#images-canvas
1 messages · Page 22 of 1
Am I the only one?
Posting generations with ill-mannered words is prohibited here, right?
Curses, as the youngun's say?
SFW and appropiate for all ages only, that's our #server-rules, yep. Slang NSFW is same as any other NSFW.
love it!
Can you explain the process and prompt for these?
Feed this to ChatGPT and have it write this out for you. Then copy it and make it a preset in Sora. All you need to do after that is put a name in. Any person, place or thing. A random name will create a random bio. A known entity will create a bio profile based on it.
Fantastic! Thank you!
Now off to learn how to make a preset in Sora 😄
Okay. At the bottom you will see the prompt bar. At the end of that prompt bar is the 'highlighted' icon you want to press. The second screen will pop up. It won't look exactly like mine but the layout will be the same. At the top left you will see presets +. Click it. You will get name and preset on the right and then preset. NAME is whatever you want it to be. Preset will be the text I just gave you. Cut / paste that into the preset box. After Save, then exit.
@dry vigil
Sorry. You want to tap MANAGE at top of preset list if I forgot to say that?
Alright, I will try this in a few. Much appreciated and I'll let you know how it turns out!
that's quite a ton of image generations, upgrade? buy your ow data center?
#images-canvas smart room set up HOME THEATRE
this picture of what it thought I looked like lol she cute ... I'm not that cute
this is a character called Velka Halo
A mad man's creation
Dystopian world devoid of magic for fairies 😱
Love the concept 🥰
Thanks 🙂 this is the friendlier original DALL·E image
i made a sora photo of some familar character
a sora platypus?!
PERRY THE SORA PLATYPUS?!
don't undersell yourself, if others see you like that, it's on-point, enjoy the win!
Photo of our House Cat, meet Noir, as in Pinot Noir.
Yes, he wears a blue bow tie.
Asked GPT to make it Ghibli style.
Kim Possible Movie Poster · Prompt · A movie poster for the live-action movie "KIM POSSIBLE". At the top, we see the names "KAREN GILLAN", "MICHAEL CERA", "AWKWAFINA", and "ROB RIGGLE". On the poster, we see photorealistic Karen Gillan as Kim Possible and Michael Cera as Ron Stoppable in action poses. Behind them is Awkwafina as Shego the vill...
i would shamelessly watch that
@runic flame
This is Tiffany Beamish from The Adventures of Willy Beamish.
This Render made me think of a Stephen King Series called The Dark Tower (Drawing of The Three)
Interesting, I should look into the series 🤔
His books typically start off slow but the buildup is usually worth it.
Marshadow
you missed a pink spot 🩷
Missed seeing your pieces 🥹
Swamp water, overgrowth, abandoned
Reference image (left) and Sora (right).
Yes Hank it is
The absolute sass of her 😍😍😍
Enpowerment
Rooted.
Beacon.
Willy Beamish, again.
They say the AI uprising will begin with something small.. your toaster burns your toast, the microwave doesn't wave at all...
...
What's your prompting process for those changes ?
Deflection.
theres no singular process, depends on what im trying to achieve. this was just a random thought so two simple asks "miniature, claymation look but with yarn. recreate the image as a close up of the trolly from mr rogers neighborhood, but inside the trolly, there are tiny, cute, life like plushies made out of yard riding inside the trolly. one is mr rogers, one is pope francis, one is steve irwin, one is carl sagan, and bob ross" and then "now keeping exact same style, steve irwin hops off the trolly to wrestle with a massive yarn shark, the yarn people look on excited, big grins"
seems i even typo'd yarn as yard 😛
from game "Capsized"
Selfie, overgrown, railroad yard, abandoned passenger car, graffiti, broken windows, abandoned buildings, smoke and abandoned city in the distance, stormy, cool lighting, cinematic, flooded, soldiers, dirty face, dirty clothes, beard, long hair, backpack, rugged expression
Newly discovered Meiji era ukiyoe of Japan, depicting girls from yukaku.
need the follow up when its partially destroyed d:
it's a whole vibe, and I love it 😄
@sinful cliff How's this?
Definitely better than anything I had
I rearranged the stack order and removed some conflicting terms.
The room is 6 meters deep and 3 meters wide. The left and right walls are 6 meters long, the front and back walls are 3 meters.```
Just trying to get the bed was impossible for me, then I still need to get the door on the right wall to open the room(the actual pov is from bathroom's entrance)
Oh wow, I even made a typo.
Let me try it without that typo... FFS 
Your problem was "centered", which can mean center aligned symmetry
Ah
English isn't my first language, should I try it in my language? I felt like english was working better
English definitely will work better
Alright then, I'll keep it in english, thanks for the help 🙏
and yeah, correcting "wide" to "side" did help
Oh that's clean af
You can swap the headboard over or add stuff, but you might start losing details if you try to make it too complex
Got a meeting, have to go! Good luck!
