 What GPT-4 is, what can it do, why is it better than GPT-3 and how much, how do you get access and what new tools are being released. So first things first, it's multi-modal which means it can actually read images. If you're wondering what you would use that for, imagine that you're blind and you can't see, just like a normal person might use a translator when they're listening to a different language, a blind person might translate their environment right in front of them. Because GPT-4 can not only take that picture and tell them what is in it, but can explain the context around that.