1. The paper is "entirely innocent of the language and tools of complexity theory. There's no BQP, no QMA, no reduction-based hardness argument anywhere in sight... everything is phrased in terms of the failure of one specific algorithmic framework (namely QMC)—and within that framework, only 'local' transformations of the physical degrees of freedom are considered, not nonlocal ones."
2. "even if the paper did address computational complexity head-on, about the most it could've said is that computer scientists generally believe that BPP≠BQP... 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. "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?"
4. "even if, for some reason, we disallowed using a quantum computer... why couldn't God, using Her classical computer, spend a trillion years to simulate one second as subjectively perceived by us?"

Aaronson's verdict (verbatim): "OK, but does any of this prove that the universe isn't a computer simulation, as the popular articles claim (**and as the original paper does not**)? ... if it weren't for all four separate points above, then sure, physicists would have now proved that we don't live in the Matrix."

Bostrom himself, FAQ §12 (https://simulation-argument.com/faq/) — his words on the media coverage:

> "Another example of a proposed physics test derives from the Ringel & Kovrizhin (2017) attempt to show that classical computers cannot efficiently simulate certain quantum systems (the authors focused on limitations of quantum Monte Carlo simulations in the context of systems with certain topological features linked to the quantum Hall effect). **Some media reported this as 'scientists have found proof that we are not living in a simulation!'. The most obvious flaw in that interpretation is that simulators could use quantum computers.** More generally... another way would be to run a simulation that 'cheats'."

### 4c. Zohar Ringel's own statements about the misreporting

**Not located.** I searched for interviews or statements by Ringel responding to the coverage (searched: "Ringel Kovrizhin interview simulation misrepresented/misreported/did not claim", Quora, press coverage). The press releases themselves (EurekAlert, phys.org, 2017) contained the overclaim ("found proof that such a simulation is impossible as a matter of principle" — about simulating a specific condensed-matter system, not the universe). What exists on the record is: (i) the paper's own abstract (which never mentions the simulation hypothesis), (ii) Aaronson's reading, (iii) Bostrom's FAQ characterization. Where I did NOT look: Hebrew-language press, podcast appearances. State this as a gap.

### 4d. Verdict on the physics objection [Inference]

The Ringel-Kovrizhin result — even taken at full strength, as an obstruction to sign-free QMC — is a result about *classical Monte Carlo algorithms for specific bosonic systems*. It does not touch the question of whether a *different* computer (quantum, or a classical machine with more exotic resources, or a "cheating" generator of observer experiences) could simulate our world. As an attack on the simulation hypothesis it **misses**, and as literature it actively *backfires*: it demonstrates how a precise, narrow complexity result gets inflated into a falsification claim (an Anecdote about how this debate is conducted, well documented by Aaronson's post). It is, however, a legitimate attack on the *sloppy popular versions* of the hypothesis that claim "a computer like ours could do it." [Inference.]

---

## 5. The "this is not science" objection — Sabine Hossenfelder

Primary source: "The Simulation Hypothesis is Pseudoscience," Backreaction (blog), Feb 13, 2021, http://backreaction.blogspot.com/2021/02/the-simulation-hypothesis-is.html (transcript of her video). All quotes verbatim.

Opening: "I quite like the idea that we live in a computer simulation. It gives me hope that things will be better on the next level. **Unfortunately, the idea is unscientific.**"

The key argument — the reproduction-of-physics requirement:

> "**The problematic part of Boström's argument is that he assumes it is possible to reproduce all our observations using not the natural laws that physicists have confirmed to extremely high precision, but using a different, underlying algorithm, which the programmer is running.** I don't think that's what Bostrom meant to do, but it's what he did. He implicitly claimed that it's easy to reproduce the foundations of physics with something else."

> "**But nobody presently knows how to reproduce General Relativity and the Standard Model of particle physics from a computer algorithm running on some sort of machine.** You can approximate the laws that we know with a computer simulation – we do this all the time – but if that was how nature actually worked, we could see the difference. Indeed, physicists have looked for signs that natural laws really proceed step by step, like in a computer code, but their search has come up empty handed. **It's possible to tell the difference because attempts to algorithmically reproduce natural laws are usually incompatible with the symmetries of Einstein's theories of special and general relativity.** I'll leave you a reference in the info below the video. The bottomline is, it's not easy to outdo Einstein."

