Why glasses behave like solids in the absence of their having any long range structural order, is a fundamental problem of statistical physics, one that has been actively researched for more than 80 years. Supported by the mean field theory of supercooled liquids and a deep connection to mean field spin glasses with one step replica symmetry breaking, the random first order transition theory offers a solution to the glass problem based on assuming proximity to an underlying ideal glass transition. In the deeply supercooled liquid the free energy landscape is dominated by metastable structural basins separated by large free energy barriers. The rate of inter-conversion between these structural states is ultimately driven by the entropic cost of remaining confined to one basin, a cost which is quantified by the configurational entropy. Both the activation free energy barrier and the number of cooperatively moving particles required to overcome the barrier diverge as the ideal glass transition is approached. The cooperative nature of the dynamics in the deeply supercooled liquid regime has been confirmed by experiments and simulations and has been the subject of intense study in recent years. In the following we explore the implications of cooperative dynamics in the random first order transition theory with particular focus on the expected behavior at the ideal glass transition temperature and at the dynamical crossover, the temperature where activated motions first become important. We also show how the general features of secondary relaxation can be recovered by adding local fluctuations to the equations describing cooperative reconfiguration. Finally, we describe how cooperatively rearranging regions modify dynamics near the surface of glasses, reducing the apparent viscosity by several orders of magnitude.