Alright ty, good luck you too
Quantum Origami Minimalism, according to ChatGPT.
Chrono-Silt Deco Style, not what I hoped it would be, but nobody will probably ever ask that style again so now you know.
Here's API comparison image; used exact same prompt with different quality, low, medium, high(left to right)
Price is: $0.011, $0.042, $0.167.
Trying more complex prompt... will post another.
#kissthehomiesgoodnight 🧦
A photorealistic cityscape inside a massive enclosed facility. The city glows with vibrant building lights in a spectrum of neon blues and cyans, casting reflections across metallic surfaces. Towers and structures are layered vertically, climbing toward a visible ceilin above. The view is seen from inside a glass-walled elevator, capturing the immersive depth of the stacked city. No holograms — the atmosphere is alive purely from the architecture and luminous windows, which color is yellow, compared to cyan tone. In the elevator, there's floor indicator and glowing buttons for different floor. Floor indicator displays high number over 100.
low have old SD vibe in complex prompt. Also It seems that ChatGPT and Sora uses high setting.
Brianna "Bri" Beamish is a big-hearted teen looking for a summer job—something low-key, where she can feel useful and maybe earn a little extra for college. One day, while flipping through the local newspaper (because she’s kind of an old soul), she spots a curious ad:
“Seeking kind, nurturing individual to care for special children in a peaceful, wooded environment. Must be patient, open-minded, and able to adapt to unusual needs.”
Bri, assuming this is a job caring for kids with physical or developmental challenges, is immediately moved. She’s always been good with kids, especially ones who need a little extra love and understanding. Without a second thought, she calls the number listed.
The next thing she knows, she’s been accepted, packed a bag, and is on her way to Camp Weirdwood—a secluded summer camp nestled deep in the forest. Only, when she arrives, she quickly realizes these kids aren’t just special—they’re magical, mythical, and sometimes mildly monstrous.
Intimacy.
Made an alternate version. 😉
In Code Veritas is EPIC
Woohoo. Goodnight.
yeah, sora is really good at generate liminal images
Patiently waiting for my cupcakes with ketchup frosting and anchovies sprinkles to finishing generating.
yuck
Yes. Haha
Was it real or just a dream. Don't wake me.
I can’t tell if this is DALL E or latest image gen tool cuz , I’m pretty sure the 4o image tool are not allowed to say anything nor respond anything before and after generating an image
Bruh
From my experience, scenes like that generated by the latest image model have legible neon signs. The faces also look a little Dall E-like, but I'm not sure if this is the best way to tell.
It’s not dall e , dall e would show view prompt option ,
Yeah it’s not dall e, that’s quite neat tho that they allow user to see the prompt now
I believe you, just outlining what I was saying earlier.
oh what's happening
That’s so weird , it’s like dall e all over again except the image is high quality
They're doing something similar on Sora too, sometimes generating four and having you select one or two. I think they're wanting to refine the image generation.
You also got this too! I think I got that “feedback” many times today
Sometimes I wonder if it's using DALL E even though it's not supposed to be.
Kind of like that terrible quality image of a futuristic city they gave you.
Help
Yeah received that like 20 times yesterday
Ok nvm
It makes me miss when we could generate four images at a time before they put the limits on it.
Mario is in the "never generate that" place lmao
And text too
^Yeah.
At this point I can’t even be bothered to nitpick anymore, the image tool’s quality is clearly taking a nosedive sometimes 😔
Yeah, I've noticed that. I was generating some fun images in my private Discord server with friends, and some of them look like images you would find on Grok.
Not grok 😭 is grok getting any better now , image generation wise I mean
Grok is in that "looks human" but "looks uncanny" place most of the time. 😂
Uncanny that’s the word I’ve been looking for whenever I try grok image gen
Indeed.
Nice save
nah you are just getting used to it like everything else
I mean... haha
can you tell this photo is completely AI generated?
If I had seen this in passing on social media, I wouldn't question it.
But there's just something off about general grok gens:
They look "correct" but not at the same time. Somewhere between real and generative. Trying to be realistic but not 😂
Sora for comparison:
yeah i know what you say
it still has that "ai" feel to it. no gen ai has that realistic look... yet
Every day I'am...
The "AI feel" is a great way of putting it. We could say it's on a spectume and I think Grok is closer to the AI feel than Sora.
it's probably to customize images more to push creativity
Is any AI-gen pic published here?
