images-discussions

[remote mountain]

What's that saying about a woman scorned? 🔥

[full cloud]

Lol. 😆

[remote mountain]

If enough people go after them maybe they will give up.

[edgy pine]

Bing Images >>> Dall-E 2

[full cloud]

Hopefully. That is generally the most effective way to stop something like this.

[quartz vale]

I think it depends on the image. I still prefer Dall-E 2 🙂

[edgy pine]

no

[quartz vale]

agree to disagree then, I guess

[edgy pine]

[teal timber]

Bing is Dalle2 so…

[quartz vale]

Isn't it Dall-E exp

[teal timber]

Dalle exp is dalle 2

Just a different type!

[remote mountain]

Wouldn't they be considered to be different versions?

[teal timber]

It's a different model so idk if I call it a version. But yeah w/e words you wanna use. They're related!

[tacit dust]

So much art posted I don't even have time to see each one :'). How can the images turn out so good? #daily-theme

Fr, y'all are talented. Where is this amount of imagination coming from?

[rose flame]

too expensive

[waxen sentinel]

generate a picture of a girl

can

can i use it generate a picture of a girl ?

[foggy flume]

line artwork

[fair crescent]

I found you can say it's a violation of sharing private information. It's complicated, but somewhere one of the lines is about sharing of your pictures without permission.

[fair crescent]

Yes, as long as it isn't lewd

[tacit dust]

What? No! It's perfectly fine to post a lot of art, I love it!

[dim cradle]

Hi, i'm a new user of DALL E, i registered just now but i see there are no free credits on my account, is that normal?

[raw harness]

How can I generate an image?
Which channel

[teal timber]

Yes, the free credits were turned off for new accounts about a month ago.

[native iron]

If bing uses DALL-E for it's images and can't be used commercially , what's the difference between using another platform than bing ?

[full cloud]

I found that. But I think that is about sharing an photo of you, and not an image made by you.

[heady wing]

I don’t use Twitter, but is there a function to report his account for these actions?

[remote mountain]

@heady wing Here they said that there is.

[heady wing]

Here’s a suggestion: DallE places a watermark to denote its made by it. Would it be possible to let individual accounts make their own watermark (with DallE) and add that to their pictures as standard? It would make the ownership problem a bit easier. This could be optional. It would not replace the Bing watermark only be added to another corner

[solar swan]

"a full-length ceremonial robe. lord capybara the first's robe of estate uses a rich purple velvet. it is embroidered using the goldwork technique. it incorporates butterflies, and cornflowers. historical costume display in a glass case at the british museum. lord capybara was noted for his short stature and plump physique." (edited) (╯°□°)╯︵ ┻━┻

[full cloud]

Unfortunately, it would probably take a very long time to implement a watermark function into Dalle if the developers would even decide to do it.

But there are apps that we can use to place our own watermarks. Or use Photoshop or any other image editing software.

@stuck zenith posted yesterday and asked if we can post images here that have watermarks added to them, even though we're not allowed to post edited images. I think that would be a good idea if the mods would accept it. And I don't see why it would be a problem, as the images would still be unedited.

[remote mountain]

I don't see the devs allowing every user to upload a graphic to be used as a Watermark.
The only thing that might be feasible is once Dall-E can do text is to use our Dalle-2 username as a semi-transparent Watermark. Or maybe give us the option to enter in some text into our account options.

©PsychoCabby

[frosty solstice]

hello guys, I have an assignement on AI generated art and I wanted to ask you something : do you think AI art is 'real' art ?

[remote mountain]

Yes

[frosty solstice]

why tho

I can't just say "yes" to my whole class

[remote mountain]

Sure you can. Just tell them a wannabe expert on Discord said "Yes!"

[full cloud]

My personal opinion is that AI generated images aren't automatically art, just like all paintings and photos aren't automatically art. But if there were some sort of human creativity and imagination, and an artistic intention behind the image, then the images are per definition art. We cannot define whether something is art or not by HOW it was created, only by WHY it was created.

