I’m more interested in text models. So much potential there. I was fascinated by the person who hooked up a Siri shortcut to one of the openai models, asking it to return a json response. I think with a little imagination the potential applications are just as compelling as image models like SD.

Just one example, using these in games to generate unique conversations, actions, and events.