AI generated art

[Fambo Number Mive]

Rontos Chips and Roy Walker

[MrsWarboysLover]

Ricky Gervais laughing at a disabled man in a wheelchair

[Tikwid]

A NEW CONTENDER APPROACHES (AGAIN)



https://huggingface.co/spaces/dalle-mini/dalle-mini

[TheGingerAlien]

"Candyman the Handyman" as rendered by Dall-e 2


Variation:


Anyone else on the public beta for this monstrosity?

[TheGingerAlien]

These are all outputs from the prompt
"rontos chips supreme ruler of the universe"

[jenna appleseed]

Nightcafe doesn't know what a Paddington or a Peppa Pig is :-(

'Paddington Bear and Peppa Pig have a fight over who loves The Queen more photorealistic superreal 70s painting'

[Bence Fekete]

[Bence Fekete]

[shoulders]

Btw, some of the text to image AI results seems to improve if you put in hi spec camera models and screen resolutions as variables.
.
Eg: "photo by nikon, photo by canon eos rebel t7, uhd, highly detailed, octane render".

[The Late Satoru Iwata]

I don't know how much of a CaB faux-pas it is to resurrect this thread but it seemed weird to start another one when this is right here.

Anyway, DALL-E 3 is out and on the whole it can do text now (when specific text is defined. I assume if you ask for a wall full of film posters, it'll create the usual AI junk text shapes). It's not 100% reliable.

 

 

[The Late Satoru Iwata]

(I assume if you ask for a wall full of film posters, it'll create the usual AI junk text shapes)

Bingo:

[idunnosomename]

Amazing to see that brief aesthetic of warped nightmarish AI images again.

These big squishy uncanny images we have now arent as fun

Also it blocks celebrity prompts

[The Late Satoru Iwata]

True enough. I don't know if DALL-E Mini is still about but that would do some fucked business.