It's a recurring theme in OpenAI research, they become more and more closed.
For instance their latest model called DALL·E hit the headlines before the release of the paper. Needless to say, the model is not available and no online demo has been published so far.
Because its winner-take-all in this research, not "winner-take-some".
Andrew Yang talked about this and why breaking up Big Tech won't work. No one wants to use the second best search engine. The second best search engine is Bing and I almost never go there.
Tech isn't like automobiles, where you might prefer a Honda over a Toyota, but ultimately they're interchangeable. A Camry isn't dramatically different and doesn't perform dramatically better than an Accord. Whoever builds the best AI "wins" and wins totally.
But they still released the CLIP model which is the complement of DALL-E and used in the DALL-E pipeline as a final filter. There are collabs with CLIP floating around and even a web demo.
Thank you for this info, as you mentioned CLIP is used for re-ranking DALL-E outputs, by itself it is just an (image, text) pairs classification network.
