A few months ago at work, for a virtual meet up activity, we played Google Draw. It was a fun way to see who could draw and have Google’s AI guess fastest.
GPT3 offered up some crazy language examples of language generation a few months ago. A shopping list from one seeded sample? Check. A whole story? Check. A book? Yes.
So now it’s offering up a (limited) example of images it can create on the fly. You can check it out here. I wish the demo offered up an ability to enter free text. Instead, it looks canned. Still, if it works as described, it’s going to be an amazing tool.
My gripe about OpenAI is that unlike its name, it’s actually not open. It’s not clear how to access except through a limited beta. Even a few demos with text have been shut down.