I’m also excited if they can make an image generator you can chat to and iterate. That would be awesome.
You might know Thai already but have you tried using ControlNet to make images with stable diffusion? It allows you to input an image that it follows along with the prompt, and there’s a bunch of ways to influence it. You can even give it a hand drawn sketch. Or have it pick out position of limbs, or use the edges of objects and keep those.
If you have something specific you want to create then it’s amazingly helpful.
The only easy to use site I know of that offers it is happyaccidents.ai (or you can run yourself if you have SD installed)
This conversation reminds me of the old Star Trek:The Next Generation episodes in the Holodeck, only they're talking to "the computer" to iterate on a holodeck scenario design.
You might know Thai already but have you tried using ControlNet to make images with stable diffusion? It allows you to input an image that it follows along with the prompt, and there’s a bunch of ways to influence it. You can even give it a hand drawn sketch. Or have it pick out position of limbs, or use the edges of objects and keep those.
If you have something specific you want to create then it’s amazingly helpful.
The only easy to use site I know of that offers it is happyaccidents.ai (or you can run yourself if you have SD installed)