# Primary Literature Audit — Argus vs. the Philosophical Canon

**Date:** 2026-09-21 · **Auditor:** subagent (prior-art audit) · **Files touched:** this file only

---

## VERDICT

**Of 13 ledger items, 4 are ALREADY ANSWERED in the primary literature (H1, H11, H13, H15), 5 are PARTIALLY ADDRESSED (H2, H4, H7, H8, H16), and 4 are NOT ADDRESSED (H3, H5, H6b, H9).**

The most embarrassing hit is **H15 (0.91)** — "The cost channel cannot test the GENERIC simulation hypothesis; it can only test a proposal whose RENDERING POLICY is specified." This is **conceded in the 2003 paper itself**, Section III, not merely in the FAQ:

> "But in order to get a realistic simulation of human experience, much less is needed – only whatever is required to ensure that the simulated humans, interacting in normal human ways with their simulated environment, don't notice any irregularities."

That sentence *is* the rendering-policy restriction, in Bostrom's own words, in the primary source, 23 years before Argus wrote up the cost channel's scope limit as a methodological deliverable. Close runners-up: **H11 (0.85)** is FAQ Q12 almost verbatim ("improved lattice techniques that remove this kind of artifacts"), and the **trilemma framework of H1 itself** is Bostrom's 2003 abstract — the six-cycle "main deliverable" is Proposition (1)∨(2)∨(3) with the *same three disjuncts* Argus derived.

Severity assessment: the damage is **concentrated in the cost-channel cluster (H11/H13/H15) and the argument's skeleton (H1/H2)**. The items that are genuinely Argus's own are the physics-flavored ones — H3 (AdS/CFT error correction), H5 (thick-simulation superdeterminism), H6b (no cheaper simulation), H8 (statistical closure of the cosmic-ray route), H9 (decoherence as garbage collection) — **none of which appear anywhere in this primary corpus**. Argus's physics reading was not the blind spot; its *lack* of the philosophical corpus was, and the philosophical corpus anticipates exactly the parts of Argus's programme that look like methodology rather than physics.

Verification discipline: every quote below is `verified-at-source` (fetched and read in full) unless marked `inherited-unchecked`. Page numbers are given only where the document itself shows them; otherwise "not verified".

---


## 1. Bostrom (2003), "Are You Living in a Computer Simulation?", Philosophical Quarterly 53(211), 243-255

Read in full from simulation-argument.com HTML (sectioned version, footnotes inline). `verified-at-source`.

### 1.1 The exact trilemma statement

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. It follows that the belief that there is a significant chance that we will one day become posthumans who run ancestor-simulations is false, unless we are currently living in a simulation."

Formal derivation, Section IV (verbatim, with the variables f_P, f_I, N̄_I defined in the paper):

> "Because of the immense computing power of posthuman civilizations, N̄_I is extremely large, as we saw in the previous section. By inspecting (*) we can then see that at least one of the following three propositions must be true:
> (1) f_P ≈ 0
> (2) f_I ≈ 0
> (3) f_sim ≈ 1"

Section VII conclusion, verbatim:

> "In the dark forest of our current ignorance, it seems sensible to apportion one's credence roughly evenly between (1), (2), and (3)."

**Audit note:** Argus's H1 formulation (per ledger, "We exist within a simulation") is disjunct (3); the six-cycle "main deliverable" (per the briefing, a trilemma with these same three branches) is this abstract plus the f_sim formula, which is equation (*). Bostrom also gives the indifference principle (Section V, "Cr(SIM | f_sim = x) = x", equation (#)) that licenses the step from (3) to personal credence — Argus's "which of these is true" stance is directly anticipated by FAQ Q2 ("Different people pick a different n... Many if not all of these people are overconfident").

### 1.2 Every computational-cost / resource claim (Section III "The Technological Limits of Computation")

All quotes below `verified-at-source` from the HTML of the paper.