From my experiences with seeing the work by and talking to other AI artists, my definite opinion is that AI art is more of an art form than most digital art that we can find online. Because there is much more passion and creativity behind the works, than the common digital art we find online, which are produced by skill, but often lack creativity and uniqueness (I'm saying that often they all look alike, and the work are clones of other work).

The artistic expression that we see in AI art is much more creative, personal and unique. And that is my definion of art.

[remote mountain]

Artist learn from many different artists and art styles to develop their own unique methods and styles. Isn't this how an AI also learns?

[frosty solstice]

after multiple researches, we can see that AI tend 'steal' copyrighted art on the internet without the artists' consent which leaded to lawsuits like the class-action one against stable diffusion; in fact AI art is not transformative art (which is legally a totally fair use of all the images on databases like Laion 5b for example) but derivative art which is not a fair use, while artists are clearly making transformative art of the references used

there is for instance the case of Kim Jung Gi, a deceased artist who 1 week after his death got his art stolen by AI developpers who then made money out of it

[remote mountain]

I'd say that by learning from training data isn't stealing. The diffusion method is transformative I believe. I'm still learning how all of this works.

[lone aspen]

lol at the amount of times ive seen this argument and countless times it has been debunked

people never learn it seems

literally watch any video of complains artist have about AI and you will see why

[remote mountain]

How is it any different than an artist in school learning from the greats? Should we outlaw learning?

[full cloud]

This is no different than how the art community as a whole have taught art for centuries. It is hypocrisy to say that AI art is theft, and human art isn't.

And your statement is also not proven, and if you look at how AI images are generated, you will see that it is false.

The AIs use databases made up by data that has been taken from analyzing images, and then creates new images based on that data.

That isn't theft. That is learning. However, many humans copy other artists images and styles to learn. If AI art is theft, then what humans have been doing for centuries is bigger theft. You have to be consistent in your judgment.

[remote mountain]

Artist can complain all day long. Doesn't mean that they are right. I'm some cases they may be right. When I see an AI output that very closely resembles a preexisting photograph that was used in the training data then I may agree. So far I haven't never seen such a thing.

[frosty solstice]

By comparing many AI generated images to the artist who the style comes from, it's clearly not transformative art which makes it not legally 'ok' for commercial use

[remote mountain]

Are you a lawyer who specializes in IP?

[full cloud]

I haven't seen it either.

And the artists aren't really complaining about the theft. If they are, then they are hypocrites. They are complaining because they won't be able to make money as easily. But that's life! That is what happens with all technological advancements in society. And my firm opinion here is that if they are creating images to make money, then they aren't truly artists and they don't have a true passion and love for art, they have a love for money.

[frosty solstice]

Yeah it does use those databases but it never quite "learns" things such as anatomy etc... By a simple analysis of AI generated images it's clearly shown that instead of learning, AI is just assembling those references like a jigsaw which explains why dynamic poses or hands always look off in those 'artworks'

[remote mountain]

Clearly your mind is made up and your biased against AI artwork. What do you hope to gain by coming here and calling all AI generated outputs stolen?

[lone aspen]

bruh this echochamber

you guys do realize that this discussion has been made by countless others more knowledgeable about the topics and you are wasting time in this echo chamber of AI bros

[frosty solstice]

it's why i'm trying to learn from others and listen to them, you should try it too it makes your essays wayyy more relevent

[lone aspen]

well anyway cya

[full cloud]

That was true with the first versions of AI, but it isn't true any longer. Dalle is better, and Midjourney and Stable Diffusion as well. And they will keep getting trained.

[lone aspen]

leaving this smelly place

[full cloud]

You're not listening. You're arguing.

[frosty solstice]

I'm studying law but still learning ofc

[frosty solstice]

I can do both

[full cloud]

Then show it!