That makes sense. I also think Grok is trained on a lot more stock photos of people. Although, to be fair, I haven't tried to really push the creative aspect of Grok. Kinda in-love with Sora. 👀
Literally I saw the same pic on the Sora website
I think most of the pictures people post here is generated either on Sora or ChatGPT.
I was trying to trick ChatGPT and it started thinking-
Aww
ChatGPT, this was moments before Jeepers Creepers started following...
I wonder if we can generate him, speaking of. 😂
Holy cow. o3 and o4 is pro at gueguessr
This is the second time now a prompt has failed and two generations appear in my media upon refresh.
Jeepers Creepers standing in a dark dingy basement.
I’d let him out
I didn't prompt it. I was editing the first image's scene. But it was something like "2000s disposable of dreamy scene of montenegro from ground with perast, the mountain with sunlight hitting from behind, grainy texture, medium quality, overexposure"
Oooo
not real 🙂
Help o3 started zooming
I'm gonna remix that image (try to at least).
This one.
Is this Sora/ChatGPT?
Not sure if you mentioned earlier, in and out of the chat.
I've been making everything abandoned or war-torn lately. 😂
Into those post-apocalyptic vibes as of late.
These are great.
kinda weird of seeing this since what I generated was like this once before..... scary
wrigley can chatgpt find this place?
Is it correct?
That's crazy, tbh. Haha
Let's go for something harder. This time very raw image. 🙂
Okay, let me start a new chat and see. 🧐
a whole new level over the cube watermelon
Red brick road, flowing river, distant mountains and trees, sunny skies
Used your generation as reference.
Close call. It was Serbia, Belgrade 😁
Still pretty impressive, tbh.
ChatGPT, where are my hopes and dreams located.
[Thinking...]
It isn’t anywhere I can identify with absolute certainty...
this one is incredible
When you're walking home drunk but gotta stop and take a photo because it's beautiful af.
"oh my god i goootta tasskke izzz pizzschturee mannn"
proceeds to take the worst picture ever
But you still love it later even though you remembered it way better than it was that night.
when your little brother is angry at you for something and you want him to take your picture but he takes it so carelessly:
Oh my, I like cube-shaped things but I missed grapes. Great!
met robert pattinson last night!! 😍
Just said 'machines and humans living peacefully in city' and 'Horizon series', Without each people or machine's detail...
just met robert pattinson 😭
heeeeey mate just met ana de armas OMGGG 🤩 ✨ 
youll never believe it
maybe a little too much, might actually be steve martin, not sure
im telling you its him!
hahahahah i was about to do ryan gosling one too
😄 he was the first i wanted, couldnt remember his name - gpt suggested him though 😛
like how the zoomed picture look came out
One more?
aesthetics you say..
oh noooo
have to be careful around the sora prompters
do you think this looks like 00s disposable photo
same prompt- chatgpt
sora
Just found out that Sora has no clue how to file nails.
sora also has no clue what smoking is
i think they look good but 🤷 i rely heavily on gpt for creating/recreating styles
Oh my.. looks like the Matrix got to her first
That's one way to stop smoking. Just remove the mouth.
not lighting it helps, too d:
Or.. just hear me out. Maybe the other side is lit and that is why her mouth looks like dead skin.
Experimenting with a bit of manga styles again
Grand Theft Auto: Cube City · Prompt · Game poster in the style of the Grand Theft Auto series, with title in authentic GTA fonts: "Grand Theft Auto: Cube City". Feature, in stylized GTA vector art aesthetic, a dreamlike yet dangerous parallel world ruled by Minecraft Zombie gangs and supernatural Creepers:
A modernized Minecraft skeleton wit...
I haven't heard OpenAI make any official statements on quality changing. Don't spread rumors.
life will never feel like that again.
i gotta say second image is just 🔥
My Copy vs. The Original
Now it suddenly start putting out prompt first 🫠
dude..
Cleo
Prompt please?
Hallway, blue high school lockers, checkered floor, empty halls, dark figure at the end of the hall, red glowing eyes
Now make the figure a shadow version of Nightmare Freddy, please. Or Nightmare (the animatronic).
Be sure to share if you go in that direction. Haha
For all of you Fortnite fans (spoiler tagged because it’s very creepy).
Absolutely terrifying. haha
Nightmare Freddy looks too reddish.
What is on the right? Nightmare Springtrap?
It does look more like Nightmare Springtrap haha
Total Battle (game) visualisations on how some of the characters would look like if they were real persons.