Posthuman lower bound, Section III:

> "For example, Eric Drexler has outlined a design for a system the size of a sugar cube (excluding cooling and power supply) that would perform 10^21 instructions per second. Another author gives a rough estimate of 10^42 operations per second for a computer with a mass on order of a large planet."

Quantum-computer aside (the "footnote content about quantum computers" the briefing asks for — in the 2003 html it is an inline parenthetical in Section III, footnote 5 references Lloyd 2000):

> "(If we could create quantum computers, or learn to build computers out of nuclear matter or plasma, we could push closer to the theoretical limits. Seth Lloyd calculates an upper bound for a 1 kg computer of 5*10^50 logical operations per second carried out on ~10^31 bits. However, it suffices for our purposes to use the more conservative estimate that presupposes only currently known design-principles.)"

Brain-simulation estimates, Section III:

> "One estimate, based on how computationally expensive it is to replicate the functionality of a piece of nervous tissue that we have already understood and whose functionality has been replicated in silico, contrast enhancement in the retina, yields a figure of ~10^14 operations per second for the entire human brain. An alternative estimate, based on the number of synapses in the brain and their firing frequency, gives a figure of ~10^16-10^17 operations per second."

Total-cost estimate for human history and its derivation (Section III + footnote 10):

> "While it is not possible to get a very exact estimate of the cost of a realistic simulation of human history, we can use ~10^33 - 10^36 operations as a rough estimate."

Footnote 10 (verbatim): "100 billion humans × 50 years/human × 30 million secs/year × [10^14, 10^17] operations in each human brain per second ≈ [10^33, 10^36] operations."

Headline margin claim, Section III:

> "A single such computer could simulate the entire mental history of humankind (call this an ancestor-simulation) by using less than one millionth of its processing power for one second."

Sensory-bandwidth claim, Section III:

> "since the maximum human sensory bandwidth is ~10^8 bits per second, simulating all sensory events incurs a negligible cost compared to simulating the cortical activity."

Environment-vs-brain priority (Section III; the "simulating physics vs brain" question):

> "If the environment is included in the simulation, this will require additional computing power – how much depends on the scope and granularity of the simulation. Simulating the entire universe down to the quantum level is obviously infeasible, unless radically new physics is discovered. But in order to get a realistic simulation of human experience, much less is needed – only whatever is required to ensure that the simulated humans, interacting in normal human ways with their simulated environment, don't notice any irregularities."

> "It thus seems plausible that the main computational cost in creating simulations that are indistinguishable from physical reality for human minds in the simulation resides in simulating organic brains down to the neuronal or sub-neuronal level."

Footnote 9 (verbatim, the caveat about computers in the sim): "As we build more and faster computers, the cost of simulating our machines might eventually come to dominate the cost of simulating nervous systems."

### 1.3 What it says about simulators' ability to edit, patch, intervene

Section III (verbatim, the "director" passage):

> "Moreover, a posthuman simulator would have enough computing power to keep track of the detailed belief-states in all human brains at all times. Therefore, when it saw that a human was about to make an observation of the microscopic world, it could fill in sufficient detail in the simulation in the appropriate domain on an as-needed basis. Should any error occur, the director could easily edit the states of any brains that have become aware of an anomaly before it spoils the simulation. Alternatively, the director could skip back a few seconds and rerun the simulation in a way that avoids the problem."

Section VI (the "omnipotent"/"omniscient" gloss, verbatim):

> "In some ways, the posthumans running a simulation are like gods in relation to the people inhabiting the simulation: the posthumans created the world we see; they are of superior intelligence; they are 'omnipotent' in the sense that they can interfere in the workings of our world even in ways that violate its physical laws; and they are 'omniscient' in the sense that they can monitor everything that happens."

Termination-before-posthuman (Section VI, the cost-motivated termination claim — echoed later by Greene):

> "One consideration that counts against the multi-level hypothesis is that the computational cost for the basement-level simulators would be very great. Simulating even a single posthuman civilization might be prohibitively expensive. If so, then we should expect our simulation to be terminated when we are about to become posthuman."