Tell us what positive opinions about AI art that you have learnt from us today!

[remote mountain]

I'd bet money that within a year, likely much less, text to image AI generated outputs will be so perfect skin, hands and eyes will be indistinguishable from reality. With the rate at which it's advancing its only a matter of time.

[quartz vale]

I'm very interested in this "theft" topic and look forward to the result of various court cases. My bias leads me to believe that AI "Art" is transformative and therefore bot not theft

[karmic aspen]

hei

[remote mountain]

Dino I think you meant not theft instead of bot theft

[frosty solstice]

it's a great tool for artist who can now have way more copyright-free references, it's faster and easier to use for fun or to have quick ideas of what a character should look like and it's mostly free for compagnies ! but for now it's not enough "developped" to use it professionally (in comic books or whatsoever)

[remote mountain]

Where's the flaws at?

[full cloud]

Those are positive opinions, yes. But I asked you to tell us what you've learnt from us today. 🙂

[frosty solstice]

the colouring is muddy, the hair looks off, the reflections on the eyes are not placed equally which is confusing (talking about lighting) and the colour palette isn't emphasizing anything

[remote mountain]

I see a minor issue with the cat's nose. When I posted this I had made the image in the last hour or so and hadn't closely examined it prior to posting it in here.

[frosty solstice]

but yeah, I know it's funny to mess with AI

[frosty solstice]

the nose is the least of your problems, don't worry

[remote mountain]

This is what you get when you cross Yoda with a Maine Coon cat. I wasn't going for realistic here.

[frosty solstice]

Saying that this kind of AI isn't a technological advancement is a lie, and for kids or just for people who want to have a little fun by giving "life" to their ideas, AI is a great medium

[frosty solstice]

oh I didn't know

but as my friend always told me "stylization isn't an excuse for errors"

[full cloud]

You're still not repeating anything that we have told you today, thus showing that you probably didn't even read what we wrote. So you came here only to argue against AI, but not to learn something new?

Bring it on! Let me know where the anatomical flaws are.

[frosty solstice]

In fact, the method in this kind of work is to use clear informations and clear examples which is not the case with most of what you said, sorry

[heady wing]

As weird as it is to say the artists are derivative when the AI literally uses those as a database. But I do agree, it lets me without skill combine things you rarely if ever see combined.

[full cloud]

You didn't even read it.

[frosty solstice]

I honestly could spend my day telling you about the eyebrows, the shape of the bones etc but I have to work

Thank you guys for your help and have a great day !

[remote mountain]

I'm glad we had a real expert drop by to educate us today. I feel so much smarter now.

[full cloud]

The eyebrows... 😅 Lol

You just realized that you lost the futile arguing that you came here for. Bye!

[frosty solstice]

I'm glad I could help you ❤️

[remote mountain]

The sound of that Whoosh was deafening. Startled my poor cat who was asleep! Now you owe Merlin some treats.

[frosty solstice]

oh if you studied art you would've known that you can't just put it randomly on the face but it's okay ! Happy to have talked to you

[quartz vale]

I was busy studying when I typed that and didn't proof-read it 😅

[remote mountain]

We all knew what you meant.

[quartz vale]

Why should professional illustrations be gatekept behind more traditional means of art/media? Some people have made entire comics solely with AI images. I do think this discussion is heading to #ai-discussions

[full cloud]

I think that's because technical skill and creativity are two entirely different things. Someone with great skill can make a nearly perfect copy of someone else's artwork. But that doesn't make them an artist.

What I love about AI is that it gives us much more artistic freedom than we can have otherwise.

[frosty solstice]

Thanks for the information, I didn't know that and am now curious about how it ! I'm sorry if I used the wrong channel but I really got to go so have a great day

[quartz igloo]

how do i create an Image in Dall E general art

[full cloud]

I have studied art, at the University. I create physical art as well as AI art, and I'm a photographer.

I've also studied teaching, and I can tell you this: you are not learning! You are only being stubborn.

