Skip to main content

On Sale: GamesAssetsToolsTabletopComics
Indie game storeFree gamesFun gamesHorror games
Game developmentAssetsComics
SalesBundles
Jobs
TagsGame Engines

So I saw today that there is a new AI text to image generator called Stable Diffusion. You can try it out here.

I ran the same test and here's what it came up with:


I think we can all agree that these AI results are much better. That being said, if I were trying to use this as concept art for a game, it's in the right ballpark, but you can tell by looking at the images that they have this "odd" look to them,  like if you tried making something with real Lego bricks and an assortment of knock-off Lego bricks. 

(+1)

I have to mention- there are apps now where you can refine your prompt / re-run certain sections of the image to get the program to 'try again' on the parts that seem off. This is how DALL-E's studio mode works. Users are able to pretty easily get rid of the mistakes/dreamy qualities of the image

Midjourney will put out a ton of images per prompt, then let you 'breed' the ones you like (selecting images, then re-running the prompt with those images as influence) until you like what you see