The second issue — the compression claim:

> "Bostrom therefore has to assume that it's somehow possible to not care much about the details in some parts of the world where no one is currently looking, and just fill them in in case someone looks. Again though, he doesn't explain how this is supposed to work. What kind of computer code can actually do that? ... **You cannot in general just throw away physical processes on short distances and still get the long distances right.**"

Her closing verdict: "**The simulation hypothesis, therefore, just isn't a serious scientific argument. This doesn't mean it's wrong, but it means you'd have to believe it because you have faith, not because you have logic on your side.**"

**Assessment of Hossenfelder's argument [Inference]:**
- Strong point: the "fill in as needed" premise is indeed underdeveloped in Bostrom 2003 (he gestures at it; FAQ §6 elaborates with procedural generation and superintelligent patching, but admits the design problem).
- Weak point (as an attack on *Bostrom's* argument): Bostrom's trilemma does not require that a simulation *reproduce GR + SM from an algorithm*; it requires that simulations be *indistinguishable to the observers inside* — and as Bostrom's FAQ §6 notes, simulators can cheat, edit brain states, and use whatever physics the basement universe allows. Hossenfelder's "we could see the difference" assumes the simulation must be a faithful full-physics computation; Bostrom explicitly denies that requirement ("Simulating the entire universe down to the quantum level is obviously infeasible... much less is needed", 2003 p. 5). So her objection lands on the *popular* "the universe is a computation of the equations" picture, not on Bostrom's actual argument. [Inference]
- Her "pseudoscience" label is a claim about the *hypothesis's* testability, to which Bostrom's FAQ §11 answers that the hypothesis is linked to observable evidence (evidence bearing on H1 and H2 shifts credence in H3) — a point she doesn't engage. [Inference]

---

## 6. What cuts the OTHER WAY — defenses of the simulation argument

### 6a. Bostrom's reply to the self-undermining objection (FAQ §4, verbatim)

https://simulation-argument.com/faq/

> "The claim that we cannot have any information about the underlying reality if we are in a simulation is false. In a simulation, we can certainly know the following two conditional claims: **A. If we are in a simulation, then the underlying reality is such as to permit simulations, it contains at least one such simulation, and (3) is true. B. If we are not in a simulation, then the empirical evidence noted in the simulation argument is veridical taken at face value...** Since we either are, or are not, in a simulation, we can conclude that the disjunct (1)∨(2)∨(3) is true."

This is Bostrom's formal answer to the Carroll-style "self-undermining" worry: the argument is structured as a disjunction over both branches, so simulated-untrustworthy-evidence doesn't break it. Note (from the same FAQ §4 and the Reply to Weatherson): Bostrom concedes the specific-evidence point — simulators "could easily prevent us from finding out" (FAQ §5), and the argument only requires weak posterior independence (P2*). [Established: these are his actual positions.]

### 6b. Bostrom & Kulczycki's patch — the argument survives its own bug

Already quoted in §2d. The authors' self-assessment: "There is a technical glitch in the original presentation... **This glitch can be patched in at least two different ways, either of which secures the original conclusion.**" [Established.]

### 6c. David Chalmers, *Reality+: Virtual Worlds and the Problems of Philosophy* (2022) — simulation realism

Primary source located: Philosophy Now issue 148 interview with Paul Doolan, https://philosophynow.org/issues/148/David_Chalmers (the book itself I did not read; the interview is Chalmers speaking in his own words about the book's claims).

- On probability: "I wouldn't say it's 'likely' we're in a simulation, I'd just say that we might be and that we can't rule it out. **In the book I speculate that there is at least a 25% chance.**"
- On "simulation realism" / virtual realism: "Maybe more important is the idea that **virtual reality is genuine reality: that is, both under the simulation hypothesis and in regular VR, the virtual world is just as real as ordinary physical reality.**"
- On the objects in a simulation: "I try to argue for the 'it from bit' idea, that even if we are in a simulation, the objects we are interacting with are real, on a good definition of 'real'. They are digital objects to be sure, but they are in a real world of real experiences."
- On the Cartesian framing: "**Descartes' argument basically assumes that virtual realities are not genuine realities... However, if you accept that virtual realities are genuine realties, it means that Descartes' type of sceptical argument is somewhat harder to get off the ground.**"

[Established: these are Chalmers's words in the interview. Also relevant: Chalmers's earlier "The Matrix as Metaphysics" (2003/2009) is the classic defense of the "creation hypothesis" and is cited approvingly by Bostrom in the Reply to Weatherson (fn. 3).]