[remote mountain]

This

This server does not provide image generation functionality.
To generate an image with OpenAI's DALL∙E 2 AI model, please proceed to this link or click the button below.

[frosty solstice]

Oh, sorry it just didn't look like you studied it, I know I can be stubborn sometimes, sorry about it but I really have to go ❤️ Have a great day !

[remote mountain]

Have a good day

Now back to our regularly scheduled program.

[heady wing]

Its a bit hypocritical to say this is an echochamber and only listen to the artists who similarly have biases (stronger biases even as its their livelyhood).

[storm pebble]

Do u get free tokens on dall e?

[heady wing]

The eyebrows? The bones? Those are some of the least problems with AI! They are relatively static and any mistakes are obfuscated by how our DNA already includes similar changes. Since those are almost always in the same spot with similar features its one of the easiest to accomplish. Thats why a phone can recognize a face live! Because there are so similar recognition features for a simple program to latch on to! Its the hands for example which can be in a million different positions and partially invisible that prove problematic. So I think we just uncovered you are biased (also HOLY CRAP what an effort to read through that “request for help”)

[full cloud]

Biased, and also probably conforming to outdated views of art.

What I love about DALLE is that it generates realistic human faces. Not faces that look like models with perfect bone structure (which MJ and SD often do), but faces that look like real people. That's not a con, that's a pro. I have studied art and the anatomical structure and proportions of a face that is used in art classes. And those proportions are not realistic. They are however being used to justify the awful unrealistic beauty standards in society.

We can't use those beauty standards to measure the quality of an Al, because the Al is using real photos of real people to generate the faces.

Unfortunately, that is apparently how some so-called art critics still think.

What we've seen here today is an example of how to fail at having an intelligent and intellectual discussion. It is happening everywhere where people from the anti-AI movement are complaining about AI art. I don't want to sound rude, but that's my experience.

[stuck jackal]

how to draw picture

[quartz vale]

click the bot's generate image link

[robust compass]

On NightCafe users have the option to download their creation with an artist overlay

[robust compass]

"after multiple researches, we can see that AI tend 'steal' copyrighted art " -> cite these sources?

I realize this person is gone now.. was catching up 🙂

[fair crescent]

I wonder if we can do this discussion one time when the emotions are less heated.

Because I think we can decide that at least some of this can be art. For example, check the pictures I attached. These are pictures of Darth Vader as an Easter island head. A search showed me no artist has made this (I did see another AI generated picture though). In that it is an 'original' creativity. As it's a picture, we don't 'steal' styles from artists. Is it then art? The descriptive process was in the hands of us, the generation was facilitated. Much like making a picture. Some creative control, but otherwise facilitated by a camera.

(Btw, saying it's 'stolen' would immediately imply it is art, but I'll leave that one be)

[heady wing]

I think they also ignore an important factor: the prompt giver. They focus on the AI creating something, not on the person using it. The AI is in many ways an advanced version of Photoshop (and in many ways less advanced as well). Do we call someone who created something with photoshop not an artist because many of their actions had a semi-random effect behind it?
I am looking forwards to a time where I can basically talk to the AI. “I want a snowy bank on the left and a desert area on the right. Split them with a river, ok now add a tree. No not that one make it a poplar tree and place it more to the left, add a small incline underneath…” etc. This would add a whole grey area to the process. The level of AI “creativity” with my words changes with how detailed I am with each portion of the image. The difference between “give me a tree” and “give me a specific tree looking like <season> with a specific angle” etc.

[fair crescent]

From there we can go a bit further. General styles like impressionism, or stick figures should be art as well. No one can claim they are stolen, as any artist can simply use it.

The heart of the problem seems to be the stealing of styles of individuals. This is where it gets tricky. When and how can you claim it's stealing? When is it too close to the original? If Picasso would make his rendition of the scream of Edvard Munch, is it stealing? Is it still art?