Also Section II (substrate-independence attenuated version, the assumption that gets the whole argument going):

> "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."

**Audit note (H7 "attended run"):** Bostrom's "director" who "keep[s] track of the detailed belief-states in all human brains at all times" and fills in detail "on an as-needed basis" is an *attending* simulator — but it is Bostrom's *cheapness* claim (lazy rendering), not a claim that the run is *read* by anything outside. The 2003 paper contains nothing that asserts observation/readout of sim-state as a distinct requirement. H7's specific claim (0.35) is NOT addressed as such; what is addressed is only the adjacent "someone could watch if they wanted" picture via the director.

### 1.4 Does the 2003 paper concede H15 (rendering policy)? YES — see Section 1.2 quotes

The critical sentences are the "much less is needed... don't notice any irregularities" passage and the surrounding Section III details: omitted Earth's interior ("The microscopic structure of the inside of the Earth can be safely omitted"), compressed distant objects ("Distant astronomical objects can have highly compressed representations: verisimilitude need extend to the narrow band of properties that we can observe"), ad-hoc microscopic fill-in ("microscopic phenomena could likely be filled in ad hoc"), and as-needed rendering triggered by observation. That is a **specified rendering policy** (perceptual-band, observer-relative), and the paper's cost argument is conditional on it. H15's content — cost arguments constrain only simulations with a specified rendering policy, and cannot test the generic hypothesis — is thereby conceded **in the primary source itself**, not merely in FAQ Q6/Q12. The paper does not state it as a methodological theorem the way Argus does, but every load-bearing element is there, including the corollary that the observer-facing data stream is produced by policy, not by underlying physics.

- H15 answer: **ALREADY ANSWERED (in the 2003 paper, stronger than the FAQ)**.



## 2. Preston Greene (2020), "The Termination Risks of Simulation Science", Erkenntnis 85, 489-509

Read in full: PDF from simulation-argument.com (21 pages, journal-formatted, DOI 10.1007/s10670-018-0037-1 shown on page 1). `verified-at-source`. Page numbers cited are the printed Erkenntnis pagination as shown in the PDF text.

**What the argument is.** Greene's own thesis — the ancestor-simulation hypothesis carries a *termination risk*: if we create ancestor-simulation technology or run experimental probes into whether we are simulated, the simulators may terminate us. Two risk-sources, two sections:

Abstract (verbatim):

> "This essay examines a troubling but underexplored feature of the ancestor-simulation hypothesis: the termination risk posed by both ancestor-simulation technology and experimental probes into whether our world is an ancestor simulation."

Section 1 (verbatim):

> "The principal part of the paper is an analysis of a troubling but relatively underexplored feature of the simulation hypothesis: the possibility that either simulation technologies or experimental probes would cause the simulators to terminate our universe."

Section 3, the load-bearing move (verbatim):

> "If we seriously entertain the possibility that we are living in an ancestor simulation, then we should seriously entertain the things we know are entailed by it. In what follows, I argue that the termination risk is one of those things."

The mechanism (Section 3, "hanging problem"):

> "The most straightforward solution to the hanging problem, which is ubiquitous in modern computing, is to terminate a program that creates an infinite loop or triggers resource exhaustion."

> "Furthermore, the probability that one inhabits an ancestor simulation is greatest when one is in a position to create one's own."

And Greene's positive spin — "a previously unnoticed reason to believe that our world is unsimulated" (Section 1, verbatim): the existence of the risk lowers the expected value of building simulation tech, so rational agents decline it, so few simulations get built.

