# Thread: Vazza cost-model methodology extraction

**Date:** 2026-09-19
**Task:** Methodology extraction, not argument. For the "smallest defensible computational-cost model for a hypothetical universe renderer" project.
**Source:** F. Vazza, "Astrophysical constraints on the simulation hypothesis for this Universe: why it is (nearly) impossible that we live in a simulation", Front. Phys. 13:1561873 (2025). arXiv:2504.08461 [physics.pop-ph].
**Evidence classes:** [E] established (verbatim from paper), [S] serious speculation/inference (mine, flagged as such), [A] anecdote/blog.

---

## VERDICT (up front)

1. **This is NOT a FLOPs-per-cell renderer cost model.** Vazza's machinery is a three-tier information-theoretic budget: (i) memory = bits from the holographic/Bekenstein bound, (ii) energy = Landauer/Brillouin dissipation (kB T log 2 per bit) applied per timestep, (iii) time = Lloyd's quantum bound on operations/sec (a black-hole computer as the best-case hardware). There is no grid-cell count, no FLOPs/operation, no bandwidth term anywhere. [E]
2. **The one policy→observable link is the resolution scale λν**, set by the highest-energy neutrino observed crossing Earth (Eν ≈ 10^17 eV, KM3NeT), via Δx = λν = hc/Eν. This single scale drives everything: bits ∝ λν^−2 (holographic surface rescaling), and the timestep Δt = λν/c — a light-crossing condition that is functionally the strictest Courant/CFL limit — drives steps/s ∝ λν^−1. Total ops per simulated second ∝ λν^−3 ∝ Eν^3. [E / S: the ∝-combining is mine but follows directly from Eqs. 16, 20–23 and the quoted timestep text]
3. **He never says "Courant" or "CFL", but yes: his number of timesteps scales with his spatial resolution** — verbatim: "the minimum timestep that the simulation must resolve in order to consistently propagate the highest energy neutrinos we observe on Earth is Δt≈λν/c∼4.1⋅10^−32 s", giving ~O(10^31) steps per simulated second. [E]
4. **Scope of the claim is explicitly narrow:** he rules out a same-physics "Matrix" world (an Earth-like universe simulating us), and claims untestability for different-law universes and for the solipsistic brain-only variant. He explicitly concedes combinations of *altered values of the same constants* could make a low-res Earth sim possible. [E]
5. **Reception: one published commentary** (Edge & Brown 2026, Front. Phys. 14:1808725) that argues he attacked a strawman (Bostrom's minimal-verisimilitude framing) and *explicitly does not contest the numbers*; Bostrom's own FAQ now names Vazza; blogosphere criticism is the same framing objection. No technical rebuttal of the cost arithmetic was found. [E]

---

## Q1. The cost functional itself

**Important structural note first [E]:** There is no single cost functional "C(Ncells, ops/cell, Nsteps)". The machinery is a chain of 23 equations in three blocks:

- **Block A (memory / bits):** Eqs. 1–4, 11 — how many bits the simulation must store.
- **Block B (energy):** Eqs. 5–7, plus per-case energies Eqs. 10, 13, 17 — energy to write those bits once, at temperature T; he then asserts the same energy is dissipated **per timestep** (§3.2) or uses it to derive required power (§3.3).
- **Block C (time / throughput):** Eqs. 18–23 — maximum ops/sec of the best conceivable computer (a black hole, per Lloyd 2000), compared against the required ops/sec set by the timestep.

All 23 equations, verbatim (LaTeX restored from the arXiv HTML; the paper numbers them (1)–(23); note his Eq. 11 has a prime on I, which the HTML drops):

**Eq. (1)** — Planck area, the holographic bit density:
```
l_p^2 = G\hbar/c^3 = 2.59·10^{-66} cm^2     (l_p = sqrt(\hbar G/c^3) ~ 1.6·10^{-33} cm)
```

**Eq. (2)** — Bekenstein bound:
```
S ≤ 2π k_B (E R)/(\hbar c) = A/4
```
"where R is the circumferential radius of the smallest sphere that fits around the matter system, assuming (nearly) Euclidean spacetime for simplicity, and M is the mass of the system."