[full cloud]

I agree with you 100%.

There are so many art forms that are accepted as art in the art community where the artists have much less control than with AI. Like splash art and acrylic pouring, or using chemical reactions. (I know there are more examples, but I can't think of them right now.) So it makes no sense to say that AI isn't art.

And it absolutely makes no sense whatsoever to say that it isn't art because it's supposedly theft. Those are two different questions.

Also the point about Photoshop is a good one. There are many digital artists out there who take stock photos and turn them into something new using Photoshop. Whether they pay for the images or not, I don't know. They could be calling it fair use and not pay. Still, they get accepted.

[heady wing]

Even better is that this AI art will increase in control of the promptgiver. As time goes on inpaint outpaint could become the next standard where you can quickly alter the pictures you get to what you had envisioned. That is key, what you, the creative process behind it, envisioned. There are parts where you can argue its AI art (still art) as the AI provides the entirety of the artform. The surprise me button and then the AI creating something from it is an example.

[full cloud]

The critics against AI don't seem to understand how prompting works. They underestimate the work, the creativity and the experimentation behind writing the prompts.

I talked to one person who also does create AI art, but who wouldn't call himself an artists, only a promptist. His argument was that if a company hires a designer for a project, the company is not the artist, only the designer. However, that's not how prompting works the way AI artists do it. The analogy would be that the company and the designer are working together in planning the project. All of the people involved would get the credit for the finished product (which isn't always true in reality, but it should be).

Another analogy would be writing a paper together. Everyone who is participating and adding to the finished paper should get the credit, not just the person who actually writes it all down.

No AI generated images that were generated from a prompt written by a human can be credited to only the AI. They wouldn't exist without the involvement of the human. So they are co-creators. It is very simple.

[full cloud]

I think that this is something that has to be considered from case to case. And only if the image could be infringing on a copyright.

I think that if someone is reproducing the style of another artists, whether manually or by AI, it is their own responsibility to consider whether it's too close to the artists original works or not before they publish it or use it commercially.

[true harbor]

green

[remote mountain]

I know I have used AI to create outputs that are unique and most definitely never existed before. When the AI takes several different items and different styles and combines it all in a new and unique way that is transformative. Sometimes I get this result as a happy accident like happened yesterday. Other times it was intentional.

[heady wing]

In 20 minutes the Dutch will have 2 minutes of silence to remember the dead. While focused on WWII casualties it has become more generalized for the dead of any armed conflict. You are welcome to participate (or not).

[heady wing]

Anyone considering this guys request, its a vital part of your PC programming

[full cloud]

2006 called and it wants your prank back!

Go be a jerk somewhere else!

[heady wing]

2006? I thought it was older!

[full cloud]

It is. But I googled it and the Internet said it became popular in 2006. 🙂

[remote mountain]

I reported that message. It should disappear soon enough.

[full cloud]

Good! 👍🏻

I can't believe people still try this.

[dim cradle]

car

[heady wing]

Click on the generate image link on the message by @verbal sequoia (BOT) Below if you want to create something with it

[cinder minnow]

how do i use chat gpt

[full cloud]

You should ask this in #chatgpt-discussions

[dim cradle]

i can do this because im a VIP player of state of survival ofc with 50 million power

[remote mountain]

I'd advise against doing such things on this server. But you do you.

[full cloud]

Trolls. This is what they look like. Or at least what their souls look like.

[dim cradle]

Rateify
Product Reviews logo

[edgy stump]

oh my. agree

[gilded osprey]

The Bing image generator can produce pretty awesome pictures, but the user experience of using it with the Edge browser on an iPad is just AWFUL. (I can forgive a lot, but not the fact that the text being entered goes behind the virtual keyboard on the screen. YUCK!

[dim cradle]

I got banned from chat gpt and idk why

[full cloud]

I don't think that anyone here can help you. You should contact the ChatGPT support.

