I think the confusion is you're reading this like it is meant to be a presentation of next-gen text-to-image models. It's more like a fancy UI iteration. And I think it can find use cases in different tools.