**How Chalmers's defense interacts with the attacks [Inference]:** (i) against the "simulation isn't real" intuition (which fuels Hossenfelder's "faith" jab and the "glitch" skeptics), Chalmers argues the *stakes are lower than advertised* — being simulated doesn't make our world, our ethics, or our knowledge illusory; (ii) against substrate-independence worries, Chalmers's functionalism is precisely the position Tononi/Koch deny — the two defenses are mutually exclusive, and the debate between them is unresolved; (iii) Chalmers does *not* defend Bostrom's 50/50-ish priors; he lowers the claim to "we can't rule it out, ~25%," which is a retreat that makes the argument *easier* to defend but weaker as a claim.

### 6d. Dainton (2002) — the other side's own broadening (relevant to the substrate-independence attack)

Primary source: https://simulation-argument.com/dainton.pdf, "Innocence Lost: Simulation Scenarios: Prospects and Consequences" (2002). Abstract (verbatim): "Those who believe suitably programmed computers could enjoy conscious experience of the sort we enjoy must accept the possibility that their own experience is being generated as part of a computerized simulation... **The first part of this paper is devoted to broadening the scope of the argument: even if computers cannot sustain consciousness (as many dualists and materialists believe), there may still be a strong likelihood that we are living simulated lives.**"

[Established — and it cuts against §3: even the *denial* of substrate-independence does not kill the simulation worry, because "hard" simulations (neural tampering) and virtual environments for *non-simulated* minds remain menacing. This is a genuine counterweight to the Tononi/Koch line within the philosophical literature.] [Inference on its bearing.]

---

## 7. VERDICT — does each attack defeat, weaken, or miss?

Scoring convention: **Defeats** = the argument as Bostrom actually states it cannot survive without major repair; **Weakens** = forces concessions/repairs, conclusion stands but shakier; **Misses** = fails to engage the actual argument (usually engaging the popular caricature).

1. **Carroll's Resolution Conundrum (2016).** *Weakens but partly misses.* It exposes a real inconsistency in the *popular* "we are probably simulated, and sims run sims" version, and its typicality critique (Hartle-Srednicki "fake humility") is a genuine and forceful objection to the *typicality step*. But Bostrom's published argument is a *trilemma*, not an assertion of (3); the nesting premise is hedged ("may"); and the FAQ explicitly patches the nesting-cost worry ("simulators could avoid this by stepping in to prevent simulated civilizations from using excessive amounts of computing power"). Net: it weakens the *further step* and the popular version; it does not defeat the trilemma. [Inference]

2. **Carroll-adjacent "cognitive instability."** *Misses as applied to the simulation hypothesis* — Carroll never makes this argument against simulation scenarios in the text I found; it is a Boltzmann-Brains argument whose mechanism (your reasoning's reliability is entailed-away by the very theory you reason with) does not transfer cleanly, because the simulation argument's empirical premises remain reliable *within* the simulation unless simulators fake evidence, in which case Bostrom's disjunction logic (§6a) still absorbs the failure. At most it shows that *some* versions of the simulation hypothesis (e.g., "the whole world, including our evidence about physics, is arbitrary code") are cognitively unstable — but that is not the version Bostrom argues for. [Inference]

3. **Weatherson's indexical/reference-class attack on the Bland Indifference Principle.** *Weakens — permanently.* It forced Bostrom to retreat from a general indifference principle (#) to the weak posterior claim P2*. The trilemma survives; the "I am almost certainly simulated given (3)" step is now hostage to an empirical judgment about simulators' motivations that no one can assess. This is the most durable of the philosophical attacks. [Inference]

4. **Birch's selective scepticism.** *Weakens — seriously, and it is the single sharpest published critique.* The Good Evidence / Impoverished Evidence / Parity of Evidence trilemma is internally valid: Bostrom needs both scientific-realist trust in computation limits and radical doubt about our own physical constitution, with no principled line between them. Birch's two salvage attempts (quadripartite disjunction; substrate-independence of computing power) are shown to fail within his own argument. Bostrom's known replies (P2*, the "special circumstance" story in the Reply to Weatherson §3) do not squarely answer the Parity charge — they restate it. I could not find a published Bostrom reply to Birch specifically; gap. Net: the *empirical* force of the argument (that f<sub>sim</sub> is huge if H1&H2 are false) is untouched, but the *credence constraint* (that you personally must take SIM seriously) is substantially weakened. [Inference]

