Raw LLM Responses

Agreed, most of the expression in this clip comes from the eyes, basically no mouth motion at all except for a slight open near the end. I'm assuming there is more to this demo and this segment just showcases eyes, which it does very well. The biggest hurdle to lifelike expressions, as i understand it, is not the animations, we have very good animators and can sequence micro expressions without needing AI. Rather, it is the shwer number of actuators and control points needed to mimic the huge range of expression the human face is capable of. Certainly, AI will make real-time expression generation much faster and more dynamic, but it will only be as believable as the android it is controlling.