Comment on "Astrophysical constraints on the simulation hypothesis for this Universe": the power bound rests on a global time-stepping assumption

Vazza (2025, arXiv:2504.08461) concludes that even a low-resolution simulation of Earth would require a power of ~10^73–10^75 erg/s, and is therefore physically impossible. We show that this figure is the cost of one specific algorithm: every one of the ~1.65×10^51 bits describing the Earth is advanced at every global time step Δt = λ_ν/c ≈ 4.1×10^-32 s, and that step is set by high-energy neutrinos, which almost never interact. In an event-driven scheme with deferred evaluation (EDDE), the cost scales with the number of interaction events, not with (bits) × Δt^-1. Free propagation is evaluated analytically, and the neutrino-imposed time step disappears. With no coarse-graining at all, that is, at atomic resolution throughout the planet, the energy dissipated per simulated second falls by about 21 orders of magnitude, to ~10^47–10^54 erg. Because running more slowly than real time lowers the power but not the energy per simulated second, we use the latter as the measure of cost. Even at the lowest temperature, the remaining cost is several solar rest-mass energies per simulated year, so we make no claim that a simulation of Earth is feasible. Our claim is narrower: Vazza's bound measures the cost of an algorithm, not a limit set by physics. Any further reduction depends on how much mass must be resolved at atomic scale, and on whether coarse-grained models can reproduce every observation of the rest. That question is open, and it can be tested with geophysical data.