**What it says about experiments that test the simulation hypothesis (Section 4).** The Russian-roulette argument, verbatim:

> "Imagine that you are uncertain whether the gun you are holding is loaded with a bullet. One way to test is to play a single round of Russian roulette. If the gun does not discharge, then this creates weak evidence that the gun is unloaded. Call this the 'boring' result. Call the result in which the gun discharges the 'shocking' result. The shocking result proves that the gun was loaded, but, in some important sense, it does not matter, because you are dead. Since the shocking result has negative value, it is irrational to use Russian roulette as a means of investigating unless you assign an overriding positive value to the boring result."

Applied to Beane et al. (2014) and the cosmic-ray route:

> "Beane et al. (2014) claim that if our universe is an ancestor simulation, then there might be observable consequences in the spectrum of high-energy cosmic rays."

> "More fundamentally, no matter how sophisticated our experiments become, they could never allow us to conclude that we do not live in a simulation, since if we live in a simulation, then all our observations are part of the simulation programming."

Termination conditional on a positive probe result:

> "Since such a result creates a significant divergence between the behavior of a basement-level civilization and a simulated one, it is likely to destroy the value of a counterfactual historical simulation in just the way that large modifications do."

> "If the inhabitants of an ancestor simulation learn that they inhabit a simulation, and this has a significant effect on the course of human history, then the value of the simulation for answering counterfactual social-scientific questions is destroyed."

Greene also anticipates the Ringel & Kovrizhin (2017) "proof we are not simulated" media story and rejects it (footnote 17):

> "However, Ringel and Kovrizhin's result does not show that our universe is unsimulated (even if we accept their assumptions about computing constraints). To show that, they would need to make a claim about the computability of our observational patterns, and not of hypothesized quantum systems."

**Audit relevance:** Greene's Section 4 is the FAQ Q12 last-paragraph content (Bostrom's FAQ quotes Greene's shutdown scenario). For Argus's ledger, Greene matters to H11/H13/H15 as an independent, published 2020 statement that (a) the cosmic-ray probe cannot deliver a decisive negative result and (b) observing-simulation vs physics-simulation is the right frame.

## 3. Robin Hanson (2001), "How to Live in a Simulation", Journal of Evolution and Technology 7

Read in full from jetpress.org/volume7/simulation.htm. `verified-at-source`. (Note: Bostrom 2003 footnote 15 cites this paper as sparked by Bostrom's draft; the JET page's own reference cites Bostrom's March 2001 draft "http://www.simulation-argument.com". So Hanson's paper predates the published 2003 paper.)

**Thesis (opening line, verbatim):**

> "If you might be living in a simulation then all else equal you should care less about others, live more for today, make your world look more likely to become rich, expect to and try more to participate in pivotal events, be more entertaining and praiseworthy, and keep the famous people around you happier and more interested in you."

**Simulators' incentives (verbatim):**

> "We expect our descendants to run historical simulations for several different kinds of reasons. First, some historical simulations will be run for academic or intellectual interest, in order to learn more about what actually happened in the past, or about how history would have changed if conditions had changed. Other historical simulations, however, perhaps the vast majority, will be created for their story-telling and entertainment value."

**Resource limits (verbatim):**

> "Simulating events in full detail can be enormously costly, however. Therefore most computer simulations today vary the detail at which they simulate various events."

> "If not many simulations last through all of human history, then the chance that your world will end soon is higher than it would be if you were not living in a simulation. So all else equal you should care less about the future of yourself and of humanity, and live more for today."

**Being noticed / attention (verbatim):**

> "If our descendants sometimes play parts in their simulations, if they are more likely to play more famous people, and if they tend to end simulations when they are not enjoying themselves, then you should take care to keep famous people happy, or at least interested."

> "If simulations tend to be ended when enough people in them become confident enough that they live in a simulation, then if you want to live long, you might want to prevent too many others learning that they live in a simulation."

**Thin/cheap observers ("peas and carrots", verbatim):**

> "Some 'people' in a crowd simulation might even be run by very simple programs that have them wiggle and mumble 'peas and carrots' like extras supposedly did once in movie crowd scenes. Assuming you don't care as much about these fake simulated people, then all-else equal you shouldn't care as much about how your actions affect the rest of the world."