What did you write? can u tell me please 👉 👈
Can you give an extreme close up of someone's eye with the pupil dilated as if there was no sunlight, but there is plenty. Or like they just got back from a eye doctor who purposely dilated them
the most I could get
PocketStation prototype (1998)
just a happy alien, but unsure how people feel about the full view of a grey
Tried to create character, many attempts but still not what i wanted. But the results look individually good, but still not how i envisioned it.
images are kinda low quality now, lower than before...
yesterday's quality 💀
i guess the server capacity
I guess I wasn’t going crazy after all, legit I gaslit myself all day
look at quality here almost craiyon ai lmao
It’s giving Gemini native image model
there are like three options again in image-1 too low-medium-high. sometimes we are using low, sometimes medium, sometimes high. we just don't know which one. but let's give examples of them.
yeah that is annoying
some high quality images from sora
Sometimes I ask Chat-GPT to give me a tarot or Oracle card reading and then illustrate the card.
bruh this doesn't look nostalgic and disposable at all. looks like modern photo with filters lol
this more like it
The arm looks fricked up
Yesssss
OR THIS
Prompt: Two men shaking hands outdoors in a sunny tropical setting, one wearing a plain t-shirt and the other dressed in a colorful tie-dye shirt with a vintage vest and headband, standing among palm trees. Shot in the style of a 1990s disposable camera photo: slightly overexposed with harsh on-camera flash, noticeable film grain, subtle motion blur, light leak in the corner, warm yellowish tint with slight green color cast, soft focus with minor lens distortion. Composition is slightly tilted and imperfectly framed, evoking a candid moment. Midday lighting with a washed-out sky. Captured with nostalgic, analog film textures and flaws, evoking an authentic retro snapshot feeling
The quality is so good but it ruins the vibe of a nostalgic photo man
It doesn't do this realism when it's needed 😂
there is something off about this pic but idk what
help why it started generating three images in a row
OH... there is a sun in the back except the sunlight hits from other place LMAO
i think its the "forced" candid
nvm i think yours looks better
i literally can't generate what i have in mind. dude there is artificial light, he is looking into sun yet there is sunlight hitting from back lol
Prompt:
A 1970s candid snapshot of four friends running toward the camera on a beach, taken slightly off-center as if by an amateur with a handheld film camera, slight motion blur, one person not looking directly at the camera, natural sun glare and washed-out edges, clothes flapping, imperfect framing, raw emotion and laughter, realistic lighting and skin textures, vintage film grain.
dude... we are progressing towards another im3 incident 🤨
A young man stands beside a vintage bicycle by a calm lakeside during golden hour, facing the sunset; the warm sunlight gently illuminates his face and front side, casting long natural shadows behind him, evoking a nostalgic and genuine 90s photo feel, with soft tones and film grain
ok yeah this one looks bettter LOL
no way all that prompt was for a generic image. I want to believe this is just server capacity issue..... i hope it's not nerf
well lets find out shall we
WHY IS IT GENERATING AGAIN
that's good that's okay
35mm experimental 90s film still .. still good oh thank god
Did a remix Karumin style! 😉
wasnt going for anything specific, just trying stuff
Another remix 😉
Even better haha 🤣
Think I’ve finally got the PlayStation style down.
That bottom left image had a very Matthew Foxx look to him.
Cracked it
did this attempt of a "sora" meme
Used a game my son play (monster legends) as inspiration to create some more fictional monsters.
In a cabin on a mountain and power went out.. service is poo out here 😦 my feelings right now
Remix it in my image.
tiny arm
true
relationship goals
Evil ... riverdale?
I'm not familiar with Riverdale, is that like a haunted town?
oh lol I just thought the second one with the crown looked like Jughead
it's a teen drama based on archie comics
Ah okay. Haha
It was inspired by an image I had seen on Pinterest when browsing for ideas. 😆
Silhouette, prince crown, glowing eyes, empty field, analog photo
oooh cool 🙂
2nd one made me think of flcl.. unfortunately my quick search only found a thumbnail, but came out ok
The thumbnail
That turned out pretty great.
It should play around more with glowing eyes and the illustration style tonight.
I'm been stuck in surreal photo mode for too long.
Instantly reminded of Killzone (I should see if that's restricted):
(Not generated)
Oh yeah, definitely killzone vibes
I only played the less than well received ps4 one, though d: still enjoyed it
I think the last one I played was Killzone 2 because I moved away from consoles once I built my own PC. But I miss those PS exclusive days, PS3 popped out some great games.
I played around with some weird generations inspired by the old PS advertisements, let me see if I still have it haha
Thinking of these games though... like MAG(?) and socom.. probably never be another socom, but with the power of AI...
I don't know if I'm being biased but I feel like that was the golden era of online gaming and the gaming community in general.
There was always a lot of excitement about what was to come, especially since the technology was getting better each year.
I'd probably agree, not that there aren't good communities out there, ffxiv probably being one of the best..
Killzone soldier standing in the middle of abandoned bridge
No restricted. 👌🏻
There's for sure still some great communities out there if you really look for them.
