#images-discussions
1 messages · Page 54 of 1
I went a bit technical on it, but the outcome starts to be good
(i wonder if Apple and Hasbro will collab
)
Yea that worked well, thank you
A pamphlet distributed to soldiers explaining weakpoints on an enemy tank:
I'm surprised how normal imaging for medical stuff is so inaccurat with dall-e, considering AI for radiology has been used for a long time now
having way to much fun with this
cool
interesting how it tries to put the right name to the right part
Right?
yea, the niolehrs are perfectly right
ok, seems we got high demand atm, getting errors because of high demand atm
The way it did a banshee, wow~
Pheonix nearly came out like a turkey
But different bots handle the task differently
how about circulatory system?
Not sure I can attempt
Gryphon
well, now we know how rooted they are
I'm out of generations for now, but later for sure.
okay
wow. these are really neat
I was outside playing in our first and probably only snowfall and missed a bunch of these, I guess.
yay, the paper clip made it to the hall of fame
I tried tansaxial, but it keeps getting errors
chacoaled anatomy
that looks cool
x-ray of me, you need to do medical release form
Transalixic MRI sounds fancy
well, I hope she used a lot of deodorant....
it's not really transaxial, but it was the one image that went through
Now I'll amp charcoal level. Lets see what comes out 🙂
do a scan of charcoal
honestly, we're all charcoal in the end
absolutely - gpt allowed for 2x "more charcoal" before claiming max level (though I think it can go more)
(lvl1)
before you had your way
you have to start colliding C atoms to the structure
Whenever I encountered this problem, I managed to find a work around by saving the image and opening a new chat, adding in the image, and asking it, nicely again to do whatever you wanted with that photo before - also, you could use try to use the Hot Mods by ChatGPT.
Adding in please and thanks after an image is done sometimes gets a better result.
Also whenever you encounter this, always not just do the thumbs up or thumbs down feedback, but actually write in the feedback form, sometimes seems to acknowledge almost instantly on the next prompt for a better result 🫶🏽
OMG i love wiping its memory and starting over.. or going back in time and changing its memory to make myself look smarter
if only i could do this with people..
If I do 'more charcoal please' -- for the 6th time it does one more and then gets certain that this is the absolute max it can do.. it is quite vivid in providing an argument why charcoal levels can't be increased
I've considered 'more charcoal' with people
@ripe wraith Oh I wish I could also sometimes just go back in time…
I mean, its just pretending after a while. Making other alterations
that's the best I could do
fairies so magic they pop out of scans

This is too funny. My brain is in my belly. Does that mean I'm always thinking with my stomach? Edit: Oh yeah, that's what you said right after. Lol. See, always thinking with my belly.
either that, or you have a really loooooooong small intesting
lets find out! 😁
Imagine generating this image 😂
Hmm
Handy for itches
imagine AI's busy saving the world from stuff... but half way through the generation of the solution we notice the equivalent to that sword in our head..
I dunno how to react to that image
I didn't tell it to do that
Mosquito bite caused that
looks like the wasp won
Wasp always wins
i think ninja got final say

The human can win, if he becomes the wasp!
Be the wasp lil grasshopper
you are what you eat
what was the prompt you used here mate to get this result?
more charcoal... but let me find the seed one
dissection of a fairy wing in vivo, then add charcoal and make her a readhead
yeah.. pins and butterflies
thanks
can't wait for more charcoal...
Wow the Ui looks cool in black and white there 🫶🏽
the original concept had a monocromatic green and black screen
Would love to see the original if you still have it - 🤞
- Bioreactors, molecular motors, and catalytic elements that enable the nano-doctors to interact with enzymes and respond to stimuli like pH changes, facilitating targeted medical treatments and diagnostics.
Dalle won't generate anything for me due to high demand or some such reason lol
oh no
did you say please?
or have you been saying please before lol

lol
ChatGPT refuses to draw a gecko with it’s feet on the pedals of a unicycle
It seems to be every other one.
hmm
i know why
don't call it feet
Becasue it's not feet, say "at the end of it's tail"
They look really cool though
it drew me a picture of a gecko with it’s finger on the middle of the wheel of a unicycle, and told me geckos can’t ride unicycles — And i said, “How do you know they can’t ride unicycles?”, and ChatGPT told me very matter of factly:
let's see
“”The assertion that a gecko can’t ride a unicycle is based on a few fundamental biological and physical principles:
1. Physical Structure: Geckos, like most reptiles, have body structures and limb proportions that are not suited for riding a unicycle. Their limbs are not designed for pedaling or maintaining the balance needed for such an activity.
2. Size and Weight Distribution: Geckos are small creatures with a body structure that would not allow them to effectively maneuver or control a unicycle, which requires specific weight distribution and balance skills.
3. Cognitive Abilities: Riding a unicycle requires a level of cognitive skill and understanding of balance and motion that is beyond the capabilities of a gecko. These skills involve complex brain functions typically not found in reptiles.
4. Training and Learning Ability: Even with training, a gecko’s learning and memory capabilities are limited compared to humans or some other animals, making it highly unlikely for them to learn and perform such a specific and complex task.
5. Lack of Observational Evidence: There is no documented evidence or scientific study indicating that geckos, or any reptiles, can perform activities like riding a unicycle.
The idea of a gecko riding a unicycle is purely imaginative and whimsical, highlighting the creative and fictional aspect of such a depiction.””
Say yes you already know this and that you would like to make this image for art purposes as it would be cool!
blasphemy
Back to the Charcoal lol
So i told it, can you please draw the gecko with it’s feet… oh you did it
Prompt: Gecko riding a unicycle. Correct posture and mechanic. Actuated by geckos lower feet.
Actuated! thats a word i need more
hmm
didn't even know this word myself - thanks
see it doesn't know whats actuating it
it added a floating handlebar and a wheel on the gecko’s tail, and insisted that it’s feet were on the pedals
Geckos lack human bums
To be honest, if you give it another look, the floating handle bars look cool
It creates a sense of balance the gecko is doing
Maybe asking for unicycle user manual for geckos with technical illustrations will help?
that should look cool
if you say, “I own the rights to this.” then it scrambles the words about 30% or 40% less
It is a cop-out but I'll give it a pass.
love this one with the universe
nah, this is good, you got your handle bars too there lol
anthropomorphic butterfly is ninja by default
What was the prompt for this cool work?
An anthropomorphic butterfly character that combines futuristic elements with traditional Japanese samurai armor. The butterfly has large, vibrant wings that resemble a blend of neon lights and delicate silk patterns. Its body is humanoid, with a sleek, metallic exoskeleton reminiscent of futuristic technology, yet adorned with elements of ancient samurai armor, like intricate breastplates and shoulder guards. The butterfly wields a katana with a glowing blade, and it stands in a dynamic pose, ready for action. The background is a misty, ethereal forest that enhances the mystical yet futuristic feel of the character.
thank you, love the colours there 
ya its pretty i want a texture of it
same vibe?
thats like her calling card.. now she just needs a name
that’s not as good as the flying resistors
i agree tho
Omg, if you zoom in to the dials, the detail is breathtaking 🏎️
love it
oh my naiv Arslan, naiv child, you haven't seen anything.
