I think I major issue with LLM based tech to users is that there really isn't a way of declaring what the tech can concretely do.

The camera app? Takes picture, can do night mode photography, boomerang, burst, etc.

The AI magic image / LLM app?? Idk just try stuff as a user until you can somewhat find the bounds of it.

This is a similar issue to what lead to great hype and then fall behind voice assistants but to an even greater extent.