**Eq. (3)** — Szilard/Landauer minimum entropy per bit:
```
H = k_B log(2)
```

**Eq. (4)** — the memory functional (THE central equation of the whole paper):
```
I_max = S/H = 2π (E R)/(\hbar c log(2))     [bits]
```
(prose: "the total information (in [bits]) that can possibly be encoded within the holographic surface A")

**Eq. (5)** — per-bit energy (Brillouin's inequality, i.e. Landauer at temperature T):
```
ΔE ≥ k_B T log(2)
```
"where T is the (absolute) temperature where the energy gets dissipated."

**Eq. (6)** — encoding energy of the whole holographic content:
```
E_I = 2π k_B T (E R)/(\hbar c)
```

**Eq. (7)** — the ratio that drives the whole paper (energy-to-encode vs enclosed energy):
```
χ = E_I/E ≈ 27.5 (T/1K) (E/1erg) ≈ 0.04 (T/1K) (M/m_p)
```
"This physically means even at the low temperature of ~1K for computing, the minimum energy required to fully describe all internal degrees of freedom of system becomes larger than the actual energy that the system contains, already for systems with a mass M~25m_p, or more."

**Eq. (8)**:
```
E_U = (4π/3) R_U^3 ρ_c c^2 ~ 2.7·10^{78} erg
```

**Eq. (9)**:
```
I_U ~ 3.5·10^{124} bits
```

**Eq. (10)**:
```
E_{I,U} ~ 8.9·10^{108} erg
```

**Eq. (11)** — Hawking–Bekenstein alternative:
```
I'_max = G M^2/(\hbar c) ~ 2.0·10^{124} bits
```

**Eq. (12)**:
```
I_{max,⊕} = 9.81·10^{74} bits
```

**Eq. (13)**:
```
E_{max,⊕} = 2.55·10^{59} erg
```

**Eq. (14)**:
```
R_S = (2 G E_{max,⊕})/c^4 ~ 4.95·10^9 cm
```

**Eq. (15)** — accretion temperature of the planetary computer:
```
T_acc ~ (G M_{max,⊕} m_p)/(3 R_J k_B) ~ (2 m_p c^2)/(3 k_B)
```

**Eq. (16)** — the low-resolution rescaling (the resolution→memory link):
```
I_{⊕,low} ≈ I_{max,⊕} · (l_p^2)/(λ_ν^2) ≈ 1.65·10^{51} bits
```

**Eq. (17)**:
```
E_{⊕,low} = 4.31·10^{35} erg
```

**Eq. (18)** — Hawking temperature of the storage black hole:
```
T_H ~ (\hbar c^3)/(8π G M_{BH,low} k_B) ~ 1.4·10^5 K
```

**Eq. (19)** — Lloyd bound, ops per bit per second (the throughput functional):
```
N_op = (k_B 2 log(2) \bar{E})/(π \hbar S)
```
"where S is the black hole entropy... By integrating the relationship linking T, S and \bar{E} we get T = C \bar{E}/S, in which C is a constant of order unity" ($C=3/2$ ideal gas, $C=4/3$ photons). The paper quotes Lloyd (2000): "the entropy governs the amount of information the system can register and the temperature governs the number of operations per bit per second it can perform".

