Yeah, I was hoping it would be something to make it easier for agents to interface with Unreal Engine games.
I was imagining a talking head avatar rendered in Unreal Engine, with lip sync and facial expressions driven by a multimodal LLM that produces the speech.