You can now access an editing view for images inside GPT. This allows you to select an area and write a prompt for what you want changed about your selection.
So far, it's a little buggy and pretty unreliable (like most text-drive image editors) but it's a sign of things to come as OpenAI moves more into the GPT UI as a sort of AI command center for other agents/GPTs, Dall-E and whatever comes next.
コメント