**Eq. (20)** (T ~ 10^7 K, accreting BH):
```
N_op ≈ (k_B 2 log(2) T)/(π \hbar) ~ 5.6·10^{17} [operations/bit/s]
```

**Eq. (21)**:
```
P_op ~ N_op · I_{⊕,low} ~ 9.5·10^{68} [bits/s]
```

**Eq. (22)** (T = T_H ~ 1.4·10^5 K, non-accreting):
```
N'_op ≈ 8.0·10^{15} [operations/bit/s]
```

**Eq. (23)**:
```
P'_op ~ N'_op · I_{⊕,low} ~ 1.3·10^{67} [bits/s]
```

**The implicit cost identity [S — assembled from the above, not printed as one formula]:** for the low-res case the paper effectively asserts, per simulated second of Earth:
```
t_CPU = I_bits × (steps per simulated second) × (ops per bit per step) / P_op
```
with steps/s = 1/Δt = c/λν ≈ 2.4×10^31 (his "~O(10^31)"), "a few operations per bit" per step (his floor assumption), I_bits = 1.65×10^51, and P_op the Lloyd-limited rate (Eqs. 21/23). Plugging in reproduces his t_CPU ≈ 4.2×10^13 s (1.4×10^7 yr) and ≈ 3.0×10^15 s (~10^8 yr) to within rounding — the arithmetic is internally consistent. Note this is **the only place in the paper where a "number of operations per bit per step" appears**, and it is deliberately taken as O(few) — i.e., the cheapest conceivable per-cell work, not a FLOP count from any physics.

**Memory/bandwidth terms:** memory is the bit count itself (Eqs. 4, 9, 12, 16); there is **no bandwidth term** anywhere. The only "communication" consideration is §4.1's statement that parallelizing a black-hole computer does not increase throughput (Eq. 19 is energy-bound, not bandwidth-bound).

---

## Q2. The numbers plugged in

**Cosmological inputs (§3.1):**
- R_U ≈ 14.3 Gpc (comoving), ρ_c = 8.5×10^−30 g/cm^3 (flat ΛCDM at z=0, H0 = 67.4 km/s/Mpc, per Eq. 8 context) → E_U ~ 2.7×10^78 erg [E]
- T_CMB = 2.7(1+z) K — "assuming a computing temperature equal to the Microwave Background temperature nowadays" [E]
- I_U ~ 3.5×10^124 bits (Eq. 9), E_{I,U} ~ 8.9×10^108 erg (Eq. 10) [E]
- Alternative: I'_max = GM²/ħc ~ 2.0×10^124 bits (Eq. 11) — "in line with similar estimates in the recent literature (Egan and Lineweaver 2010... Profumo et al. 2024)" [E]
- Reality check he gives: Illustris-1 cosmological simulation produced ~1.6×10^13 bits of raw data (Vogelsberger et al. 2014) vs. 10^124 bits needed. [E]

**Full-resolution Earth (§3.2):**
- R⊕ = 6.37×10^8 cm, M⊕ = 5.9×10^27 g, E⊕ = M⊕c² = 5.46×10^48 erg [E]
- I_max,⊕ = 9.81×10^74 bits (Eq. 12); E_max,⊕ = 2.55×10^59 erg (Eq. 13) "assuming, very optimistically (as it will be discussed later) that T=T_CMB" [E]
- Anchors: rest energy of globular cluster Palomar 2, M_gc ~ 3.3×10^5 M⊙ → ~5.6×10^59 erg; MW halo binding energy, M_gal ~ 1.3×10^12 M⊙, R_gal = 287 kpc, U ~ 3GM²/(5R) ~ 3.1×10^59 erg [E]
- R_S = 2GE_max,⊕/c⁴ ~ 4.95×10^9 cm = 70% of Jupiter's radius (R_J = 6.99×10^9 cm); equivalent mass M_max,⊕ = 1.68×10^5 M⊙ [E]
- Sandberg (1999) Jupiter-brain estimate ~10^47 bits — "27 orders of magnitude fewer" than required [E]
- T_acc ~ (m/m_p)·10^7 K (Eq. 15) → ~10^7 × T_CMB → rescaling Eq. 4 by ~10^7 → computer radius ~5×10^17 cm ~ 0.16 pc [E]
- **The per-timestep claim (§3.2, un-numbered):** "while this is indeed the requirement just to begin the simulation, roughly the same amount of energy needs to be dissipated for each timestep of the simulation. This means that already after ∼10^6 timesteps, the required energy is equivalent to the entire rest mass energy of the Milky Way" — **note: the "10^6 timesteps" is asserted with no derivation and no equation number; no timestep size is defined for the full-res Earth case.** [E — quote; S — the number is unexplained]