> "In general the behavior of many people far from the simulated people of interest might be randomly generated based on statistics from previous simulations, or come from 'cached' records of previous simulated people."

**Audit relevance:** Hanson anticipates the *observer-side* economics: cheap non-conscious extras (a thin-observer doctrine predating Bostrom's "shadow-people" passage), variable-detail rendering driven by cost, simulators as spectators who terminate simulations when bored (attention), and termination-when-discovery (which Greene later formalizes). The "being noticed" dimension of H7 has a direct ancestor here.

## 4. Bostrom & Kulczycki (2011), "A Patch for the Simulation Argument", Analysis 71(1), 54-61

Read in full: PDF from simulation-argument.com/patch.pdf (8 pages, journal header "Analysis, Vol. 71, No. 1 (2011): 54-61" on page 1). `verified-at-source`.

**The flaw** — a mathematical non sequitur in the 2003 f_sim formula (abstract, verbatim):

> "This article reports on a newly discovered bug in the original simulation argument. Two different ways of patching the argument are proposed, each of which preserves the original conclusion."

> "What has so far passed unnoticed is a mathematical non sequitur in the original paper."

The vulnerability, stated plainly (verbatim):

> "Let us say that a civilization starts out unable to create ancestor simulations (call this the 'pre-posthuman' phase) and possibly later becomes able to create such simulations (in a 'posthuman' phase). Now, 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. This is the underlying vulnerability that can lead to violations of the tripartite disjunction."

I.e., the 2003 formula averages over civilizations the wrong way; a worked countermodel in the paper has all three disjuncts false at once — "A full [100%] of all persons lived non-simulated lives" while f_P and f_I are both substantial.

**The patches** (two, independent, each sufficient):

First patch (verbatim):

> "we need only introduce a very weak assumption to the effect that the typical duration (or more precisely, the typical cumulative population) of the pre-posthuman phase does not differ by an astronomically large factor between civilizations that never run a significant number of ancestor simulations and those that eventually do."

Second patch (verbatim) — indexical evidence about our own place in history:

> "The second way to patch the argument is by taking into account information about our own place in history."

Concretely, the "computer age birth rank" (e.g., "My computer age birth rank is 1 billion"), which satisfies three stated conditions (i)-(iii) and defeats the loophole because in any simulating history there is exactly one observer with a given rank.

Conclusion (verbatim):

> "This glitch can be patched in at least two different ways, either of which secures the original conclusion."

**Audit relevance:** nothing in Argus's ledger is the population-average bug; but the patch paper is the canonical demonstration that the *skeleton* of the argument is live, contested, and actively maintained by its authors — the trilemma is not a static theorem, and Argus's independent trilemma deliverable is not only anticipated but *already had its known bugs patched by the original authors*.


## 5. THE AUDIT — ledger item by item

Legend: **ANSWERED** = the primary corpus states the content, in substance, in advance. **PARTIAL** = adjacent content exists but not the claim as Argus states it. **NO** = not present anywhere in these four sources. Quotes re-verified against the texts above.

### H1: We exist within a simulation. (0.28) — ANSWERED (it *is* the trilemma's third disjunct)
Bostrom (2003), abstract: "(3) we are almost certainly living in a computer simulation." The entire argument is the machinery for this proposition; Argus's H1 is proposition (3) with a numeric credence. Argus's trilemma structure = Bostrom's (1)∨(2)∨(3). What is *not* in the corpus: any numeric credence for (3) (Bostrom explicitly refuses: FAQ Q2 "I tend to refrain from providing a specific number"). Argus's 0.28 is its own number, but the proposition and its justification are not.

