Reply to thread

Message

<blockquote data-quote="Jfdlsjfd" data-source="post: 9467809" data-attributes="member: 42856">And you're exactly right. It's a redesign. You've been misled by Microsoft's 100% human intelligence here (the intelligence wanting you to spend 20 bucks in order to use it more) in thinking the product you're using actually modify (as in, inpaint elements on top or merge new element into) the former image. It is just regenerating a prompt with more elements extracted from the &quot;discussion&quot; with the LLM, and creating a totally new image using this prompt. The AI you're using isn't designed for modifying an image, only for generating them from text. I am not aware yet of workflow where an LLM inpaints for you. There was an early attempt called instructpix2pix but it was like 2 years ago and no new effort was made in this direction as far as I know ever since.The pro version can do some very limited edit (but it's more like a collage of independantly drawn elements, so not an answer to your need).</blockquote>

[QUOTE="Jfdlsjfd, post: 9467809, member: 42856"] And you're exactly right. It's a redesign. You've been misled by Microsoft's 100% human intelligence here (the intelligence wanting you to spend 20 bucks in order to use it more) in thinking the product you're using actually modify (as in, inpaint elements on top or merge new element into) the former image. It is just regenerating a prompt with more elements extracted from the "discussion" with the LLM, and creating a totally new image using this prompt. The AI you're using isn't designed for modifying an image, only for generating them from text. I am not aware yet of workflow where an LLM inpaints for you. There was an early attempt called instructpix2pix but it was like 2 years ago and no new effort was made in this direction as far as I know ever since. The pro version can do some very limited edit (but it's more like a collage of independantly drawn elements, so not an answer to your need). [/QUOTE]

Verification