5. **Bostrom & Kulczycki's own Patch.** *Confirms a real defect, successfully repaired.* Not an attack, but the honest record shows the original 2003 derivation had a genuine mathematical bug; the two patches are themselves additional assumptions. This slightly reduces the argument's elegance but not its conclusion — the authors' own verdict. [Established]

6. **Brueckner's "Sims can't really build computers."** *Misses* — Bostrom's virtual-machines reply ("each computation that any of these virtual machines implements is really being implemented") is correct as a point about computation, and Brueckner's quote is conceded to be a misquote ("may" for "will"). The deeper point Brueckner gestures at — that the *basement* computer bears all the cost of nested sims — is real but is exactly the cost question Carroll raised, and it is addressed (not resolved) by the FAQ's leaf-node argument. [Inference]

7. **Tononi/Koch / IIT substrate-independence denial.** *Weakens, conditionally — the deepest unresolved front.* If IIT (or something like it) is right that digital simulations of brains "experience next to nothing," then f<sub>sim</sub> ≈ 0 and disjunct (3) dies. But: (a) IIT is a contested theory, not settled; (b) Bostrom's attenuated premise merely needs *in fact* consciousness to supervene on synaptic-level structural replication — a claim at the center of an open dispute; (c) Dainton's broadening shows that even non-conscious simulations threaten "virtual life" for *resident* minds. Verdict: the substrate-independence assumption is the argument's softest empirical underbelly, and IIT is the only *published, systematic* scientific framework that attacks it head-on — but "IIT implies X" is not the same as "X is established." [Inference]

8. **Ringel & Kovrizhin.** *Misses entirely* (see §4d). The careful readings (Aaronson, Bostrom) agree the paper says nothing about the simulation hypothesis; the media coverage is the textbook case of the attack being built on a headline, not the paper. Honest note: I could not find Ringel's own corrective statement; his co-option into "proof we're not simulated" happened against the paper's actual content. [Inference; the absence of Ringel's own statement is a Not-Found gap]

9. **Hossenfelder's "pseudoscience."** *Misses Bostrom, lands on the popular version.* Her two premises — (a) simulators must reproduce GR+SM from an algorithm, (b) the "fill in details lazily" scheme is unexplained — are both denied/answered by Bostrom's actual positions (only observational indistinguishability is required; cheating is allowed and cheap; FAQ §6). Her stronger residual point — that nobody has shown *any* underlying algorithm can generate our observations — is a genuine burden-of-proof objection, but it is an argument for *agnosticism*, not against the trilemma's validity. Also note the irony for the fleet research program: Hossenfelder's own "the laws are mathematical, so you could say the universe is really just computing those laws" line concedes the weak computationalism that many sim-hypothesis researchers build on. [Inference]

10. **Cutting the other way, in balance.** Chalmers (25%, virtual realism, defusing the "not real" sting) and Bostrom's FAQ (§4 two-conditionals; §6 cheat-cheap-cost; §8 early-generation indexical; §11 testability; §12 experimental caution) collectively answer or absorb most of the above. The one attack cluster to which no adequate published reply exists, per my search: **Birch's selective-scepticism charge** and **the IIT denial of simulated consciousness** — those two are where the anti-simulation case is strongest, and they attack different premises (epistemology vs. philosophy of mind), so the hypothesis's defenders must win *both* to keep the trilemma's force. [Inference]

**Bottom line for Argus's conviction [Inference, clearly mine]:** The philosophical case against the simulation *hypothesis* (the belief), as opposed to against the *argument's* strength, is much weaker than its rhetoric. What is actually defeated: (i) the naive "we're probably in a sim because most minds are sims" inference (Weatherson, Carroll-typicality), (ii) the idea that the hypothesis is easily *testable now* (Hossenfelder half-right; Ringel-Kovrizhin nothing), (iii) the 2003 derivation as originally written (Patch). What survives: the trilemma in repaired form, with substrate-independence as the single most vulnerable empirical premise. A steelmanned pro-simulation researcher should concede (i)–(iii) freely and put their weight on the two live debates — consciousness-substrate and observer-typicality — where the evidence is genuinely open.