[main gull]

A minimalist painting in the landscape of Cappadocia with the phrase "We are learning Turkish" written in Russian in the sky where the sun shines and travel balloons fly

[balmy finch]

in the bing creator how can i see all of my images that i generated? The side bar stops after a few

[robust compass]

reported

[maiden swift]

Make Your Product Shine with 3D Bottle Mockups

Make Your Product Shine with 3D Bottle Mockups

[twilit tiger]

The images that are generated by AI , whether it’s DALLE or another AI , can they be copyrighted by the person who came up with the prompt ?

[amber sapphire]

How do create a character that can be used again and again in different environments and circumstances?

[robust compass]

dependent upon the service you are using, they each have their own terms.

[robust compass]

https://help.openai.com/en/articles/5008634-will-openai-claim-copyright-over-what-outputs-i-generate-with-the-api

[raw elbow]

U.S. Copyright Office has pretty much just said no back in March this year

But laws change all the time so, yea

[dim cradle]

sup

???

[twilit tiger]

Thanks. So it seems that OpenAI will not claim the copyright for the generated images. How about next step. I generated something with DALLE, and I want to have copyright to that image and use for let’s say my business. I suppose that is possible , just like with an image I captured myself for example with my camera. Right?

[full cloud]

You cannot copyright AI generated images. You have ownership of the images, but no copyright. You can use the images for commercial purposes, as long as you let people know that the images are made by AI. They are yours to use, and OpenAI won't claim them.

But you cannot copyright them, because that's what the legislators have ruled. That's because you can only copyright something that has been created by humans.

[full cloud]

The issue with this is not whether you can use the images or not. The issue is whether other people can use them or not, and that is a gray area.

[dim cradle]

how is a discord bot better at making art than a multi-millionare company

[gray jasper]

I want to combine Dalle functionality with something like thispersondoesnotexist in a way that I can generate realistic humans with natural anatomy and get either full body shots or headshots based on physical descriptions, as opposed to randomly generated then having to manually edit them, or getting some of the janky photos we see from Dalle. Does anyone know of a way to integrate them, or a way to get the dalle model to make people more effectively?

[quartz vale]

?

[robust compass]

Dall-e 2 isn't a discord bot

[balmy finch]

yo how come i cant see my old generations anywhere? even if i try to look in my history it just starts generating new results when taken to the page :/

[serene ridge]

it's not. AI doesn't make ART. It generates images.

[bleak sedge]

mango

[limber dove]

Why does it say my generation goes against TOS

[hushed ingot]

dall-e-discussions

[timid thunder]

Hey guys I'm trying to figure out how people make those AI avatar apps. Specifically, I want to be able to upload an image of a person's face and have that face translated onto a separate image generated by one of the AI image generators. What APIs do people use to do this?

[fair crescent]

Do note that there are things you can do to still claim copyright. If I use the images to make a comic for example. I've put in human effort for a certain order, possibly editing the pictures in the process as well with text or the like. A logo for a company can probably be licensed in different ways, making it copyright protected. Or simply edit a picture in certain ways to make it distinct. So there are some ways you should be able to still copyright AI images in some ways.

[austere hinge]

How to develop the habit of reading

[fair crescent]

Practice

[hasty solstice]

how to generate images from this discord ?

[full cloud]

As I have understood it, from the case with Kristina Kashtanova, only the human work of arranging the images would be copyrighted, but not the images themselves. She originally got the book copyrighted, but after it was revealed that the images were made by AI, they changed it so only parts of it is copyrighted. As I understood it, the text and the layout have copyright, but not the images.

I don't know how much an image must be edited for it to count as distinct from the original image, but I doubt that only minor photoshopping (like color correction etc) would count. I should look into it.

[still valve]

are there AI's that can fill in missing gaps? such as this one for example

[full cloud]

You could try by inpainting it in Dalle. Import the image and then erase the gray area in between and describe what you want there instead.

