Field observations of maturely slipped faults show a generally broad zone of damage by cracking and granulation. Nevertheless, large shear deformation, and therefore heat generation, in individual earthquakes takes place with extreme localization to a zone <1-5 mm wide within a finely granulated fault core. Relevant fault weakening processes during large crustal events are therefore likely to be thermal. Further, given the porosity of the damage zones, it seems reasonable to assume groundwater presence. It is suggested that the two primary dynamic weakening mechanisms during seismic slip, both of which are expected to be active in at least the early phases of nearly all crustal events, are then as follows: (1) Flash heating at highly stressed frictional micro-contacts, and (2) Thermal pressurization of fault-zone pore fluid. Both have characteristics which promote extreme localization of shear. Macroscopic fault melting will occur only in cases for which those processes, or others which may sometimes become active at large enough slip (e.g., thermal decomposition, silica gelation), have not sufficiently reduced heat generation and thus limited temperature rise. Spontaneous dynamic rupture modeling, using procedures that embody mechanisms (1) and (2), shows how faults can be statically strong yet dynamically weak, and operate under low overall driving stress, in a manner that generates negligible heat and meets major seismic constraints on slip, stress drop, and self-healing rupture mode.