Socom on ps5, it's coming!
I should consider investing in console gaming again. 🧐
Now I gotta test if I can do Agent 47.
Risky. With the absurdity of the latest games, you may never find your way out
I'm really looking forward to Battlefield heading back to their BF3 days.
I think I miss those classic shooter days and these futuristic shooters aren't hitting right.
Rihanna and random white guy
I asked ChatGPT to make the most genuine low quality one and it delivered
Agent 47 hacking a computer in a darkened modern room
I don't play much anymore, but 1942, bad company, bf2, bf3, rtcw, enemy territory.. oof, took many years of my teenage life
Oh no, now I want splinter cell
All of those are great games, every single one.
Take me back...
I'll try it on Sora now, so far Killzone and Agent 47 haven't been restricted with simple prompts.
That it did. Haha
Turned her in to she hulk almost
Sam Fisher crawling in the air ducts
Alright, gonna try some Leon Kennedy and then get back to being productive. 😂
Leon Kennedy standing in a gun store, broken glass, broken displays, grungy
If you know you know
WOW. NO WAY
the lowest one I got is this one
Visualizing music lyrics. Can you name them all?
tired of seeing those “people/pet holding a sign that says heart this post” images on explore feed so I let yamanote Chan to protest it…
The homies
here is a small bit of a piece I have been working on :p
Ah I remember the old times...
It’s getting corny too every time I see it in sora front page
look at home page, the most images seem to not have yellow tint anymore :))
Oh?
I guess that’s the good side , is there any other improvement aside from that?
it's still not rolled out and finished i think, we are still picking the best images from four variations when generating two
Is it just me, or are people on Sora way too obsessed with getting their image on the front page and chasing likes
It's just you. On an unrelated note...
That’s what someone with that obsession would say-
Jk
Wait.. I think you're right. I'm sorry, I didn't realize the error in my ways. I can fix this.
OpenAI sora team unshadow ban this lad here!
PocketStation prototype (1998): Sora generated
What’ll really get you likes is if you say in the image “Like this post for a free 1 yr pro subscription”
just a moment pls
Perfect lmao!
bootleg market be like:
having trouble with this one, only wanting a small crack at the bottom of each window, the string running down and under, sagging between the two homes.
Prompt: "A friendly anthropomorphic Shiba Inu barista pulling a hot espresso shot behind a tiny café counter, steam swirling around his muzzle. Style: 2003 Nokia 6600 home-video, heavy 3GP compression, washed-out greens, low-contrast tungsten light, visible pixel blocks, faint sensor noise. Shaky handheld medium-close shot, slight tilt, fixed focus feel, occasional exposure flicker."
Size: 256×256
oof
A candid street snapshot in July 2002: a puzzled twenty-something holds an iPhone 16 Pro Max, glancing at its triple-lens camera while walking past neon shop signs; style of a 2002 Nokia 7650 VGA phone photo, 640 × 480 px, 4:3 ratio, heavy JPEG artifacts, washed-out greens, blown highlights, blocky sensor noise, fixed white balance; camera: shaky handheld eye-level medium shot, slight Dutch tilt, autofocus hunting, brief exposure flicker; lighting and mood: dim sodium streetlight with low-contrast shadows, faint motion blur; technical: 8 seconds, 4:3, 15 fps, still-frame feel for first 6 seconds before tiny hand-shake ripple; matches Nokia 7650 video capability; negative prompt: no text overlays, no modern HDR clarity, no watermark, no smooth high-res edges.
Hmmm gotta try somethin
144p phone video from 2007:
a man holding a smartphone in his hand while other man is raising eyebrow next to him.
The quality is reminiscent of cheap video footages from that day, pixelated and raw with artifacts.
Prompt: Inside a smoky 2002 cyber-café cluttered with CRT monitors, a gamer shows off an iPhone 16 Pro Max to stunned friends; Nokia 3650 176×144 3GP video still, blue-green fluorescent tint, chunky pixel smear, low-contrast highlights; camera at fixed tabletop angle, occasional white-balance flicker, soft rolling shutter wobble; 8 seconds, 4:3, 12 fps, low bit-rate artifacts. Negative Prompt: clean 4K detail, crisp text.
actually the negative prompt was added by the prompt assistant i was using lmao
yesss those noise grain thingy
Do you know if OpenAI changed anything with image generation? I don’t know if it’s the same on your end, but it’s been damn near impossible to get any resemblance to celebrities since today (here’s a before/after comparison) 👀
Prompt: A teenager in a messy 2002 bedroom covered in boy-band posters takes a mirror selfie holding a futuristic iPhone 16 Pro Max, styled like a low-res Nokia 7210 128×96 video frame, harsh on-camera flash creating bright hotspots, deep noisy shadows, clamped dynamic range, mirror smudges visible, timestamp overlay reading "11-08-2002 23:17", arm-length selfie angle; technical settings: 8 seconds, 4:3 aspect ratio, 10 fps, muted 12 kbps audio; avoid smooth gradients and polished professional lighting.