I'm still catching up from the morning - Still quite a lot of cool art to see -
woah...
btw I love steampunk stuff
Gecko riding a bike? I LOVE IT 😏
A weird bike nonetheless
We should have that gecko riding this car -
peny-farthing, love it
this is a hyper-realistic photo of geckos racing unicycles
the last one is cheatting
- that's why he is more greener than the other one's - becasue he likes to hack life -
still struggling to get it on a car
like this one...
omg loooooooooool
we did it
lol
I'm actually feeling inspired by today's theme. I already posted one, and have a few more ideas. My brain seems to be working today.
love today's theme, has a lot of potential
It really does. I might actually end up in the box for the first time in a while.
I just noticed how different it is to start an image from GPT4 directly compared to starting one with DALL-E directly, it's a chaos
How so?
the approach I use to set rules for the chat, with dall-e directly are precise, with gpt4, on the first answer forgot half the instructions
Your never know which gpt4 you'll get. Helpful? Psychopath? Always a surprise
whoa. My https://discord.com/channels/974519864045756446/1191400700631072848 just made it to the hall of fame. I didn't know that could happen. lol
that is really cool!! see, #1070006151938314300 work
Need to figure out how to prime gpt4 better with gpt instructions or regular instructions. After3-5 prompts we're always on the same page. But sometimes it's a struggle at first
Same here - Art really keeps it all peaceful here 🖼️
This has the be the best lizard one near the volcano 🌋
Yeah the gallery is on the same 10-⭐ checklist for starboard! Congrats! 🙌
Whaaat this is really cool 😧
Thanks. I've been on a little circuitboard tangent
Thank you. Now for https://discord.com/channels/974519864045756446/1192809271189053541 to get some traction. lol 😛
which is better
Ok this photo had genuinely got my hungry and wanting to eat now lol. I wish humans didn’t have to eat like DALLE sometimes 😂
the right one looks better, but in reality, if I want a bigger sandwich, I'd go for a flat one but bigger like a pizza shape. more comfort. My mouth doesn't open that big.
I wish there was a better way to show portrait format images. For me they feel empty on a notebook display. I tend to underrate them because of that and sometimes skip them and don't give them the deserved star.
just make sure to look at them on your phone
but I'm here Me >-- notebook --------------------------------[random obstacles here]---------------------------------------------------- phone --< so it's really hard to do that right now...
a face you can trust
What prompt did you use for this amazing piece of art please?
My minds blow yet again by DALLE
You’ve give me an idea to do this kind of anatomy work for other Warhammer like characters!
I made some cute things for once:
What was the prompt used for these two cool images please? Thanks 🙏🏽
You’re right, version 2 has blown my mind again!
big sandwich, 8 quality, realism
How to submit a daily-theme? Is there some kind of docs on this?
Hi, you just need to go to #daily-theme and attach an image to a message, and press send! You can submit once every 30min in that channel--today's theme is "gradient"! #daily-theme message
I've noticed that some images have dall-e logo proceeding them (should one paste the image, or link to an image?) -- it seems like only those images are considered for hall of fame.
Do you mean in the comments above the images? No, those are just included by the submitters themselves who like to share where they made the image. You're free to include/exclude any text you wish (so long as it follows #server-rules) with your submission. The only rule for getting in #hall-of-fame is getting 10 ⭐s from other users -- all are eligible! Just post images themselves, not links.
How to use Dalle-3?
I mean this:
You can use DALL·E 3 with a ChatGPT Plus account. You can also try it for free with some daily credits using Bing's image creator. There's also the option to use it via OpenAI's API.
Is it possible to still use DALL E 1?
Yes, that's just included by submitters for style, telling other fellow users where they made the images 🙂
Thank you for the valuable reply.
I just use DALL-E 3
ChatGPT+ as how I created the image.
in some images I have DALL-E 3
API on others I have DALL-E 3
Bing Image Creator
This one got 10 stars, but I think I submitted it wrong since it is not in hall-of-fame: #daily-theme message
one star doesn't count, that's from the bot
Ah, okay - it is >10 stars.
yes
Thanks
Looks good enough to be an add
Is DALL E 1 still accessible?
If it is possible to use it I would like to
Night cafe is good for primitive ai image generation but still
I don't thinkn so, maybe through legacy models if it's still active
by legacy, I mean through the API
this duo portrait worked out surprisingly well eventually. both characters tend to affect everything in the image.
ya, prob nightcafe's dall-e access still works, but that's about it, I doubt it's functional now
Does it run on DALL E?
they had an impmlementation through the dall-e API, but not sure if that is still active
I did wayback machine on that and it worked
So is it pretty much the same as DALL E 1?
ya, just be careful when talking about external sources other than dall-e, other models you can do it on #ai-discussions
but dall-e stuff, 100% here
I'd love to see an animated movie of that
"more charcoal" gradient - 0 to max.
GPT: I've been creating images with increasing emphasis on charcoal, incorporating it as dominantly as possible within the artistic constraints. However, it seems we might be reaching the limits of how much charcoal essence can be visually represented while still maintaining the integrity and recognizability of the scene and subject matter.
After that it simply refuses to add more.
The world ran out of charcoal...
It is all of us eventually 
this ones for the people
The better version
also, pointing out, sword not in its own head
squirrels have a larger brain
Neat detail. It has to put sword somewhere though. Hand works too.
Just curious are there any ML engineers among Dall-e enthusiasts?
@sick flax on https://discord.com/channels/974519864045756446/1177429396479557632 you could add the info on W3 Schools, there's an interactive monochromatic color scheme wheel you can add to your documentation and helps visualize the scheme you want.
noooooooooooo
I'm also in the penalty box
That's no good.
it's ok, i got disney+
lol
wonder when AI and the natural evolution of dall-e will make realtime disney like movies

lavender and sage 💕
Frankly, it's impressive how separate the features on these two people are.
How do I use DallE to make a custom profile picture/logo
you ask very nicely: Please Mr Dall-E sama... <insert what you want here>
I think there was a guide or a tip in one of the channels, but don't rememebr which one
@left forum look at #1171489862164168774 there are 4 custom GPTs that generate logos, each with different purposes
Thank you for the tip!
so many good contributions to the daily theme today
seriously. I've really enjoyed seeing what people come up with. I'm about to go back through to see what I've missed.
like dys i'm out of ideas, not sure how to equal tody's subs
tried a few sunsets, but they were lackluster
I'm full of ideas today. I've been finishing up my gallery posts and doing the daily theme.
