Happy Fireworks EditionPrevious thread: >>101222587>What is DALL-E 3?It's a text-to-image generator made by OpenAI.>Linkshttps://www.bing.com/images/createhttps://designer.microsoft.com/image-creatorhttps://copilot.microsoft.com/Bing AI Slop (0.7.0)https://pastebin.com/raw/0rre6VvS (embed)Designer AI Slop (0.5.1)https://pastebin.com/raw/eUSVC7Ew (embed)>How does it work on Bing?In Bing's Image Creator, "boosts" are used to speed up the image creation process, each user starts with 15 boosts. When you use a boost, the AI creates your image quickly. If you run out of boosts, the image waiting time increases, ranging from seconds to minutes.>Resetting BoostsTry clearing your search history at https://www.bing.com/profile/history>Landscape and Portrait regenerationFollow the instructions here: https://rentry.org/bicedit>Prompt Generatorhttps://tipseason.com/dalle-prompt-generatorhttps://files.catbox.moe/o4yihf.pnghttps://files.catbox.moe/320rm1.pnghttps://files.catbox.moe/8gge6o.pnghttps://files.catbox.moe/dcokjf.png>Related Boards>>>/v/dall-e>>>/co/dall-e>>>/trash/aibf>>>/aco/dall>>>/tg/slop
>>101252114Why are her shorts that up high, almost touching her chest?Does she think dressing like that will hide her fat?
>>101252428hehe
>>101252458what was the art style
>>101243853I suppose something similar is possible, but, absolutely just in case, I forgot to mention that you need to click on the "description" text field in order to see that size button, it's hidden when you load the page.Yeah, I remember the photos that I've seen recently of that black wolf in Alaska, Romeo, playing with dogs, he was certainly significantly larger than any dog in those photos, but he wasn't so large that he completely dwarfed a labrador. Funnily enough people were even discussing riding wolves in that thread, apparently their shoulder bones aren't really suited for that? But I guess a wolf as large as that one in your picture wouldn't have any issue.>>101254804Very cute.
Here's what the AI outputs if it's anything other than a person looking straight at the viewer, the more it censors the worse it gets.
>>101256029That has nothing to do with censorship. Dall-e 3 and other AI image generators struggle with multiple characters in poses like that all the time.
>>101257079Before it used to struggle with hands and fingers. My point was that if you censor AI from the get-go then it's limited and won't ever get past a certain threshold.
>>101257116Dall-e 3 isn't even censored on a model level, dumbass. It's not SAI tier nonsense.
>>101252114>People still making gens of this girl after so longCute. Glad there's some sort of lingering Thread Thing that hasn't gotten censored away, even if it's just a silly girl.
>>101257395>Dall-e 3 isn't even censored on a model level, dumbass. It's not SAI tier nonsense.It is censored at least something is, guess we are still learning what. Something about maybe seed numbers are removed or blocked? or the ref id numbers or that other number I forget what it is. If its blocking the people then it could be blocking reference data about hands?Maybe that other anon is right. AI doesnt really know what a human is, and in the end its just an image generator. For instance if I type person with suitcase, the AI doesnt care and their not real people just pixels. What I mean is because its not a real person it could just have the person standing near the suitcase or fold them up inside the suitcase or a person as a suitcase. Ive seen this when using a closet and seen people stuffed in the shelf of closets like nightmare fuel. If their blocking say celebs then do we know if its blocking the model data gathered from their millions of images and if it is then maybe that is why hands are deformed? Anyone got more info on the 3 ids found the Seed id the ref it and that other id? I dont think these can be used in any way I dont think their fixed or static available say for reuse if you like a scene or character, i think their just informational but useful to us as users in say recreating the scene using the numbers without the prompt even if we punched it into some api tool. This is theoretical since we dont know but given that services like bing and google are using the entire image database as their checkpoints I assume this would be more like a LARGE scale ai not a small one with limited datasets of checkpoints like SD. Maybe there is no method as of yet using the numbers to say recreate a similar scene or character using them I think it might be per instance? It would be nice in the future if the next api version had some kind of hashtag system to use the same character, art style and scene, useful for artists and developers.