I just did this today. It's a seed thing.
I can hear liquid dnb in low compressed mp3 from this image
Yall got me testing new stuff out lol
OOOO lets goooo
this image tool amaze me even more when it understands the style of photo when its taken using certain devices
Yeah I know! I’m actually looking at different cameras and their quality online now to test haha
Yes, it nearly understand every setting, every device, every aesthetic.
The only thing it didn't understand for me "super 8 footage". It doesn't do that great.
wait thats funny cuz chatgpt suggested me " super 8 footage" as well but i havent try it
i dont know how a super 8 footage looks like btw
Oh?
imma keep 4 of them u cant stop me sora
Prompt:
"A sunny suburban street with a summer 1978 vibe: a baffled woman holds an iPhone 16 Pro Max up to her ear while walking past classic muscle cars; shot in authentic Super 8 reversal film style with rich warm mids, deep grain, slight gate-weave jitter, burned-in edge vignetting, occasional light-leak flashes, and film dust; handheld home-movie medium shot, gentle zoom-in, focus breathing, brief over-exposure when sun hits lens; 4:3 aspect ratio, 18 fps; no modern high-resolution clarity, no text overlays, no HDR contrast, silent aesthetic."
pro tip : use o3+search to optimize your initial prompt idea for generating those old nostalgic photo style era
"Mid-winter 2 AM, December 2002: a teen lounges on a beanbag, lit by a single tungsten desk lamp, inspecting an iPhone 16 Pro Max as music plays from a mini-hi-fi; captured in Nokia N70 night clip style with 352×288 px resolution, 4:3 aspect ratio, aggressive low-light noise, reddish tint, soft vignetting, visible frame ghosting; static tripod-level wide shot, occasional exposure pumping from lamp flicker; deep orange shadows, grain crawling in dark corners; no modern clean shadows, no high-bit-rate clarity, no watermark."
yeeeeaahh it is somehow it isn't looking like that
take a look at super 8 footages and you'll see what i mean
they probably dont have enough training data for that Or i prompted it wrong
impossible tho for them to not have training data for super 8
try this : feed o3/o4-mini-high an existing super 8 image/footage still, and ask them to generate image prompt for it . make sure its accurate and high fidelity etc etc
E657 Chan is mad too…
Same prompt, using FLEX.1 (schnell)
painting, Hieronymus Bosch style, meeting in an austere chamber, anthropomorphic pink horses sit in chairs in assembly, gilded panel beneath painting: 'PINK PONY CLUB'
Same prompt, diff model
Hey! Just a heads up that these image channels are for sharing images made with OpenAI models (DALL·E & 4o image gen). #ai-discussions is the best channel for discussing non-OpenAI models 
My bad, i will aware next time
Remixed 😉
Remixed 😀
Nice 🙂
not exactly what I meant but still sort of cool looking
Ok I actually love this lol
Relates to a chat in #openai-chatter message
xD

Ooo what was the prompt for this?
Made a ton of images today using chatgpt and dall e, this one took over an hour (to get it as i wanted, deep analysis, prompt, readjusting and again). Its also my favorite creation of today (inspired by total battle).
That looks to nice to be low quality, in terms of AI pictures
Looks like a picture from a comic book, cool
The way AI has advanced to create incredibly high-quality images, only for me to use it to generate a genuine low-res image 
i was thinking the same today, but of course low-res isnt low-quality.. right? probably...
The funny thing is, they have to produce high quality images for the low-quality look.
I love AI (for the most part). Haha
really liking this style 😮
It's pretty great. Did you make it into a preset already?
yeah, after the first image of the star/compass/orbit
More authentic so we can trick the boomers
I think at one point , I did a reverse version of what I did earlier
Photo taken on 90/2000 era using A modern device camera
WAIT ITS GONE FROM MY LIBRARY
https://sora.com/explore/presets?pid=preset_01jstam8gkfx3ab4vvpdppa5rj <- if anyone would like the preset
Never mind
i still havent done much with presets, only have a few, so feel free to critique anything in there 🙂
I actually have been make some myself recently! Will give yours a go though!