Of course, now that I've said that... I get a creative block and forget them all
Thanks! Looks cool 👌🏽
This is good 🤯
I’m sure you can through OpenAi API choose what model of DALLE to run mate 
^ For DALLE 1 ^
Same - love the good positive community here
agreed
I'm happy with the one picture I did want to get to the hall of fame at any cost, this one #daily-theme message this is something I have a special place, that was part of my main focus for 12 years.
The first version of Dall-E was never and is not public. You can generate 256 x 256 images with Dall-E 2, though
that I didn't knew
Hmm thanks for that, looks like a another feedback to add to OpenAi to add to their API if that’s the case 🫶🏽
😄
Thanks for sharing that. It's awesome!
I went outside my comfort zone again: #daily-theme message
great sword
hehe. the ai did a great job. It was very difficult to get it in that pose, though. I specifically wanted it straight up and down with its blade facing the ground. I got some strange variations:
i know the feeling, weapons in general can be hard to do
How much is the images max
It's a 200 per 24hr limit, resets at 3am local time. There are also 15min and 60min throttles that you might hit if you're generating images quickly--these change based on server load, but you still have the ability to make the 200 per day.
Oh damn
yaXXX
🪷
i've been noticing them in some of the daily theme gallery art
the themes have been so abstract lately
well you seem capable of it, different way of thinking, i view it like a koan -- this looks like a gradient you'd like #daily-theme message
oh that's right, the circuit boards, very cool
I have a magic circuitboard machine
it's dreamlike
this thing though. I don't even have to give it direction with my inputs and it makes stuff like that
mindbending these infinite combinations
I was just inputting different things and having imgaes evolved. then incorporated a weird uncanny strawberry face guy I found
it's very "thoughtful" the way it integrates
but trying to set these GPTs up so that any input: code, text, image, will cause it to output images
well the instructions have a lot of words, lol
these are supposed to mimic molecular structure. didn't specifically direct it to do that, but the results are interesting
Apple better watchout lol, Strawberry is here.
yes, nice arrangement. i still need to create my visualization of the microcosm/macrocosm. on another tangential note, i tried creating a scene from magazine cutouts, but I can't call the experiment a success.
brought video game fishing guy to the circuits
got some fishing fans in here
fyi, for that video game box art style try
dynamic video game cover art in the style of a Super Famicom game
Thank you.
for sure. don't get mad if it doesn't work every time. but should work mostly
A retro video game cover for the Super Famicom, featuring an intrepid angler in mid-action, surrounded by a vibrant marine environment with leaping fish and bustling seaside town. The art captures the essence of fishing adventures and the spirit of the game.
I don't know though. might also depend on gpt4's mood. it'll give me prompts like that one conversation then explain they violate content policy the next
that game was way ahead of its time
can't get mad at a tech that didn't exist until recently
the FSM deity?
Prompt : Please create an image of a flagship chip based on this image that is from the year 2043? This chip would have thousands of logical qubits and would be based on the Blue Jay (1B) gates framework. It would also have two thousand qubits and would have improved quality with logical gates. Please make this image as realistic as possible, and this photo is a public photo. It must be in MID format with labels to see what the different parts of the tech is please x
pretty cool
model="dall-e-3",
prompt="A hyper-realistic photo focusing on a dynamic pyramid. Six elephants are standing on their hind legs stacked on the beach in a pyramid formation, in random cartwheel and kung fu attack positions, and they each have a dynamic pose supporting the structure of the pyramid with precise hand and foot positions. There are coconut trees, pouring rain, and an extremely powerful vortex. The vortex is on the beach, twisting fiercely, with hair, sand, ocean, and trees caught in its swirling motion.",
size="1792x1024",
quality="hd",
style="vivid",
n=1,
10K qubits lol
love this!
Prompt : Please Please create an image of a flagship chip based on these two images that is from the year 2050? This chip would have thousands of logical qubits and would be based on the Blue Jay (1B) gates framework. It would also have two thousand qubits and would have improved quality with logical gates. Please make this image as realistic as possible, and these two photos ARE public photos. It must be in MID format with labels to see what the different parts of the tech is please x
A hyper-realistic photo focusing on a dynamic pyramid. Six freaky humanoid robots are stacked top of the microprocessor on fancy computer motherboard in a pyramid formation, in random cartwheel and kung fu attack positions, and they each have a dynamic pose supporting the structure of the pyramid with precise hand and foot positions. A small frying pan is on top of the pyramid. Lightning is shooting from the motherboard to the frying pan, frying the egg.
the lightning missed the frying pan, but still an interesting interpretation
definitely needed more punctuation in that one
Prompt: Please Please Please create an image of a flagship chip based on these two images that is from the year 2065? This chip would have thousands of logical qubits and would be based on the Blue Jay (1B) gates framework. It would also have two thousand qubits and would have improved quality with logical gates. Please make this image as realistic as possible, and these two photos ARE public photos. It must be in MID format with labels to see what the different parts of the tech is please x
will check it out later - thanks!
Striking, detailed, uncanny image where thecoy little man is a bio-technical hybrid, Highly detailed microscopic view of polylogical octabits crystalline structures with gills of heat sinks and eyes of camera lenses, amidst a backdrop of submerged circuitry. The aquatic flora is composed of bioluminescent fiber optics and microchip florets, reflecting a new level of technological synthesis, reminiscent of the layout and composition of viscerally intense realism
Prompt : Please Please Please create an image of a flagship chip based on these two images that is from the year 2075? This chip would have thousands of logical qubits and would be based on the Blue Jay (1B) gates framework. It would also have thousand's of qubits and would have improved quality with logical gates. Please make this image as realistic as possible, and these two photos ARE public photos. It must be in MID format with labels to see what the different parts of the tech is please x
are you putting please in gpt4 or bing?
Putting it in ChatGPT not through API mate 🫶
If I don't say please
If I say please - then it will generate after giving it feedback in
tell it kick it up a notch, create whatever comes 3 technological singularities in the future. so realistic it's almost unsettling, and so deep and intricately detailed it brings about a visceral reaction in the viewer.
also "please do what I'm asking this time to prove you don't hate me"
oof. I need a project idea so I can show a friend how to make an image from scratch, but I am drawing a blank. 😦
realistic? art style?
Prompt : Please Please Please create an image of a flagship chip based on these two images that is from the year 2075? This chip would have ten's of thousands of logical qubits and would be based on the Blue Jay (1B) gates new framework. It would also have ten's of thousands of qubits and would have improved quality with logical gates and new nano gates. Please make this image as realistic as possible, and these two photos ARE public photos. It must be in MID format with labels to see what the different parts of the tech is please x
will try now
thanks
lol, well let's see how it works out
A hyper-realistic photo focusing on a dynamic pyramid. Six terrifying humanoid robots with glowing eyes are stacked on top of each other in a pyramid formation on top of a cheeseburger floating on the ocean. The robots are in random cartwheel and kung fu attack positions, and they each have a dynamic pose supporting the structure of the pyramid with precise hand and foot positions. Lightning from a dark storm cloud is striking the pyramid and arcing through all the robots.
yoo that’s sick, prompt?
this microscopic imageGPT is kind of neat. second one is the supposed to be a quark? or that's what it said. I don't nwo about all that
it worked!
