The phenomena that have drawn the most attention in molecular magnets such as Fe_8 can be understood at the single molecule level -- a large spin in an anisotropic crystal field. Yet that is only half the story. The slow and glass-like demagnetization or magnetization of the solid as a whole, its response to a slowly swept external field, etc., entail collective behavior arising from the dipole-dipole interaction between the spins of different molecules, and pose a challenging theoretical problem. The talk will discuss how this behavior can be understood in terms of a long-range Glauber-like model, how this model can be justified microscopically, and how it can be formulated in terms of a kinetic equation. Monte Carlo simulations of the model, and numerical solution of the kinetic equations display several (but not all) aspects of experimentally seen behavior, and many questions remain open.