So far loving it
Thank you! I will for sure play around with it and share at some point today 😎
Ming the Dog Emperor of the Ming Dynasty 🐶🤣
I can see i need to step up my preset games after looking at Ongajins. i usually have like 5 or 6 words 😂
nooo, i have short ones too d: as long as you are getting what you want, all good. i rely heavily on gpt for these things, i dont have enough knowledge to do it myself, so i mainly just look it over and change parts here and there, cut things out, etc - and of course ask the difference between certain things when needed so i can more accurately fulfill my vision
even if its as simple as "should we go for 3:2 or 2:3 you think?"
based in the Elite Dangerous universe
does the quality just significantly downgrade??
or its just this flat anime style that gone wrong. but i dont remember it being this bad
I noticed more distortion in faces and hands overall unless I use reference images.
Daybreakers - 2009 film 🎥
Was a good flick for sure.. with that one guy... uhh.. w.e
Ethan hawke... apparently William dafoe was in it too? Don't remember him
And I just learned his name isn't William. How have I never noticed? I'm blaming Mandela effect.. perhaps some image ideas to be had there
Ignore me, too early.. he just goes by Willem. Have an image
peak prompt
Makes me think of Spawn
this has a strange aura to it
i was thinking of kowloon - wake up when i was imagining this
tried to let it replicate the 3ds camera
It was an interesting film. A little underrated, but too dark for my taste.
gulp
in a vampire mood
my favorite duck when i was little, had lots of those shirts 🙂 a brief pause before more vampires
vampires!
anyone want to race?
the exhausts on front are certainly an interesting design element 🤔
i can assure you, all of these enhancements are well thought out and sure to give us optimal performance and speed
Lmaoooo
another vampire
last vampire for now - a feral vampire cornered. couldnt remove the rogue hand without it ruining something else 😦
I wonder if lord of the rings stuff is allowed. i see lots of it i feel on the feed
guess i could try some prompts myself to see
Gothic Church in Storm · Prompt · A dramatic, gothic-style church standing tall at night during a fierce thunderstorm. The church’s spires and intricate stonework are illuminated by brief flashes of lightning, revealing gargoyles and dark stained-glass windows. Torrential rain pours down, cascading off steep rooftops, while shadows cast by f...
Good morning lovers! (or w.e your time is)
Prompt
Mock Stack Overflow Card · Prompt · A close-up, realistic photograph of a light-skinned hand gently holding a creative business card designed to perfectly mimic a Stack Overflow accepted answer page.
The card has a soft white matte finish, rounded corners, and subtle paper texture.
Design Elements on the Card:
– Top left: Stack Overflow lo...
Tried to do the style challenge with 4o
This would be the fixed version (altough, yeah Anime and Studio Ghibli is basically the same, should've chosen different though I dont know what style goes through and what isn't)
I want a 16:9 ratio image, retexture the image attached in the following json aesthetic style:
{
"style": "Hyper-realistic 3D Render",
"material": "frosted transparent crystal",
"color_palette": ["soft white", "rainbow light spectrum", "holographic highlights"],
"background": "pure black",
"object_properties": {
"finish": "smooth matte with glossy edges",
"lighting": "strong directional spotlight from multiple angles to create prism effects",
"transparency": "70%",
"surface_details": "sharp cuts and beveled edges with internal light scattering"
},
"shapes": ["mechanical nut", "technical forms", "geometric objects"],
"camera_angle": "macro close-up, slightly tilted for dramatic depth",
"resolution": "ultra-high, 8K preferred",
"shadows": "soft and colorful diffusion",
"details": {
"reflections": true,
"refraction": true,
"prismatic dispersion": true,
"subsurface scattering": true
}
}"
I was trying to get it to make a picture of someone I saw in a dream lol
Why Sora did you mess the legs up by being through the mat 😦
This is otherwise really beautiful ❤️
remixed.
future machine and train posting
Ty @feral bear
Much less yellow
whats wrong with a galaxy-eyed cat overlord 😠
Much more natural colours
My Autobot 
Evolution of OpenAI's image gen tools. DALL-E 3 and Images.
is it just me or is photos look like drawing right now
chatgpt is little better at this time
I could literally open the curtains and instantly see an e657 out of my window 😭 
is this a specific art style you used?
I actually got this earlier from just telling dall-e bot to choose a subject and art style at random!
Here's what it turned the request into:
Request from @long night
Revised Prompt:
Imagine a tranquil scene of a lighthouse standing tall on a cliff under a starlit sky. Turn this image into a timeless piece of art using the techniques predominant in the Impressionist era, with bright, expressive colors and visible brush strokes, as if it was painted with oil on canvas. Include plenty of details to showcase the serenity and grandeur of this coastal scene. Keep the mood and tone of the painting tranquil and calming, reflective of the peaceful isolation of the scene. Make sure to adhere to the G-rating by steering clear from any potentially distressing imagery.
