This is interesting:
I logged onto ChatGPT, but somehow I was in the version 3.5 world, which does NOT connect to DALL-E.
I provided the following image prompt:
Widescreen image. A happy cute robot reads his morning newspaper while drinking hot chocolate on the patio of a home high up in the mountains. The morning light is warm and golden. Birds flutter in nearby trees. The style is akin to that of Ed Emshwiller.
Because 3.5 does NOT do images, it responded with this:
I’m just a text-based AI, so I can’t create visual images, but I can certainly help you visualize this scene with a description:
In this picturesque widescreen image, a delightful and adorable robot with a friendly smile sits comfortably on a wooden patio chair. The robot’s metallic body is sleek and polished, reflecting the warm, golden morning light that bathes the high mountain home. The air is crisp and refreshing, and the mountains stretch majestically in the background.
The happy robot holds a morning newspaper in one of its dexterous robotic hands, and with the other, it delicately sips from a steaming cup of hot chocolate. The cup is adorned with a charming pattern, and the cocoa’s aroma mingles with the clean mountain air. The robot’s digital eyes are filled with curiosity as it peruses the newspaper’s pages, absorbing information and enjoying a moment of leisure.
All around, the trees rustle with life as colorful birds flutter among the branches, creating a harmonious symphony of chirps and tweets. Their vibrant plumage contrasts beautifully with the serene mountain landscape.
The scene, reminiscent of Ed Emshwiller’s artistic style, captures the tranquility of a morning in the mountains, where even a robot can find happiness in the simple joys of life.
Hmm, I thought. Wonder what it will do if I switch to ChatGPT4, which DOES do images, and feed it this much more elaborate prompt ?
I did so, and it came up with this lovely image.
I need to explore this further, and do some compare and contrast tests.
Per Spock, "Fascinating . . . "
Anyone have info on this 3.5 prompt generation thing, before I go off and spelunk a bit ?? For example: is it just revealing what 4 does before talking to DALL-E ?? Or something else ???
– stan
