I am amazed by level of comprehension that DALL-E shows. At first glance the results also look amazing. Zoomed in though it turns into the stuff of nightmares. Quite literally actually. I am an enthusiastic practicionor of lucid dreaming and this stuff really feels similar to what I see when I closly observe details in lucid dreams. Ostensibly everything looks real but this reality falls apart when actually observed.
FYI, openai has eased the rule on realistic face generation. Now you can generate and publish Photorealistic faces. They will internally filter those to make sure they don't match famous faces.
I would love to see what comes out with certain aspects of the prompts negated.<p>- "lemon gelato that’s been shaped to look like a heart, on a handmade waffle cone being held up to the camera in a cobblestone courtyard somewhere in italy" ... what about "somewhere not in italy"?<p>- "Diorama made of clay of a group of computer programmers looking disapprovingly at their CMO who has just given them diet pepsi instead of mountain dew" ... "looking approvingly"?<p>- "friends gathering around a tabletop “shichirin” grill where an assortment of meats and seafoods are being grilled over glowing binchotan charcoal; everyone is happy." ... "everyone is unhappy"?
the way it generates specific art styles and textures is amazing. it's interesting to try and spot out subtle details that it misses/ignores entirely (who is drinking diet coke out of a mug? why are the mummies interpreted as skeletons in sock-raincoats?)