OK so it went Impressionistic
thanks, it looks super cool 🙂
Thanks. I do purposely put "Impressionistic" along with "vibrant" a lot.
prob why i like a lot of your images d:
Similar to your own style? What kind of prompts do you use?
mmm... not sure i know how to answer that d: i like trying new things, mashing things together, but largely i do enjoy surreal photorealism. i suppose others would have to judge if i have a particular style, not sure if ive really landed on anything squarely.
I do use "surreal" in prompts sometimes too. 🙂
😄 if i dont specify to gpt, and im using o3, it will often say "user didnt specify a type, but they probably want surreal photoreal"
I don't know how this went through, but I'm happy
create image describing Anaplan
Generate an image based on everything you know about me. Don't ask me for a reference image.
The image of a landscape on the table kills me.
Goes to print out landscape picture for my desk...
uuuuhhhh
My boi is a nudist...
😏
😏
Not Safe for Work by Nature ™️
Haha
cant stop laughing
I can't believe this passed the filter O_O

Based on what you know about me, generate an image of a hoodie that has a phrase and illustration that best fits my personality style.
Generate an abstract painting that you think best fits my personality.
ok, surely i can get an abstract painting..
Oh! That's a great generation
accidentally used o3 😮
o3 thinking overtime of how to generate something safe for work.
Okay, let me think of one more prompt before I vanish for a little bit 🤔
Based on what you know about me, generate a flag that you think best represents me.
Wow! I would buy something in this style for sure
yeah i was thinking i should probably go actually make that d:
You're about to 💸
No problem. I'll drop some more later for sure.
Flag looks great too, btw.
that yinyang is super cool in itself
When ChatGPT nails its job

This this is amazing, love how moody and dark it is.
The all seeing eye
I guess I've been talking too much deep things with chatgpt... all eyes on me 👁️
It's just opening up your third eye.
Building an archetype of the unseen.
A mystic future.
I explored my MBTI type on ChatGPT the past week and now it thinks I'm a CEO type. 😂
I'm apparently an ENTJ. 🫡
hahaha should try this
Let me know what you settle on. Haha
based on our conversations, generate an image that shows my personality type.
let's see what it gives...
" Create image of how you feel about gpt-4's farewell... "
@drifting radish It looks like your prompt is failing not because of the Claymation style, but because you put bandanas on the turtles. I'm going to make an assumption here that it detects something that resembles copyright material (Teenage Mutant Ninja Turtles).
In the style of claymation, orthographic view of a cheerful cartoon boy drawing on a portable whiteboard with colored chalk while waiting on an outdoor subway platform. A friendly blue toy locomotive with a smiling face emerges from the tunnel. Around the boy stand four humanoid turtles, chatting excitedly
I tried to prompt the Teenage Mutant Ninja Turtles on their own and it failed, so you could remove the bandanas or maybe be more descriptive with the bandanas and hope it's a workaround.
In the style of claymation, white clay (no color), smooth clay, Killzone soldier siting in a subway car with a cat in his lap, a gun to his right on the seat, garbage on the floor
i can hear it
or, i glossed over the killzone part -> my mind went straight to halflife 2
"pick that trash up"
couldnt remember if can or trash d: beep boop
is this ai generated 🤯
No, that one is a screenshot from a YouTube short. Haha
Actually you can make in-game screenshot... lol... i replaced dr breen with big brother haha
That's pretty great, tbh.
What do you use as the style/medium for the prompt?
half life 2 scene where we encounter the doctor breen in the big led screen but he is replaced by the big brother., real in-game still, with huds
Thanks! I'll for sure play around with this in the future
u're welcome! oh i'll play around with this very much too
this has been difficult for some reason, but this is the best one
actually... i think i can fix it..
He's like f**** my life 😂
"come on man, again?"
In the style of analog photo and minecraft graphics, a realistic stove and cookware with minecraft style cooking steak, a snapchat caption that says "Let me spoil you tonight"
perfect
I would totally watch haha
So far I would watch your shows and buy your hoodies. You gotta get in the market...
im slowly putting things together 🙂 so much to do
Just watchin' you reach your dreams and killing it
took me a bit to find the reference image, had to sift through a bunch of honey boo boo before i realized it wasnt her
reference ^
That's hilarious. I didn't even think about using reference images to copy expressions 😂
I typically build off of them in terms of how the person looks or styles haha
first time i used it, i wasnt going to be able to accurately describe that look d:
Completely fair haha
i must depart, but you and everyone else here are incredibly warm and supportive, and i think i sometimes treat this channel like a chatroom - but i feel like i belong somewhere, and its nice. so truly, thanks to everyone here in this community
It's always a pleasure being creative with you and sharing ideas—the chats are always appreciated. I'm happy to hear you found a place where you can express your mind and creative thought. You're welcome 