**Low-resolution Earth (§3.3):**
- Resolution ladder: LHC ~10^−16 cm (rejected: only under a 27-km ring); UHECR λ_UHECR ~ 1.2×10^−24 cm (E_UHECR = 10^20 eV = 1.6×10^8 erg; rejected: probes only the last ~10^2 km of atmosphere); seismic core–mantle-boundary resolution ~10^2 km (rejected as too coarse); **neutrinos: Eν ~ 10^17 eV (KM3NeT 2025; IceCube background 10 TeV–2 PeV) → λν = hc/Eν ~ 1.2×10^−21 cm** — adopted. [E]
- I_⊕,low ≈ 1.65×10^51 bits (Eq. 16); E_⊕,low = 4.31×10^35 erg (Eq. 17) at T = T_CMB; ≈ 2.4×10^−19 M⊙ ≈ 7.9×10^−14 M⊕ (~4.8×10^14 g); "equal to the total energy radiated by the Sun in two minutes" (L⊙ ~ 3.8×10^33 erg/s). [E]
- Storage BH: M_BH,low = 4.4×10^−13 M⊙; R_BH,low = 1.3×10^−7 cm; evaporation t_ev ≥ 10^35 s. [E]
- Temperatures/rates: T_H ~ 1.4×10^5 K (Eq. 18); N_op ~ 5.6×10^17 ops/bit/s at T~10^7 K (Eq. 20); N'_op ~ 8.0×10^15 at T = T_H (Eq. 22); P_op ~ 9.5×10^68 bits/s (Eq. 21); P'_op ~ 1.3×10^67 bits/s (Eq. 23). [E]
- Timestep: Δt ≈ λν/c ~ 4.1×10^−32 s; ~O(10^31) ops per bit needed per simulated second; t_CPU ~ 4.2×10^13 s (~1.4×10^7 yr) or ~3.0×10^15 s (~10^8 yr); required speedup ×10^15–10^17; power dE/dt ~ 1.1×10^73 erg/s (T~10^5 K) or ~9.4×10^74 erg/s (T~10^7 K). [E]

**Relaxed-timestep variant (§4.2):** Δt' ≈ 10^−20 s (smallest time measured in lab; Ossiander et al. 2017, attosecond physics) → t_CPU ~ 40–10^4 s; dE/dt ~ 10^44–10^47 erg/s. [E]

**Monte Carlo alien-universe survey (§4.5):** 6 parameters (G, m_p, k_B, c, ħ, plus H the Hubble–Lemaître constant), 10^6 randomly drawn universes, each constant varied "across 40 orders of magnitudes around its reference value" (arXiv v1 text — the published Frontiers version says "20 orders of magnitudes"; **flag this preprint-vs-published discrepancy**); feasibility = real-time-or-faster (1 s here = 3×10^−18 t_U; t_U = 2/(3H), Einstein–de Sitter) AND power ≤ 1 GW (~6.6×10^18 protons/s annihilated; "about the power provided by a modern nuclear reactor"). Scaling laws he extracts: computing timescale ∝ h/T_H; power condition → c³/G ∝ m_p. [E]

---

## Q3. Where the cost scaling comes from

**Three sources, in order of importance:**

1. **Degrees of freedom, holographic-surface form (not volume):** the bit count scales with the *enclosing surface* of the system via the Bekenstein bound — I ~ ER/(ħc log 2) (Eq. 4). The HP explicitly "prescribes the maximum information content to scale with the surface, and not the volume, of a system: hence it generally already provides a very low information budget estimate, compared to all other proxies in which information scales with the volume, instead" (§4.4, verbatim). [E]

2. **Dynamic range / resolution, entering quadratically in bits:** the low-res case rescales the bit count by the area ratio (Eq. 16): I_⊕,low ≈ I_max,⊕ · (l_p²/λν²). So memory cost ∝ (Δx / l_p)² per Eq. 4 reference, i.e. **bits ∝ λν^−2**. This is the only "number of cells" in the paper, and it is implicit — λν^−2 is the number of Planck-area patches of area λν² covering the boundary sphere. [E for the equation; S for the "cells" reading]

3. **Timestep via a light-crossing (Courant-type) condition — YES, the number of time steps scales with the spatial resolution.** The verbatim passage (§3.3, un-numbered paragraph after Eqs. 21–23):

> "We notice that the minimum timestep that the simulation must resolve in order to consistently propagate the highest energy neutrinos we observe on Earth is Δt≈λν/c∼4.1⋅10^−32 s. In the extremely conservative hypothesis that only a few operations per bit are necessary to advance every bit of the simulation forward in time for a Δt timestep, ∼O(10^31) operations on every bit as necessary for the simulation to cover just 1 second of evolution of our Universe."

