There's definately a bit of an uncanny valley in the land of top tier diffusion models. A generative video of someone smiling is way more likely to illicit this

There's definately a bit of an uncanny valley in the land of top tier diffusion models. A generative video of someone smiling is way more likely to illicit this response for me than a generative image or single frame. It definately has something to do with the movement.