Predictions: Future Versions of DALL-E 2, Midjourney, and more

If you missed it, in September, I proposed new ideas about what a future version of DALL-E, Midjourney, or Stable Diffusion could look like. I believe these predictions will be possible as early as next year! I spent most of this summer conceptualizing what could be possible next.

It’s Part II of my series GPT-X, DALL-E, and our Multimodal future, I renamed it to GPT-X, Diffusion, and our Multimodal Future:

I also created an accompanying book as well which outlines the ideas and is available free as a google doc.

Key Ideas

Recombinant Influence control:

Natural language prompt editing capabilities:

Advanced natural language edit prompt understanding:

New ways to prompt image to text models:

Something I call, “Logical Variations”

… and so much more!

If you’re interested in what next year could bring for Creative AI tools, I can’t think of a better, more comprehensive resource outlining what’s possible!

Also, some of these predictions have already started to happen, such as Stable Diffusion approaching real time image generation speeds.