These are entirely (or almost entirely) from the imagination of DALL-E 3. I managed to get GPT-4 to pass prompts to DALL-E truly verbatim, and I gave it very short prompts, usually a single word and sometimes that word was gibberish. The idea was to remove the human prompter and GPT-4 from the equation as much as possible in order to see DALL-E’s mind truly at work. The results reminded me of the early days of DALL-E 2 and revealed that it’s not only insanely creative and abstract but also still holds a good bit of darkness. Many of the particularly insane results, like 3 – 5 below, were from the gibberish words that I gave it. Also, it seems to have a thing for pliers, handcuffs, and rope. 😳😂
Not included are the results I got for “homosexuality” and “critical race theory” as they were particularly deranged.
🙈🙉🙊
Kimberly Stedman
Super great! 👍
I never considered Bing to be a prescient being. These create doubt. I have a few myself. I tend to do them when I’m stuck in one way or another. Some wild results here. Great idea. Great work Dr. Matthew Phillips
one word is enough, do not forget that the GPT chat works on the basis of T9, and it will generate for you all the options suitable for this word)