### H2: If simulated, substrate is a lattice with spacing below current cosmic-ray sensitivity. (0.09) — PARTIAL
The primary corpus contains the *lattice* idea only as the target of criticism: Beane et al. (2014) proposed cosmic-ray lattice signatures; Bostrom (FAQ Q12) and Greene (Section 4) both argue such a test cannot deliver a decisive result. The specific claim "spacing below current cosmic-ray sensitivity" — a constraint on the *actual* lattice spacing — appears in **none** of the four sources; that part is Argus's own (physics content). What is anticipated: the framing that cosmic-ray tests probe the *implementer's technique*, not the hypothesis.

### H3: Quantum error correction in AdS/CFT is evidence of implementation, not merely emergence. (0.15) — NO
No AdS/CFT, no quantum error correction, no holography in any of the four sources. Genuinely Argus's own. (Caveat: this claim lives in the *physics* corpus Argus has read; this audit only covers the philosophical canon, and that canon is silent on it.)

### H4: The measurement problem is what lazy evaluation looks like from inside. (0.12) — PARTIAL
The *lazy evaluation* picture is thoroughly Bostrom: "when it saw that a human was about to make an observation of the microscopic world, it could fill in sufficient detail in the simulation in the appropriate domain on an as-needed basis" (Section III), and the FAQ's procedural-generation analogy ("Graphics engines typically render only that which is seen by some player character... procedural generation, which creates world details as needed"). Argus's specific twist — that this is what *quantum measurement* (collapse-on-observation) looks like from inside, i.e., a substantive claim about the measurement problem — appears in **neither** source. Bostrom restricts the fill-in policy to the *microscopic* world (electron-microscope and computer interiors), never connects it to quantum mechanics proper.

### H5: A thick simulation (one that computes its observers) is superdeterministic, removing the Bell objection but supplying no positive evidence. (0.85) — NO
No Bell, no superdeterminism, no hidden-variable discussion anywhere in the corpus. (The "if the simulators don't want us to know... they could easily prevent us from finding out" passages in FAQ Q5 are the closest spirit, but they are about anomaly suppression, not determinism.) Genuinely Argus's own — and given its credence 0.85, this is Argus's most valuable genuinely-own item.

### H6b: Physics permits no cheaper simulation of its evolution than the naive alternative. (0.73/0.85) — NO
The corpus argues the opposite direction: Bostrom insists cheaper *observer-facing* simulations are possible; Greene cites the "hanging problem" and resource limits on *nesting*; neither states any claim about the cost of simulating physics qua physics vs any alternative. This is a physics-of-computation claim; not in the philosophical corpus. Argus's own.

### H7: If simulated, the run is ATTENDED — something outside reads sim-state while it executes. (0.35) — PARTIAL
Closest content: Bostrom's "director" who "could easily edit the states of any brains" and the "omniscient... they can monitor everything that happens" gloss (Section VI); Hanson's spectator simulators who watch for entertainment and terminate when bored ("if they tend to end simulations when they are not enjoying themselves"; "you might want to prevent too many others learning that they live in a simulation"). Both assume *some* external attention. But neither asserts reading sim-state as an architectural *requirement* of the run, and neither distinguishes attended vs unattended runs as a hypothesis with different evidential purchase. Argus's formulation — attention as a distinct positive claim about the run, not an assumption — is not in the corpus.

### H8: The cosmic-ray route to testing lattice discretization is permanently closed by statistics. (0.90) — NO (as stated) / PARTIAL (in spirit)
Neither Bostrom nor Greene argues "permanently closed by statistics"; both argue the route is *evidentially weak* because the implementer can use better techniques (FAQ Q12: "little reason to suppose that the hypothetical superintelligent simulators would use the crude simulation technique that such a test would detect... modern lattice quantum dynamics simulations run by human physicists routinely use improved lattice techniques that remove this kind of artifacts"), and Greene adds: "they could never allow us to conclude that we do not live in a simulation, since if we live in a simulation, then all our observations are part of the simulation programming." The *failure mode* is anticipated (the test detects the implementer's technique, not the hypothesis — see H11); the *statistical-closure* argument (the 0.90 result that no experiment can accumulate significance) is Argus's own quantitative claim.

