Dark energy from the string axiverse

String theories suggest the existence of a plethora of axion-like fields with masses spread over a huge number of decades. Here we show that these ideas lend themselves to a model of quintessence with no super-Planckian field excursions and in which all dimensionless numbers are order unity. The scenario addresses the "why now" problem, i.e., why has accelerated expansion begun only recently, by suggesting that the onset of dark-energy domination occurs randomly with a slowly decreasing probability per unit logarithmic interval in cosmic time. The standard axion potential requires us to postulate a rapid decay of most of the axion fields that do no become dark energy. The need for these decays is averted, though, with the introduction of a slightly modified axion potential. In either case, a Universe like ours arises in roughly 1 in 100 universes. The scenario may have a host of observable consequences.