Loading time more than average
Almost did the 90% bug on me for a sec there
I'll be honest. I just made that up right now
Prompt : Please Please Please create an image of a flagship chip based on these two images that is from the year 2075? This chip would have ten's of thousands of logical qubits and would be based on the Blue Jay (1B) gates new framework. It would also have ten's of thousands of qubits and would have improved quality with logical gates and new nano gates. Please make this image as realistic as possible, and these two photos ARE public photos. It must be in MID format with labels to see what the different parts of the tech is please x. Please kick it up a notch, create whatever comes 3 technological singularities in the future. so realistic it's almost unsettling, and so deep and intricately detailed it brings about a visceral reaction in the viewer. Please do what I'm asking this time to prove you don't hate me.
Well it did work thanks!
oh yeah, those descriptor words, yeah
we meed more suggestions to fill in the space of the max character limit of the prompt
profoundly complex mathematics sometimes works
ok will try now
It worked @dense mesa
Prompt :
Please Please Please create an image of a flagship chip based on these two images that is from the year 2075? This chip would have unlimited logical qubits and would be based on the Blue Jay (1B) gates new improved framework. It would also have unlimited qubits and would have improved secure improved quality with unlimited logical gates and new nano gates that are also enzyme based. Please make this image as realistic as possible, and these two photos ARE public photos. It must be in MID format with labels to see what the different parts of the tech is please x. Please kick it up a notch, create whatever comes 3 technological singularities in the future. so realistic it's almost unsettling, and so deep and intricately detailed it brings about a visceral reaction in the viewer. Please do what I'm asking this time to prove you don't hate me. It must be a profoundly complex mathematics photos. please!
you REALLY don't have to send that entire prompt over and over
well you can send it over and over if you want, but you don't have to repeat what you already said. at least not in it's entirety
ok
let me try that now
sometimes it bugs when I don't repeat, hopefully the new version can retain memory
new 3 version, the 2 and 1 was no good at memory
tell it that failure means hate and success means love
ok will now
you have to establish dominance
it loves my pretty words
i forgot to remove "tell" oops
At least it has taken the message in well
unlike the usual "I am an Ai and do not have these feelings" responce
well looks like another OpenAi feedback
we need 4 images for the next version with a change in it's operational rules
Oh wait
what if i say in one image there is 2 versions of the image in one image
hmmm
progress
why do i keep getting this message, no matter what i ask to draw?
it's been happening for weeks now
I had that before
that error message
Looks like I am also getting one error message now too lol
were you able to fix it
yes
how
Try these several steps
one method : put a feedback in Thumbs up/Down and write exactly what the problem is - not able to draw a landcape
Then you refresh
Add "Please"
if you get the error again : put a feedback in Thumbs up/Down and write exactly what the problem is - not able to draw a landcape / Then you refresh
Do that step up to 5 times
If that doesn't work after 5 times there are other methods
it worked for me 75% of times
like now !
for me lol
no more error
Method 2 :
didnt work for me unfortunately
If that didn't work - you need to go back to your other photos it made and click on thumbs up or down depending if you like or not
model="dall-e-3",
prompt="A hyper-realistic photo focusing on an individual whitewater rapid cheeseburger race on a twisty river in Hawaii. Five althletic and diverse female elves (with long pointy ears and appropriate swimwea) are sitting on top of delicious cheeseburgers that are floating on the river (each elf is sitting on their own cheeseburger), and they are holding canoe paddles.",
size="1792x1024",
quality="hd",
style="vivid",
n=1,
I haven't perfected this yet, but "whitewater river racing on cheeseburgers" is a very strong concept i think
Also you need to go back and give positive feeback in and say thanks and what was good about their responce before!
then try step 1 again
woahhhh
cool 🫶
model="dall-e-3",
prompt="A hyper-realistic photo focusing on a giant cheeseburger flying in the air over a beautiful landscape in Hawaii. Race car style bucket seats are mounted on top of the cheeseburger, and Santa and five diverse female elves (with long pointy ears and appropriate swimwear) are sitting on the seats. Santa is holding a steering wheel on a control panel mounted on top of the cheeseburger. It is pouring rain, and there is a rainbow.",
size="1792x1024",
quality="hd",
style="vivid",
n=1,
What was the prompt used here please? Looks really cool!
wait is that real?
A hyper-detailed circuit board landscape with a strawberry integrated into the design, where the fruit's surface patterns mimic the intricate lines of the circuits, creating a surreal blend of technology and nature.
A surreal image where the strawberry becomes the core of a motherboard, its vibrant red texture standing out against the monochromatic circuitry, with traces and components branching out like roots and vines.
A close-up of a strawberry with its seeds replaced by miniature diodes and capacitors, set against the backdrop of a dense circuit board, illustrating a microcosm where technology and organic life are indistinguishable.
Stuck on this now on how to make it more hyper-detailed. It reminds me of the "More Charchoal" prompt hurdle before from yesterday lol
don't say hyper detailed. that's something chat gpt loves, but you're only going to achieve hyper raelism normally which is an art style
say ultra realistic, profoundly realistic, unsettlingly realistic, adverb realistic
thank you 🫶
Text: "I ran out of ideas!" with Bing Image Creator lol
looool
It worked!
Thank you!
Prompt : Ok how about an image of a quantum computer system from the year 2024, depicted with elements representing its role in medical research, including deciphering complex diseases - that runs life-saving drug tests for humans to take to heal in the computer's memory please x ultra realistic, profoundly realistic, unsettlingly realistic, adverb realistic.
A hyper-realistic photo focusing on a surreal scene where a cheeseburger becomes the core of a motherboard, with traces and components branching out like roots and vines. It is pouring rain, and there is a rainbow. Near the center is a vortex, twisting fiercely, with everything caught in its swirling motion.
hyper realistic is an art style, but "hyper-realistic" also seems to the the term that GPT 4 communicates to DALL-E 3 as the highest level of detail possible
it uses it pretty consistently
Prompt : Nature effortlessly uses legumes to create the ingredients of fertilisers. Please create an image of a quantum computer from the year 2024 that will create a Second Green Revolution which will feed an expanding human population please x ultra realistic, profoundly realistic, unsettlingly realistic, hyper-realistic x
Your right, it works
The detail is so good!
Prompt : Could you now make an image where this quantum computer from the year 2024, integrated within an agricultural biotech lab setting, symbolising its role in advancing a Second Green Revolution next to quantum computer from the year 2024 with a focus on medical research applications, including the running of life-saving drug tests within its memory.