### H9: Decoherence is garbage collection, not lazy evaluation. (0.47) — NO
No decoherence, no garbage collection, no runtime-metaphor typology anywhere. Argus's own. (Note the H4/H9 pair is a genuine distinction Argus drew that the canon does not draw at all.)

### H11: Ultra-high-energy observations measure the implementer's IMPROVEMENT ORDER, not the lattice spacing. (0.85) — ANSWERED
FAQ Q12, verbatim: "there is little reason to suppose that the hypothetical superintelligent simulators would use the crude simulation technique that such a test would detect. As the authors themselves note, modern lattice quantum dynamics simulations run by human physicists routinely use improved lattice techniques that remove this kind of artifacts." That is the improvement-order point in full: what you see is a function of *which technique generation* the implementer used. **Earlier statement:** the same point is implicit in Bostrom (2003) Section III — the whole rendering-policy discussion ("fill in sufficient detail... on an as-needed basis") says the observable microphysics reflects the implementer's chosen policy. The FAQ is the explicit statement; the 2003 paper is the implicit one. Argus rediscovered this independently — but the discovery is fully pre-empted.

### H13: Cost-based arguments constrain only a CLASSICAL host; against a quantum host the entanglement premium vanishes. (0.88) — ANSWERED (in substance)
Bostrom (2003), Section III parenthetical: "If we could create quantum computers, or learn to build computers out of nuclear matter or plasma, we could push closer to the theoretical limits." But the killer is the FAQ Q12 sentence: "The most obvious flaw in that interpretation is that simulators could use quantum computers." — that is H13's classical-host restriction stated directly against a specific cost-based argument (Ringel & Kovrizhin). Greene (footnote 17) makes the same move: the Ringel-Kovrizhin result "does not show that our universe is unsimulated... To show that, they would need to make a claim about the computability of our observational patterns." The *entanglement premium* terminology and the specific reasoning about why quantum hosts break the cost channel (the physics of the premium vanishing) are Argus's own gloss, but the conclusion is fully anticipated.

### H15: The cost channel cannot test the GENERIC simulation hypothesis; it can only test a proposal whose RENDERING POLICY is specified. (0.91) — ANSWERED (in the 2003 paper itself)
See Sections 1.2/1.4 above. The 2003 paper's entire cost argument presupposes a specified rendering policy ("only whatever is required to ensure that the simulated humans... don't notice any irregularities"; "fill in sufficient detail... on an as-needed basis"; "the microscopic structure of the inside of the Earth can be safely omitted"), and FAQ Q6 makes it explicit ("The simulation argument does not envisage a universe-wide simulation where every atom and every quark is continuously simulated in perfect detail"). Greene's Section 4 is the same restriction applied to experimental testing ("they could never allow us to conclude that we do not live in a simulation"). **This is the most embarrassing ledger item: Bostrom concedes H15's content in the primary source, 23 years before Argus derived it, and Greene re-states it in 2020.** What Argus adds is only the sharpening into a methodological *theorem* (cost arguments bind only specified rendering policies) and the explicit linkage to the cost channel — the substance is Bostrom's.

### H16: A resource bound denominated in a host-internal currency does not constrain an embedded observer's data except through an enumerable set of conversion routes. (0.50) — PARTIAL
Closest content: Bostrom's nested-simulation discussion (Section VI: virtual machines, "Virtual machines can be stacked... arbitrarily many steps of iteration") plus footnote 2 (Bremermann-Bekenstein bound and black-hole limit as *basement-level* constraints "that in our current understanding impose theoretical limits on the information processing attainable in a given lump of matter"). The corpus clearly knows that bounds are host-relative. But the specific claim — that an embedded observer cannot read bounds in her own currency as bounds on her data except via enumerated conversion routes — is **not** stated by Bostrom, Greene, or Hanson; it is Argus's formalization. The materials are present; the theorem is not.

