“Gaurav Misra identifies ByteDance's recent video model as a key milestone, being one of the first large generative models from a major company to successfully produce expressive dialogue-based video.”