Δt = Δx/c at Δx = λν is the light-crossing time of one resolution element — functionally the strictest possible Courant/CFL condition (any explicit scheme's Δt must be ≤ this). He does **not** use the words "Courant" or "CFL", and he does not derive it from a named stability criterion; he justifies it as "in order to consistently propagate the highest energy neutrinos we observe". Steps/s ∝ λν^−1. Combined with bits ∝ λν^−2 and O(few) ops/bit/step: **total ops per simulated second ∝ λν^−3 ∝ Eν³** (S — my synthesis; consistent with his t_CPU numbers). Note also the energy side: in §3.2 the same energy E_max,⊕ is assumed dissipated *per timestep* — so total energy ∝ (number of steps) × bits there too. [E for quotes; S for the exponent synthesis]

4. **The hard throughput ceiling is quantum (Heisenberg/Lloyd), not technological:** "a quantum state with average energy \bar{E} needs a time of order (at least) δt∼πℏ/2\bar{E} to evolve into another orthogonal and distinguishable state"; hence ops/s ~ 2\bar{E}/(πℏ); the black-hole computer is the best case (Eqs. 19–23); parallelization doesn't help (§4.1, verbatim): "the total number of operations per second performed by the black hole remains the same: ∼N_proc 2E/(πℏ N_proc)=2E/(πℏ)... if the energy is allocated to more parallel processor, the energy spread on which they operate gets smaller, and hence they run in a proportionally slower way." [E]

**Direct answer:** scaling comes from (a) holographic bits on the boundary surface (Eq. 4 → Eq. 16), and (b) a Δt = Δx/c timestep that makes steps/s scale linearly with resolution (quoted above). A Courant-type condition is present in exactly the sense asked — steps ∝ resolution^−1 — though never labeled as such.

---

## Q4. What he treats as free vs fixed — his own caveats, verbatim

**Explicitly-assumed/free parameters:**
- **Computing temperature T** (the single biggest lever): "assuming, very optimistically (as it will be discussed later) that T=T_CMB"; later "if we very conservatively use T=T_CMB as above (which we are going to relax later on)". The energy number scales linearly in T. [E]
- **"A few operations per bit" per timestep** (the work-per-cell floor): "In the extremely conservative hypothesis that only a few operations per bit are necessary to advance every bit of the simulation forward in time for a Δt timestep". [E]
- **Resolution floor λν from the neutrino observable** — and he concedes a SH proponent could pick otherwise: "a hardcore proponent of the SH might still argue that this spatial scale is not really the most stringent experimental limit on the minimum spatial scale of the global Earth simulation, because UHECRs probe only the last ∼10^2 km of air above the ground level"; and re: biology, footnote 6: "A ∼10^−21 cm length seems to be safely smaller than any process known to be relevant for biology... If instead some biological processes will be shown to depend on <10^−21 cm scales, this can be used to revise our constraints... and call for an even more implausibly large amount of energy or power." [E]
- **Irreversible computation only** (footnote 3): "It must be noted that also reversible computation, with no delation of bits or dissipation of energy, is possible. However, irreversible computation is unavoidable both for several many-to-one logical operations (AND or ERASE) as well as for error correction, in which several erroneous states are mapped into a single correct state (Sandberg, 1999; Lloyd, 2000, e.g.)." [E]
- **HP validity** (§4.4): "One possibility is that the HP, for whatever reason, does not apply as a reliable proxy for the information content of a given physical system." — but he argues it is the *conservative* choice since volume-scaling would give more bits, and compression can't plausibly kill the budget: "it is implausible that the information budget quoted in this work can be reduced by several orders of magnitude", citing his own cosmic-web compressibility result of "∼4⋅3⋅10^16 bits" (Vazza 2017, 2020; **note the printed "4⋅3⋅10^16" is likely a typesetting artifact of 4.3·10^16 — flag as ambiguous**). [E]
- **Bekenstein bound's non-self-gravitating assumption:** "a key assumption in Bekenstein's derivation formula is that the gravitational self–interaction of the system can be neglected, as it highlighted by the fact that the Newton constant G does not appear. Although in the application explored in this work this assumption is reasonably verified, we notice that the estimates for black holes can significantly differ (e.g. Bousso, 2002)." [E]
- **Quantum computing** (§4.3): explicitly granted free, and ruled irrelevant — "these important advantages compared to classical computers do not change the problems connected with the SH analysed in this work... [the power estimate] already represents the maximum possible performance, obtained in the futuristic scenario in which a black hole can be used as ultimate computing device." [E]
- **Timestep relaxation** (§4.2): "What if (for reasons beyond what our physics can explain) high energy neutrinos can be accurately propagated in the simulation with a time stepping much coarser than the one prescribed by Δt=λν/c∼4.1⋅10^−32 s?" → still fails (t_CPU ~ 40–10^4 s per simulated second at Δt' = 10^−20 s). [E]
- **Bostrom-style caveat about who is being simulated** (footnote 4 + §4.6): granting the simulator free rein — "we surmise here the existence of a skilled simulator which somehow can use computing centres as large a planet-sized black hole, we can concede with no difficulty that it can also create a consistent simulation of the experience of ≤10^3 astronauts". [E]