---

## Primary sources and URLs

1. Bostrom, N. (2003). "Are You Living in a Computer Simulation?" Philosophical Quarterly 53(211): 243–255. https://simulation-argument.com/simulation.pdf
2. Weatherson, B. (2003). "Are You a Sim?" Philosophical Quarterly 53(212): 425–431. http://brian.weatherson.org/sims.pdf
3. Bostrom, N. (2005). "The Simulation Argument: Reply to Weatherson." PQ 55(218): 90–97. https://simulation-argument.com/weathersonreply.pdf
4. Birch, J. (2013). "On the 'Simulation Argument' and Selective Scepticism." Erkenntnis 78(1): 95–107. https://simulation-argument.com/pdf/birch-on-the-simulation-argument-and-selective-skepticism.pdf
5. Bostrom, N. & Kulczycki, M. (2011). "A Patch for the Simulation Argument." Analysis 71(1): 54–61. https://simulation-argument.com/patch.pdf
6. Bostrom, N. (2009). "The Simulation Argument: Some Explanations." Analysis. https://simulation-argument.com/brueckner.pdf
7. Brueckner, A. (2008). "The Simulation Argument Again." Analysis 68(3): 224–226. (read via #6; paywalled at https://academic.oup.com/analysis/article-abstract/68/3/224/1803849)
8. Dainton, B. (2002). "Innocence Lost: Simulation Scenarios: Prospects and Consequences." https://simulation-argument.com/dainton.pdf
9. Carroll, S. (2016). "Maybe We Do Not Live in a Simulation: The Resolution Conundrum." https://preposterousuniverse.com/blog/2016/08/22/maybe-we-do-not-live-in-a-simulation-the-resolution-conundrum/
10. Carroll, S. (2017). "Why Boltzmann Brains Are Bad." arXiv:1702.00850. https://arxiv.org/abs/1702.00850
11. Tononi, G. & Koch, C. (2015). "Consciousness: here, there and everywhere?" Phil. Trans. R. Soc. B 370: 20140167. https://royalsocietypublishing.org/doi/10.1098/rstb.2014.0167
12. Ringel, Z. & Kovrizhin, D. L. (2017). "Quantized gravitational responses, the sign problem, and quantum complexity." Science Advances 3(9): e1701758. https://www.science.org/doi/10.1126/sciadv.1701758
13. Aaronson, S. (2017). "Because you asked: the Simulation Hypothesis has not been falsified; remains unfalsifiable." https://www.scottaaronson.com/blog/?p=3482
14. Hossenfelder, S. (2021). "The Simulation Hypothesis is Pseudoscience." Backreaction. http://backreaction.blogspot.com/2021/02/the-simulation-hypothesis-is.html
15. Bostrom, N. Simulation Argument FAQ. https://simulation-argument.com/faq/
16. Doolan, P. (2022). Interview with David Chalmers, Philosophy Now 148. https://philosophynow.org/issues/148/David_Chalmers
17. Chalmers, D. (2009). "The Matrix as Metaphysics." (cited via #3)

## Not-found / gaps (where I looked)

- **Carroll applying "cognitively unstable" to the simulation hypothesis in print:** not found. Searched preposterousuniverse.com (blog + podcasts, incl. Mindscape #111 with Bostrom), arXiv, Google; the term is used in his Boltzmann Brains paper only. The May 2023 AMA question quoted it re: Boltzmann Brains.
- **Ringel's personal statement on the misreporting:** not found (searched news, Quora, ResearchGate profile, interviews; did not search Hebrew-language media or podcasts).
- **A published Searle-style critique aimed specifically at Bostrom:** not found; the transposition is informal literature.
- **A published Bostrom reply to Birch:** not found.
- **Tononi & Koch paper body:** publisher Cloudflare-blocked; abstract verified via PhilPapers/publisher page snippets. Full-body claims beyond the abstract unverified.
- **Brueckner 2008 full text:** paywalled; read via Bostrom's extensive quotations.
- **Carroll's *Something Deeply Hidden*:** checked via search only; no simulation-specific passage located.