Some guy's thread about Dall-E.

Thread Starter


Joined Apr 2, 2020
I saw for the first time earlier today. You can ask it to create an image of anything. Add the word photorealistic or sketch or rendering to get the kind of image you want.
I asked for a chicken riding a bicycle on a beach...

View attachment 275203
It gets confused by some language issues - the server is overloaded right now as well so some strange things get returned. Each request gets nine different versions (images). Post your request and results if you try it.

Moderation: Restored this first post so that the posted replies have some context.
Last edited by a moderator:


Joined Jan 27, 2019
Craiyon (previously DALL•E mini) is amazing stuff. It is a fork of DALL•E which verges on incredible, and much higher resolution. The DALL•E project asked them to change the name to avoid confusion, hence Craiyon.

Some of DALL•E’s output is frighteningly real, and like Craiyon it’s purely the response to a text prompt.

Here are a some DALL•E mini pictures I got followed by some DALL•E pictures my son sent me (he has access to the beta).

DALL•E mini / Craiyon

Prompt: “World’s Largest Duck”


Prompt “Beauty”


Prompt: “Giant squirrel on a hill in the distance”


These Images are all from the prompt “Big Chungis* ad”. I wish I had some of the photorealistic output, but I can‘t find any. I will get some and follow up.
Note the AI model does not understand words that appear in pictures, that is, it can’t understand the semantics but it does recognize how some aspects like form and color suit some prompts, so you get very odd not-quite-right words. The word ad or advertisement puts things in the context of advertising and generates some surprisingly ad-like results.
* “Big Chungis” is an odd meme that uses an early, much larger and fatter Bugs Bunny that didn‘t end up being the final version of the skinny Bugs we all know.






Joined Jan 27, 2019
MY thread is about a free, user-accessible AI image generator - ( Please don't drag it off-topic with images from a version that is INACCESSIBLE (except for your son). This thread was not meant as another place for you to compete - just a place to share/discuss/laugh at some AI images that users define and created in 2-minutes or less by an AI image generator.

Please be creative and show us the phrase you asked for and the image.
It is not “only my son” that has access to the beta, you can apply for it if you’d like. They are just limiting the rate at which they provide access.

Craiyon. as I said, is a fork of DALL•E, and used to be called DALL•E mini. The first pictures are from that. It is the same model given a smaller learning set and less CPU time. It is not something different. I was adding information about what you posted, totally on topic. Maybe you didn’t read the whole post.