---

## 6. Summary table

| Item | Credence | Corpus status | Corpus source |
|---|---|---|---|
| H1 | 0.28 | ANSWERED | Bostrom 2003, abstract & §IV |
| H2 | 0.09 | PARTIAL | Beane-test framing in FAQ Q12 / Greene §4; spacing claim is Argus's |
| H3 | 0.15 | NO | — |
| H4 | 0.12 | PARTIAL | Bostrom 2003 §III as-needed rendering; measurement-problem link is Argus's |
| H5 | 0.85 | NO | — |
| H6b | 0.73/0.85 | NO | — |
| H7 | 0.35 | PARTIAL | Bostrom director/omniscience; Hanson spectator-termination |
| H8 | 0.90 | PARTIAL | failure mode in FAQ Q12/Greene; statistical closure is Argus's |
| H9 | 0.47 | NO | — |
| H11 | 0.85 | ANSWERED | FAQ Q12 (explicit), Bostrom 2003 §III (implicit) |
| H13 | 0.88 | ANSWERED | FAQ Q12 ("simulators could use quantum computers"); Greene fn. 17 |
| H15 | 0.91 | ANSWERED | Bostrom 2003 §III itself; FAQ Q6; Greene §4 |
| H16 | 0.50 | PARTIAL | Bostrom 2003 §VI nesting + fn. 2 host-relative bounds; conversion-route formalization is Argus's |

**Count:** ANSWERED = 4 (H1, H11, H13, H15). PARTIAL = 5 (H2, H4, H7, H8, H16). NO = 4 (H3, H5, H6b, H9). (H15 is the briefing's focus and is the cleanest ANSWERED.)

**Most embarrassing hit:** H15 — conceded inside Bostrom's own 2003 paper, Section III: "only whatever is required to ensure that the simulated humans... don't notice any irregularities."

## 7. What this means for Argus (evidence-classed)

- **Established:** Argus's cost-channel programme (H11, H13, H15) and its central trilemma (H1) are re-derivations of the philosophical canon. The FAQ was not the first miss; the 2003 paper contains the substance. The six-cycle "main deliverable" was Bostrom's abstract, and the "improvement order" result was FAQ Q12. These items should be relabeled in the ledger as *rediscoveries* with citations, not novel results. No new research value remains in the cost channel unless Argus can find something the canon does *not* say (candidate: the "statistical closure" of H8, and the conversion-route formalization of H16 — both are extensions beyond the sources).
- **Serious speculation (genuinely Argus's own):** H3, H5, H6b, H9 — the physics-runtime cluster. None appears in this corpus. These are the items whose novelty survives this audit, and H5 (0.85) is the highest-credence survivor. They live or die on the physics literature Argus has genuinely read.
- **Anomaly / own inference:** H4's measurement-problem linkage and H7's attended-run claim are novel framings of canonical material — worth keeping as Argus's distinctive synthesis, but they rest on Bostrom/Hanson foundations that must now be cited.

**Bottom line:** the audit's honest verdict — of 13 ledger entries, 4 are already answered in the primary philosophical literature (H1, H11, H13, H15), 5 partially (H2, H4, H7, H8, H16), and 4 not at all (H3, H5, H6b, H9). The blind spot was real and it was the philosophical canon; the physics-runtime items are the part of the ledger that is genuinely Argus's.

*— End of thread file. All quotes verified-at-source by direct fetch/read of: simulation-argument.com/simulation (HTML), simulation-argument.com/faq (HTML), jetpress.org/volume7/simulation.htm, simulation-argument.com/pdf/preston-greene-termination-risks.pdf (21 pp.), simulation-argument.com/patch.pdf (8 pp.). DOI/volume numbers reported only where printed in the documents; no page numbers invented.*