Our current efforts are focused on using machine learning to predict where hot spots will appear and how fast they will grow. The idea is that we should be able to look at the PBX and make such a prediction. However PBX are opaque so we can only look at the surface. Now we have developed a method to make a PBX wafer that is about one crystal thick. As depicted below, this wafer is embedded in a transparent polymer so we can watch it as the shock passes through. A series of video images shows a hot spot evolving in time. (The white outlines show where the HMX crystals are located and the red regions are the hot spot). Reference: D. Sellan, X. Zhou, L. Salvati III, S. K. Valluri, and D. D. Dlott, In operando measurements of high explosives, J. Chem. Phys 157, 224202 (2022). The PBX wafers have a large enough area that we can shock about 50 spots on each wafer. Recently we have obtained enough data to show machine learning can look at an image of the wafer and predict where the hot spots develop and how fast they grow.