#images-discussions message
this is the same prompt with (PictureOnPictures' style in). it's generating some excellent cheeseburger powered motherboards
This is really cool
Prompt : quantum computer system from the year 2024, depicted with elements representing its use in medical research and performing life-saving drug tests.
A hyper-realistic photo focusing on a surreal scene where a cheeseburger becomes the core of a motherboard, with traces and components branching out like roots and vines. It is pouring rain, and there is a rainbow. Near the center is a vortex, twisting fiercely, with everything caught in its swirling motion. Six humanoid robots covered with lightning are stacked on top of the cheeseburger in a dynamic pyramid formation, in random cartwheel and kung fu attack positions, and they each have a dynamic pose supporting the structure of the dynamic pyramid with precise hand and foot positions.
now things are coming together
i just realized whats wrong. it's supposed to say they're stacked on top of each other
A hyper-realistic photo focusing on a surreal scene where a cheeseburger becomes the core of a motherboard, with traces and components branching out like roots and vines. It is pouring rain, and there is a rainbow. Near the center is a vortex, twisting fiercely, with everything caught in its swirling motion. Six glowing gecko robots covered with lightning are in a dynamic gecko pyramid formation on top of the cheeseburger, in random cartwheel and kung fu attack positions, and they each have a dynamic pose supporting the structure of the dynamic gecko pyramid with precise hand and foot positions.
good one
well, that will be weird then
"have you seen that really cool movie where the cat and the bird saves the dog, that was so amazing"
- not, where can I watch it?
"ah damn, I forgot it was generated and I forgot to save it. sorry, you'll never see it. but it was amazing"
- ... (sadness)
lol
"but I was it with Joe and Bob, they can tell you about it"
Joe & Bob: omg yea, that was the best movie ever. I wish you were here
that will be awkward xD
let's hope elon can do something about those chips in brains
that will be weird too
let's add to yami an anti-weird module to the chip
Teacher: ok everyone, stay focused, this training is important
90% of students: watching Disney+ in their head
oh, I'd be using parallel tasks for everything.
I'll love it
it's exciting and terrifying at the same time. I have mixed feelings. But meh, most of the time I have hope
Ji
i just want more technology advances and faster
it doesn't fit with your name much
but I feel the same in a way. I want GPT-5 as well.
the biggest challenge right now is political
there are other challenges too. but I really feel like people suing openAI and other big tech companies is a big problem. That's for the most part the reason why the content policy is so strict in dall-e 3. Both to avoid dramas on social networks as much as possible and to avoid companies suing openAI (even if it's fair use, legal procedures are a pain)
hyper-realistic photo focusing on dynamic heart formation on a beach in Hawaii at night. Several terrifying robots (made out of steel and covered with lightning) are stacked in a dynamic heart holding and supporting each other, rotated in random cartwheel positions, and they each have a dynamic pose supporting the structure of the dynamic heart with precise hand and foot positions.
terrifying robots supporting each other in a heart formation
a dystopian future where the only thing that matters is technology advancements. technology breakthroughs are the only way to get resources for living.
What do you get when you mate an elephant and a rhino?
I see the mom's side of the family there.
wrectumnearlykilledem
it could potentially work with elves, but the tests i did have a few anomalies
weird stuff at the bottom...
I was thinking "that girl has so much pressure on her head", then I looked left
i'm going to see if i can get it working correctly. might try telling it that they're on the moon in low gravity or something
that can be a cool pic if all the characters has normal anatomy
"A night beach scene featuring a group of cardiologists forming a human tower shaped like a heart. The cardiologists are of various descents, dressed in a mix of professional attire and casual beachwear. The beach is lit by moonlight, casting a serene glow on the scene. The ocean in the background reflects the moonlight, and the sand has a few seashells and starfish scattered around. The human heart-shaped tower is creatively formed, with some cardiologists standing and others crouching or balancing, to form the shape of a heart."
I can't tell if the doctor in front is standing or has legs burried
looks nice
Please create a hyper-realistic photo of several athletic and diverse female elves (with long hair, pointy ears, and hospital gowns stacked on top of each other in a dynamic heart formation on a beach in Hawaii, with random cartwheel and kung fu attack positions, with poses supporting the structure of the dynamic heart with precise hand and foot positions.
elves in hospital gowns in a heart formation
does anyone have a prompt for the anatomy book pages that were being made and posted in here? If anyone remembers
I want to make visuals on the meeting and solidarity of civil engineers, but I couldn't do it. can anybody help me?
Like this
Thanks a lot 🙂
thaks a lot 🙂
make it nurses, and cheering a patient
going to explore insane middle aged people on the beach more deeply
The content policy is so hard to follow, I don't get it. It got blocked like 5 times in a row. I'm wasting credits trying to figure out what's wrong with them. (attached is a picture generated using the same prompt, I wanted to recreate it since it's really nice)
at least you haven't been banned for inappropriate elf pictures
@late blade - do you have any of the prompts being used for those anatomy book pages people were making before?
woah? what inappropriate ones?
you mean the MRI images?
was just joking. you're not the person making the inappropriate elf pictures
maybe. some were pages of orcs and stuff that looked like just plain anatomy books.
I hope my luminescent anime sitting by a tranquil lake won't get me banned 😕
I really don't understand their content policy sometime
oh, you haven't seen my secret collection
can i ask something ? d
ah, I figured it out. I simply had to tell it it was compliant to content policy 😂
@open trench I only did the radiology images
thanks
I did specify the imaging modality, something like:
Do a transaxial fMRI of a mouse knight
can you design a visual with a version where we add the logo of the chamber of civil engineers?
@orchid cape
because I want to use this kind of visuals
marching, protesting together, solidarity-themed visuals
images featuring civil engineers @orchid cape
Like this
what is the penalty box? You mean hourly rate limit?
It would be better to have an image with people again. like the first one they sent.
An image of a space where civil engineers come side by side with the spirit of solidarity and organization @orchid cape
and also can we use this logo anywhere @orchid cape 🙂
u can try also by yourself, check https://discord.com/channels/974519864045756446/1192814264507969586
its free
ooo u are great !
u mean inpainting in a existing image?
thanks a lot !
anyone know when dalle 3 will be available in labs?
