Generate an image of a long-overgrown suburban street in Okinawa, with the central road winding up a hill and disappearing as it rounds a corner in the mid-ground. Let there be an anime-style girl in a painterly style standing in the middle of the road in the near ground, wearing a barn coat and a backpack, with dark brown hair. Use a 1:1 aspect ratio.
I wonder if ChatGPT pulls some Google Street View images to use as reference, because this actually legitimately looks like one of the small back alleys in Ishigaki.
Or did you give it some photo as reference that we're not seeing in the metadata?
No, this is totally all on ChatGPT. I guess it just has enough training data and contextual inference to prompt DALL-E very well. I've been to Oki as well, and south Kin looks just like this.
No, this is totally all on ChatGPT. I guess it just has enough training data and contextual inference to prompt DALL-E very well. I've been to Oki as well, and south Kin looks just like this.
I'm going to remove the self-upload tag until they explain because they're commenting about iodoff's bot putting self-upload on their other Drawfag posts and this is probably why.
Edit: Changed my mind, I'm actually going to remove the drawfag artist tag on the posts that aren't sourced to 4chan and seem to be actually made by them.
I tag myself as drawfag if it's my own gens simply because I want to remain anonymous and have no desire to be a followed artist, but do want to contriubte to the culture. My name here is literally randomly generated.
Drawfag serves that purpose for 4chan anon's, so I'm hitching a ride. Alot of my uploads are not mine, anon posts from 4chan.
I understand wanting to isolate drawfag to 4chan, but the wiki cites it as "anonymous posters on western imageboards". By technicality, danbooru and aibooru are western imageboards, and allowing anonymous users to self-upload under drawfag overcomes any potential artist_request call. Your guy's call, and I'll happily comply going forward.