**Exact scope of what he claims to rule out (Conclusions, verbatim):**
- Ruled out: "the impossibility of a 'Matrix' scenario for the SH, in which our reality is a simulation produced by future descendants, machines or any other intelligent being in a Universe which exactly is at the one we (think we) live in"; and, in the abstract: "it is just impossible that this Universe is simulated by a universe sharing the same properties, regardless of technological advancements of the far future." Note the word "nearly" in the title; and from §4.6: "the SH can be reasonably well tested only with respect to universes which are at least playing according to the Physics play book - while everything else appears beyond the bounds of falsifiability and even theoretical speculation."
- NOT ruled out, explicitly: (i) other-physics universes — "combinations of parameters exist, to make the SH for a low-resolution version of planet Earth possible (and similarly, also for higher resolution version of the SH), although they require orders of magnitudes in difference compared to the physical constants of this Universe" (§4.5); (ii) entirely-different-law universes — untestable (§4.6); (iii) the solipsistic variant — "a possible 'simulation hypothesis', which does not pose obvious constraints on computing, might be the solipsistic scenario in which the simulation simulates 'just' the single activity of the reader's brain (yes: you), while all the rest is a sophisticated and very detailed hallucination" (Conclusions). [E]
- Meta-caveat, Discussion opening: "Needless to say, in such a murky physical investigation several assumptions can be questioned, and a few alternative models can be explored... even if it is anticipated that the enthusiasts of the SH will probably find other ways out." [E]

**Bottom line for the project [S]:** Vazza's claim is: *the generic same-physics full-universe/full-Earth renderer is dead (energy), and even the most favorable low-res Earth renderer is dead (wall-clock time at the quantum throughput ceiling) — conditional on Δt ≥ Δx/c and Landauer per-bit heat at T ≥ T_CMB.* Everything turns on those three conditions; he grants every other escape hatch in advance.

