The standard model of (at least low-mass) star formation envisages collapse of cold molecular cloud cores after substantial diffusion of magnetic fields over a period of several Myr. Strong magnetic fields were presumed necessary to slow gravitational contraction and thus avoid overproducing stars in molecular clouds which last for some tens of Myr. This picture is in direct contradiction with observations of stellar populations in nearby molecular clouds, which are never older than a few Myr. The evidence of stellar ages demonstrates that molecular cloud formation in the solar neighborhood is generally rapid, that star formation follows immediately upon cloud formation, and that the molecular gas must also be dispersed rapidly. In this dynamic picture, the low rate of star formation is due to a low overall efficiency of converting gas into stars rather than slow collapse. Drawing upon large-scale simulations of the dynamical interstellar medium, I outline a physical scenario of formation of molecular clouds by large-scale flows which can satisfy the observational constraints. I conclude with a detailed analysis of star formation in the Taurus molecular cloud which suggests protostellar cloud core formation by gravitational fragmentation in filaments.