Comment by umutisik

And very likely this is coming to computer vision too with multi-modal GPT.