Astronomers have long inferred the Solar System's acceleration from stellar proper motions, yet a new Gaia-based analysis of quasars shows prior error bars were too narrow. The preprint (arXiv:2605.30367) applies the pseudo-C_ℓ formalism to both the Gaia EDR3 quasar catalogue and the Quaia sample, then uses simulation-based inference to jointly fit the dipole while marginalising over higher multipoles. This yields a best-fit acceleration amplitude of 5.72^{+0.53}{-0.52} μas yr^{-1} from Quaia, with components (g_x, g_y, g_z) = (0.40^{+0.70}{-0.70}, -5.09^{+0.54}{-0.54}, -2.40^{+0.55}{-0.58}) μas yr^{-1}; the credible intervals widen by factors of 1.5-2.5 once degeneracies are accounted for. No redshift dependence appears, supporting a purely kinematic origin tied to the Sun's orbit around the Galactic centre rather than any cosmological dipole. Earlier Gaia Collaboration results (2021) and the 2018 VLBI-based determination both reported tighter uncertainties without fully modelling scanning-law systematics or stellar-density correlations now diagnosed here. The study remains a preprint and relies on two catalogues whose selection functions differ, leaving residual cross-correlation power that future DR4 data may resolve.