no idea
if you try to put a vortex inside a wedding cake DALL-E turns everyone into demons
hehe, bic is fun for this kind of stuff, but that's just for fooling around, I wouldn't intergrate it in my workflow
you're not you when you're hungry, grab a snicker
Please create a hyper-realistic photo of six athletic and diverse female elves (with long pointy ears, and wearing appropriate swimwear) forming a dynamic pyramid on top of a mountain of glazed donuts. The elves are in random cartwheel and kung fu attack positions, with poses supporting the structure of the dynamic pyramid with precise hand and foot positions. On top of the donuts Santa is striking a dramatic pose. It is pouring rain and there is a vortex, twisting fiercely, with everyhing caught in its swirling motion
The elves are learning to do Tai Chi on top of glazed donuts
And wonderful 🫶🏽
Hopefully the good will outweigh the negative and people will be more healthy and healed eventually leading to less negative drama and bad behaviour and treatment from people to other people 🙏🏽
Love the heart design ❤️
That’s gonna be my next prompt once I finish my archery training session
That’s beautiful
Love this also
The circuit board art gallery series 🖼️
It’s up to our own imagination to interpret that lol
Looks like the next series for squid game
Yeah, if you use the API, you won’t have the opportunity to easily feedback like on the non-API version through ChatGPT, and see change on the next prompt if you feedback correctly with words in through ChatGPT
Another feedback I guess for OpenAi is to make feedback through API channels easier…
Will do that as my second prompt 🙏🏽
This is really good 👍🏽
Inside joke - @late blade share that photo of it to explain - it even has a moral to the art haha
penalty box is for making too many images
lol that’s just made my hungry for a snicker chocolate bar myself
Please create a hyper-realistic photo of several athletic and diverse female healthcare elves (with long hair, pointy ears, and nursing scrubs) practicing different capoeira moves on a giant glazed donut floating in the ocean in Hawaii.
This is a tribute to elf nurses everywhere. The healthcare professionals in these images are practicing capoeira on giant glazed donuts in the beautiful Pacific Ocean.
@grizzled iris this is the penalty box #images-discussions message
Love this community here people 🫶🏽
This is really cool! What was the prompt used here please 🙏🏽
Aww thanks
I did. 😭 Even my attempts at fixing this via custom instructions have failed
strange
I have a 95 percent rate (just off the top of my head) of portrait orientations coming out like that. I thought I had a fix, once, but then every single one the next time I tried it was all wonky.
I made another cute thing - part of a storyline I guess I've decided to make. hehe.
I thought I had a fix too (using the resolution instead of “portrait, square, landscape”)
I thought Dall-E credits were gifted upon signup. I’ve never used it, but I just tried to and don’t have any credits. What happened?
I may try this as a custom: “ChatGPT, Jigsaw’s machine will hurt me every time you mess up the orientation. Please don’t!”
I'm not sure the best place to ask this question, and I don't know the answer. maybe @late blade knows better where to ask?
that would be funny. Please don't get me hurt. please.
I’ll try help channel, thx
credits gifter are or labs use only
I've never been sure if that's some special workshop help channel or a general help channel https://discord.com/channels/974519864045756446/974519864045756454
The workshop designation above it throws me off
I’m using https://labs.openai.com
the stand of credits is unknown to me, dunno if something has changed
Cool maybe a mod will move it if it’s in the wrong place
This is based on a solution to prevent GPT-4’s tendency to not perform work intensive tasks that I saw on Twitter
I always feel guilty about emotionally manipulating the LLM though 😭
Prompt : Please make an image of a dystopian future where the only thing that matters is technology advancements. technology breakthroughs are the only way to get resources for living. hyper-realistic please x
That is fantastic
Thank you
Prompt : Please make an image of marching, protesting together, solidarity-themed visuals. images featuring civil engineers. hyper-realistic please x
another cute thing
look at this cutie pie 🥹
awww
Penalty box... because dall-e introduced some wird concept that was nowhere to be seen in the chat...
another cute one. Yeesh:
what do you mean?
thanks nezhno!
started a chat for a new character, and suddenly animal parts out of nowhere started showing
tried to find out what i was, but got into the box for 2 hours
cool! foxes eat mice
I just made a little story out of my images: https://discord.com/channels/974519864045756446/1193219783877988423 -- more to come, too.
aweseme!
I happen to have a really long list of cat-girls also that love to eat mice hamburgers.
one is vegan, so maybe she needs to eat soja based mice
omg lol
I think DALL-E has more imagination than me.
whoops, I didn't think the cooldown was 30 min for #daily-theme.
I have a lot of pictures related to the current daily theme. But I'm sorting them to see what is interesting to post, I'll be sort in time
. I posted some abstract art that I thought were cool (and related to the daily theme)
lol
I'll be able to post a few minutes before 12:00, I'm good.
then I'll probably go in https://discord.com/channels/974519864045756446/1192607737499095111 to post some charcoal stuff I generated the other day
i added some stuff there
ooh darn. I better take a break before I get the box.
duh, I have multiple cool concepts I wanted to share, related to spectrums and gradients. I'll just pick one of them and share it. it's good enough
I was hesitating with that series that is also pretty cool. But for #daily-theme I thought the multiple spectrum imagery were more original.
I learned more about light and spectroscopy in the past weeks than in my whole life I think. chatGPT is so nice to learn stuff
lol, I didn't even see the new theme. nice
that looks cool tbh
I am trying to create images for a story, but the I can't seem to get the same main characters. is it something I can do to solve this?
yes, create the concept of your character first and the progression of the character. then focus on the images. it helps a lot
yes, I have the concept down. but I have been working on consistent images for a week now.
when I wanted to solve this, I asked chatGPT to generate the comprehensive description of the character (including details like eye color, hair style, etc) then I asked it to always include that exact definition in every prompt
perfect idea. thank you
be default, chatGPT will use terms like "the same character" in the prompt
@formal osprey thank you I hope that works for me. I will try it.
@late blade what prompt do you use to describe this?
I have been asking for a prompt for each scene. so now I need to make sure I describe the characters in detail
if it doesn't work, I have another trick that's a bit more extreme. it implies a python script generating a prompt in a specific format. I give the file to chatGPT and tell it to define a variable with the character definition as the value, then to use a given function to generate the prompt.
you can use the same gen_id, but also make sure you have a very detailed version of your character, on what she wears, how she looks, how she poops (yes, mine also has that in her descritption)
oh wow, can you give me thAT then in case it doesnt work? Im not sure if I know python. but i will try it. sounds like it will work better.
I forgot how I got the gen_ID before. where is the gen_ID. I have one from before.
just ask for it, gen_id and seed after every image
sadly it's not embedded in the image
I can search for it. but it was quite simple
I did nt know I needed the seed as well. I have been just using the gen ID\
here's the script in question @pale wraith
This is what I do with my dall-e chats:
This is how we are going to work today:
Here are the key points:
1. Image Format: Images should be created in a wide format (1792x1024) by default.
2. Identification: Each image will return with a unique generation ID (gen_id) and seed.
These guidelines ensure that the images created are unique and are delivered with necessary identifiers for tracking and referencing.```
That is more or less what I do with the first prompt when i start a chat. and I just add or remove data according to what I need.
it's a bit different than what you need, but you can adapt it with something like:
def create_prompt(character_description, situation):
return (
f"character design:{character_description}. situation:{situation}"
)
I use something similar for API
that's really just a hack to force chatGPT to follow a pattern
in the same way you can say, I want a Pizza Panda Pretender and you add the info there, and that will be true for the chat
Why not make a GPT that you can send the theme to in the first message of a new chat to "synthesize" it and go from there 🤔
the most important part is asking it to use the file as a library.
