Can Bitcoin's Hashrate Fix the Kardashev Scale?
A 60-year-old assumption says civilisations grow their energy use exponentially. Six decades of real data say otherwise — and the fix runs through Bitcoin's hashrate.
Correction (11 July 2026): An earlier version of this article said the linear model was statistically preferred over the free-rate exponential (ΔWAIC = 5.5). That comparison omitted its standard error. Computed from the pointwise predictive densities, SE(ΔWAIC) ≈ 18, so ΔWAIC/SE ≈ 0.3 — well within one standard error of zero. The correct statement: the linear and free-rate exponential models are statistically indistinguishable on the 1965–2024 data. Unchanged: the falsification of Kardashev's 1%/yr assumption (posterior r = 2.01% ± 0.03%/yr). If anything the Conundrum sharpens — the data cannot distinguish functional forms whose extrapolated Type II timescales differ by ~12 orders of magnitude, so the choice of functional form, not the data, drives the extrapolated future. The text below has been updated accordingly; the preprint was corrected on 1 July 2026 (v5).
