Will you be adding iterative image generation, similar to what is available in ChatGPT? The ability to add multiple images as references and then modify the same image even after it's generated, keeping it as a reference, would be a great game-changer for all users. It would make life so much easier if you could incorporate this feature.