We talk about NVIDIA’s DLSS 4.0 multi-frame generation (MFG), NVIDIA’s weird decisions on testing tools, differences in transformer vs. CNN (convolutional neural network) models, benchmark performance, and generated frames. Frame generation commonly gets referred to as “artificial” frames, “generated” frames, or commonly online, “fake frames.” This video delves into the topic of NVIDIA’s MFG, which is absolutely worth learning about, but also digs deep into its decisions to take ancient PresentMon code and then write what we think is a misleading review guide about how it works, updating it seemingly after we raised the alarm in discussions.