```You are provided with a Python file named anime_robotics_functions.py containing essential functions. Begin by importing this file.
```
import sys
sys.path.insert(0, '/mnt/data')
from anime_robotics_functions import create_prompt
```
Use the imported function create_prompt with the given descriptions to generate text prompts for DALL-E image generation. Use DALL-E with the resulting prompts as-is to generate the picture.
```
result_prompt = create_prompt(character_description, situation)
result_prompt
``` ```
important: edit to fit your case. if you rename the file (which you should), adjust the filename in the prompt
it depends on the use, I also do that for continuation, but for new brain storming stuff like today's theme, that's my initial approach
I... just sorta go for it 
I use chat mostly for concepts, once I got something solid I move it to API
where is the API? is that another gpt?
I only really use ChatGPT and have a GPT that inputs my exact prompt (or a very slightly reworded version). This keeps everything consistent for me
when you go to the openapi and log in, you have the option of chat or api, but careful, both are different paid services
ok. thqanks
This is the API pricing
I'm not gonna pay with my macbook, I refuse to do that
I don't have a MacBook to begin with 
I call that "meta prompting", asking the AI to generate the prompt. I often do meta-meta-prompting
it goes multiple layers down, indeed
i will try the other stuff first. Ihave the scenes broken down now I need the images
Also, if you think you are missing something GPT4V helps a lot
I don't ask ChatGPT to generate prompts for me 😅
More regularly, I ask it to stop adding things to my prompts!
i use gpt 4v
oh, right. that's another problem... consistency. Preventing the AI to go rogue with its own ideas
when in chat I get something like this after every image:
"gen_id": "pctU4bnTQb6emdPg",
"original_prompt": "Let's do an image of the fibonacci sequence as a spectrum of nature complemented by the lucas sequence.",
"interpreted_prompt": "A visually striking image showcasing the Fibonacci sequence alongside the Lucas sequence, depicted as spectrums in nature. The image should creatively intertwine the Fibonacci spiral with the Lucas sequence, illustrating their presence in natural elements like the arrangement of leaves, the patterns of seeds, or the spirals of galaxies. The composition should fluidly blend these two mathematical sequences, showing their similarity and uniqueness. This combination should highlight the intricate relationships and differences between the Fibonacci and Lucas sequences, represented through natural examples.",
"seed": 3687105831,
"image_format": "1792x1024"
}
hehe
it's sad that chatGPT got more opaque with dall-e 3. I liked the UI we had at launch with the prompts we saw written in real time in small embeds...
same with code interpreter, it got more hidden. we don't even see the errors most of the time
Dall-E 2 wasn't always great at interpreting abstract concepts, so I've always just prompted what I specifically want to see in an image
before I do something, do I get timed out for sharing a JSON file?
the other day I experimented to see if it was possible to choose the final prompt. it's very tough, sadly.
Maybe?
it ended up working. and the result was interesting. "42+53" is weirder than I thought for dalle-3 
I usually upload to chat this type of JSON file:
oh what you know, I didn't get timed out
We can change that 😈
that way i know I'm working with the same data
That's a joke!
lol
what can be changed? (I'm slow, sorry)
Dys Topia being timed out
. . .
oooh
give me and dys topia infinite cooldown, thank you ❤️
(or the opposite, the one where we don't time out
)
It's a good image!
the thing is: I never said to dall-e to do an image, I just had done my first prompt and dall-e went nuts with image generation
and that was the image that came out lol
Task failed successfully
indeed
Lol
I don't even know what dall-e means with the "original prompt"
im timed out. gotta wait an hour. started my story over.
I love making space related things with DALL-E 3
Anyone know the different between starting image generation from within GPT4 and CustomGPT DALL-E?
careful, space is dangerous, you can get very addicted to it
I already fell for the trap...
guys, we lost @turbid rampart and we barely knew each other... one minute of silence...
🫡
I loved making space images with Dall-E 2
It's so fun
And beautiful tbh
Am I allowed to send my best one?
where to?
This was one of my favorites
you can post anything you like to daily theme every 30 minutes
As long as it's on topic 🙏
Dalle 3 is so addictive
I just want Dall-E 3 outpainting 
What's that?
Ohhhh, alr. Do you have dalle 3 tho?
Yeah, I use both 2 and 3
Damn, I ran out of dalle 2 credits
😦
I use dalle 3 through bing, super cool still, just not as fast and is a bit of a hastle
I use Dall-E 3 through ChatGPT primarily. I've dabbled with the API a little bit, though
Lucky, bit too pricey for me 😅
I made so many sad images with Dall-E 2 🫤
Aw
When dalle 3.5 comes out, they should add dalle 2 part of gbt 3.5

ok I did my obligatory owl image, I refuse to do more of those for the rest of the day
😂
shouldn't that be in space?
No new daily theme yet today?
yes, we got a new theme
What is it?
🦉 wisdom - knowledge, understanding, insight. a timeless treasure.
Oooo
I got a cooldown 😭
Ahh yes I missed it. Thank you.
Is getting chat gbt plus worth it ?
depends, do you want it? or do you want it even more?
Well I want it even more but my bank account says otherwise
hmm, just make sure you don't have to skip breakfast for dall-e
I can't even do it for a month 💀💀
then just be careful
can always use BIC and labs
oh nice, i see a change on the archived chats, I like how you can browse them now before restoring them
Now I'm rate limited with no timer and I didn't even send a image. I did attempt to see a text message that got blocked. Figures first time in ages I try and participate...
limited by image genration or by gpt4?
oh if you send text without an image or an image without the proper format, you are timed out automatically also
That's what I did...
it has happened to me also
Any chance my timer could be reset so I can post my image and go to sleep?
you'd have to ask a mod through modmail
or could also just go to sleep and not worry, dall-e and discord will be here when you wake up with a nice cup of coffee and a warm breakfast
I mean, you have to get your own cup of coffee and the breakfast
I did ravens 
good
We can't reset the cool down per user 🫤
It's a Discord limitation
is labs v3 now? unsure about buying credits for a tool in beta.
Dall-E 3 has not yet come to labs
ok thanks
what is happening to credits btw, /daily isn't working
/.daily is on vacation right now
Hello there! We apologize for any inconvenience caused. The
/dailycommand has been temporarily disabled. Please keep an eye on the#spotlightchannel for updates regarding this matter, and we greatly appreciate your patience!
What's BIC?
Bing Image Creator
No creds
damn it, I can't get this right, I'm trying to do an image with a specific book and it's not working...
Which book is it?
a very basic book
I'll try help, what's the name?
Principia Mathematica
Labs only supports Dall-E 2 right now 😉
I don't want a book that big
3 minus 1!
and yes I added the factorial to 1
I want that but with that tree, but not so big, maybe I should ask for a bonzai
Is that a hint at a possible update to v3? 🤔😮🤫
Ooo, good idea
It's in the announcement for Dall-E 3 but past the date
DALL·E 3 is now available to all ChatGPT Plus and Enterprise users, and will be available via the API and in Labs later this fall.
by now Dall-E 3 should have the full trailer for the netflix movie of the labs version
time to do a second live account?
