In this Letter we examine the claim by Baganoff & Malkan that the timescale of brightness variations of quasars increases with decreasing optical wavelength, and that this effect can offset a time dilation effect at high redshift. They base their case on a thin accretion disk model that implies that timescales become shorter as the observed passband becomes bluer. We show that light curves of both Seyfert galaxies and quasars conflict with the predictions of their model and that neither the timescale nor the structure of AGN light curves changes with color. We discuss the consequences of this result for accretion disk models of quasars and the relation of our result to the microlensing mechanism for quasar variability. We also look at the implications of quasar microlensing for the nature of dark matter, and use some recent results on quark masses to support the idea that the dark matter is in the form of primordial black holes.