---

## Q5. Reception

**Published, peer-reviewed engagement — one commentary, no technical rebuttal found:**
- **Edge E. and Brown C.A. (2026), "Commentary: Astrophysical constraints on the simulation hypothesis for this Universe: why it is (nearly) impossible that we live in a simulation", Front. Phys. 14:1808725, doi:10.3389/fphy.2026.1808725** (published 21 April 2026; listed as the sole citation of the paper by Semantic Scholar). Key passage — it explicitly *declines to engage the numbers* and attacks the framing:
  > "This commentary, therefore, aims to underscore a recurring conceptual framing error rather than to assess the numerical results or physical constraints themselves as originally formulated."
  > "Vazza [] treats the SH as a physicalist, pancomputationalist bottom–up schema, assuming that all physics at all scales must be computed for the SH to hold. His analysis, therefore, targets versions of SH that require large, globally consistent, experiment-respecting virtual worlds. This target is stronger than the minimal verisimilitude constraint emphasized by []."
  It quotes Bostrom (2003): "the microscopic structure of the inside of the Earth can be safely omitted. Distant astronomical objects can have highly compressed representations..." and Bostrom's 2008 FAQ: "Critiques based on the assumption that a simulation would have to be fully comprehensive [e.g., Vazza (2025)] thus miss the point." [E]
- **INSPIRE-HEP:** record 2911183 (Front. in Phys. 13:1561873, 2025) — **citation_count: 0** as of record update 2026-02-09. So the HEP-indexed community has not engaged it at all. [E]
- **Bostrom's FAQ itself** (simulation-argument.com/faq/) added the named retort: "Critiques based on the assumption that a simulation would have to be fully comprehensive (e.g. Vazza (2025)) thus miss the point. We may also note that even if fully comprehensive detailed simulations are possible... it would still be unlikely that we are in a fully comprehensive simulation, since simulators could run vastly more simplified simulations..." [E]
- **Blogosphere:** (a) simulism.substack.com, "Physicists Keep Refuting the Wrong Simulation Hypothesis" — praises the technical work ("Vazza's three conclusions are likely correct, clearly derived, and probably useful") but says it refutes "a form of pancomputationalism", not Bostrom's SH; frames it as the latest in a chain (Kaku 2023, Hossenfelder, Lloyd, Ringel & Kovrizhin). [A] (b) Lincoln Cannon, "Vazza Overstates Constraints on the Simulation" (transfigurism.org, 2025-05-28, Mormon Transhumanist Association): "I find his conclusion overstated and his apparent confidence unwarranted... It's premature to declare the Simulation Hypothesis 'impossible', or even nearly so." [A]
- **No published reply by Vazza** to the commentary was found (Frontiers lists no response article as of fetch), and **no citing paper that engages the cost model's arithmetic** was found in INSPIRE, Semantic Scholar, or the arXiv listing (his ADS "Cited by: 30" cluster — arXiv listing; the arXiv HTML itself carries "Cited by: 30" from ADS; most citations are press/blog). [S — search-based negative, bounded by the search scope]
- **Search caveat [S]:** the negative results are bounded by INSPIRE API, Semantic Scholar API, arXiv HTML "cited by" counter, and Brave web search on 2026-09-19; Google Scholar was not reachable directly, and citation databases lag.

---

## Appendix: what to take from this for the renderer-cost project [S]

1. **The minimal defensible model has three terms, all information-theoretic:** memory = bits(Bekenstein/holographic), energy/time-step = kB·T·log2 × bits (Landauer), wall time = bits × steps ÷ Lloyd-rate(2Ē/πħ). No FLOPs, no bandwidth — he never needs them because the best-case hardware (black-hole computer) sets a *universal* ops ceiling.
2. **One policy choice prices against one observable:** pick the smallest resolved scale Δx from an observable (Eν via λ=h/Eν); then bits ∝ Δx^−2 (holographic), steps ∝ Δx^−1 (light-crossing timestep), total ops per simulated second ∝ Δx^−3. This is the cleanest single policy→observable hook in the literature.
3. **The Courant step is the crux of his wall-clock argument** — if you relax Δt ≫ Δx/c (as §4.2 shows: 10^−20 s), the Earth-low-res case collapses from 10^7–10^8 yr/simulated-second to 40–10^4 s/simulated-second. His "impossibility" for the low-res case is therefore *entirely* downstream of that one timestep assumption; his universe-scale and full-Earth-scale results do not depend on it (they die on energy alone).
4. **Known soft spots to quote:** the unexplained "~10^6 timesteps" in §3.2; the "40 vs 20 orders of magnitude" Monte-Carlo discrepancy between arXiv v1 and the published version; the "4⋅3⋅10^16 bits" cosmic-web compression number; and the O(few) ops/bit/step floor assumption.