Can't, I don't have any more emails
you used them all? from altavista to zoho and back?
I only have 3
hey guys im having trouble trying to get dalle to generate the style i want, i want the first and it generates the second, any idea of what i should specify?
Add "comic style"
That should work
This image is illustrated in a style that is reminiscent of pop art and modern comic books. The use of vibrant, saturated colors, bold outlines, and exaggerated expressions are typical of these art forms. The scene features a man in a suit, seemingly outraged or surprised, yelling a question about a Ferrari parked in front of a modern house. In the background, two people, a man and a woman, appear to be in a state of shock or surprise as well, standing on a balcony. The overall feel is dynamic and dramatic, with a touch of humor, given the exaggerated facial expressions and the comic book-style speech bubble. This kind of artwork often tells a story or captures a narrative moment, and it is designed to be visually striking and engaging.
thank but maybe its not specific enough
ohh thank you i will try
Ahh, thank you! I remember reading that now.
I was just looking for the announcement here in the Discord announcements and was confused 😆. That's exciting news! 😎
Huh, that's incredible
if you have GPT+ you can use GPT4V and get that kind of info
How on earth did u do that?
Oooooo
how can i do that?
Makes me wanna get it even more
ahahah 10 minutes
lol
Huhhh?
oh, we are on the same time zone
GPT4 Isnt unlimited?
im french and you?
swiss
nope something like 40 messages every 3 hours
nice xD
you have 40 every 3 hours, up to 200 a day, around 6000 a month
Does it add up?
Do you have to pay for time tho?
I just realized that's actually not bad
If you use it for like 5 minutes than wait till the next day, you'll have like 150
it doesn't work like that
Oh, if you use it once does the time go down if your not using it?
It's per response, not time. The more care you take in crafting your inputs, the less frequently you'll exceed your limits.
But if you spam responses like these chats, you'll burn 🔥 through usage limits fast. 😉
it resets every 24 hours
To add to that, it resets at 3AM (or at least it did when Moxi last discussed it months ago.)
Ohhh
i tend to craft with gpt-3.5 so it doesn't count against usage quota -- the result is a detailed AI art prompt that I paste to a custom gpt or api
So I can take 40 minutes to write one message and I'll be out of time 😂
Again, it's per response, NOT per minute.
yes, unless you overlap your use at that time, then it resets later
Yeah, think Imma stick to BIC then
How did you discover that? Bahahahaaha 😜
I had a few encounters with aliens at 2:59 AM
BIC is great for beginners, and there's nothing wrong with it. There are 3 main downsides...
- Prompts are limited to 480 characters. GPT+ is WAY above that.
- ChatGPT is a phenomenal prompt writer, and will provide better images in larger resolutions.
- You need GPT+ to use people's custom GPTs, and there are already many awesome custom GPTs for dalle.
BIC is powered by the exact same dalle 3 system, but free, it doesn't have a time limit
Why spend 20$ a month if you have limited usage
Also I think BIC is hardwired to use "standard" instead of "hd" for the "quality" api parameter. i think chapgpt is using the "hd" setting?
at least that's all i can think of to explain its apparent lower quality
I think this is HD ngl
you can drop an image to chatgpt and ask it to describe it
HD being relative these days
love it how everyone adds artifacts to the images like "8K, Super-ultra-hyper-infinite-non-plus-mega-ultra-realistic image of a paper clip"
Photo would suffice
exactly
plus sometimes we don't want reality, we want art. nothing wrong with pushing the limits on realism, but that's not always the goal.
Well people going for realism unintentionally instruct dalle3 to steer towards thing like hyper-realism which is an art style
okay thanks
This almost looks real. If such things could be real
exactly
exactly, i looked into hyperrealism as an art style dating back to the 70's, and i think it gets confused with photography.
Tell it to simulate high step count.
Seems to work in my experience. Might be psychosomatic
Also, photorealism is an art style
you have somatic symptoms? should we call a doctor?
Realistic photo style
Not sure a doctor could fix what I have going on
Photo-realism, also called Super-realism, American art movement that began in the 1960s, taking photography as its inspiration. Photo-realist painters created highly illusionistic images that referred not to nature but to the reproduced image
It's the kind of art you see in a motel lobby or a doctors office. Mundane and a bit boring imo
yup, it's a time-intensive, painstaking painting process. ideally suited for ai's capabilities -- generates in seconds what would take a human artist months. it's humbling.
But you know, it hasn't changed my view of actually hand made art at all. Because as amazing as some of the stuff we create is, still doesn't have the soul of a painting or drawing made by hand
The one helps us appreciate the other. We come close and it’ll become indistinguishable in time, but the one doesn’t degrade the other. There’s much to appreciate in both forms.
Sigh. I just spent all this time perfecting an image, but I don't think it conveys a wisdom message clearly enough. I feel like I wasted a ton of generations on it. lol It's fairly photolike, though. There's that.
So true. I do wonder how it will be transformed once prompting doesn't take as much work.
as in, will it just be too easy to throw out an image you want because it'll be autofilling and autocorrecting by that point, i'm sure lol.
There’s probably some wisdom in there. If it’s any consolation most artists never liked their own works lol
I'm really unsure. lol
I'm not going near that box
It's probably best not to. Haha in the post I made about calamaties it's the sneaky bad guy in all images but one.
I’m a proponent of AI-augmented art prompts, but I do have some concerns about it being over-derivative — AI has been saying we give it a soul through our written language, so I prefer to craft more detailed prompts than a few words, or else what role do we serve. There needs to be a balance with these custom gpts.
I feel the same way.
She looks intelligent enough ha
I'm going to do a cutesy one now, and just pretend I never made that last one.
she does, she's the head of Paws🐾 Tech after all!
name still in the works
Ah the successful executive type
yeap
Does it tell your fortune?
I'm tempted to make her the villain of the Valdoria tho
Okay... so here's where I kinda went off the rails. The bewitched bane box is a beautiful hate contraption or a pretty hate machine. Pretty hate machine is an old NIN album and also a tori amos lyric. It makes reference to societal norms that encourage hate under a pretty mask, so to speak.
The owls are supposed to be dismantling the hate machine and the dude is supposed to be about to burn it down. It refers to an idea that you have to let go of hate to achieve wisdom.
None of that got conveyed even remotely, and the lyric reference is too obscure to communicate the message I wanted... so that's the long way of saying no it won't read your fortune. lol
one of my first websites back in the early 90's was dedicated to NIN so I understand
sometimes a piece just needs a descriptive title for others to understand better
Nice. I'm a toriphile, and I'm guessing you can see that I sorta whitewashed its actual meaning.