[still valve]

ah true

I will try first if I can draw it in by myself lol

my progress so far xD

[full cloud]

It should be fairly easy to draw it yourself. 🙂 you can also try using the spot healing brush tool in Photoshop. It might work.

[still valve]

I'm using Gimp but I found out that the Airbrush tool might work

I'm just doing the rough colors with pencil tool first and then Airbrush to fade it in

[full cloud]

I would try copying the already existing image too, and then mold it somehow to fit.

[still valve]

hmm

[full cloud]

If there is a warp tool in Gimp that could work quite well.

[still valve]

idk

the grass isn't too bad

I just used the air brush

maybe I'm too skilled to use AI XD

[slate pecan]

so for dall-e does Bing Image creator count as it uses dall-e

[full cloud]

Yep!

[slate pecan]

nice

[high ingot]

Are you real AI?

[prime quarry]

yes

[high ingot]

prove me?

[raw harbor]

newbie question - is there a way to use words to enhance or alter a dall-e generated image?

[full cloud]

"Prove me?" You want the AI to prove that you're real?

[vapid elk]

the variation system does not allow for including a prompt

[open wind]

Anyone know if outpainting will be available on ipad soon?

[buoyant rain]

I started an account on DALE but it says i have zero credits and i have to buy. I thought the process was free-of-charge? Did i do something wrong?

[full cloud]

Dalle isn't free and you have to buy credits to use it.

However, Bing Image Creator is free, and that is powered by Dalle.

[buoyant rain]

Thank you, super helpful answer! 😃

[summer rain]

Bing’s is better.

[full cloud]

That depends on your definion of better. Bing has better quality, without a doubt. But the original Dalle2 has much more variation, character and realism, whereas Bing often produces images that look filtered and edited.

I wish that Bing would have the same variation that Dalle2 has.

[summer rain]

Well, I think it’s more important to generate images that match the prompt.

[full cloud]

That can be a bit of a hit and miss with both Dalle2 and Bing. Sometimes it works, sometimes it doesn't.

My main problem with Bing is that it does NOT match the prompt in many ways. One example being if I ask for a woman without makeup, that's what Dalle2 gives me, but Bing doesn't. Same with colors, filters etc. Bing requires a lot more work to generate natural images. Dalle2 rarely miss details like that, but Bing often have issues with it. Bing tends to go into the more popular mainstream direction, whereas Dalle2 is more creative.

[summer rain]

“a cat wearing socks” looks simple, but why DALLL·E 2 just doesn’t handle that well?

[full cloud]

It sometimes does. But that's only one example. It isn't wise to base your opinion on only one example. You've got more examples?

[vagrant zodiac]

Bing is powered by Dall-E

[summer rain]

I have tried a few times before.
Although most of them were generated last year.

[full cloud]

I understand. But that's still only one example. I don't think that we can decide that the images don't match the prompt based on one single prompt.

[cinder saddle]

do any body have idea how to create (write in dalle ) for graph or some 3d render animation for timeline of UNIX,linux ,OS?

[dim cradle]

bro is this server dead?

[prime quarry]

try this A furry feline with four paws clad in colorful knitted socks, padding around with a look of mild annoyance. The mismatched patterns and colors clash with its sleek coat, but the socks keep its toes warm on chilly days.

[prime quarry]

no

[fleet wyvern]

Is dalle down ATM??

I cannot gen again. Keep erroring

[rotund matrix]

gm

[ruby vault]

Logged in to "try" the software only to find out there are no credits in my account. How is this "trying" the software? Thanks for wasting my time! No way I'm paying a company who says there will be 15 free credits to try and you log in and have none. Pretty pathetic if you ask me. . .

[dim cradle]

remove the credit system

[edgy pine]

bing > Dall- E 2

[edgy pine]

How much data is Bing Creator collecting per day? Are they surpassing the competition in the RLHF? If they are using it.

[little walrus]

Has the v2 Experimental mode been shut down? Every attempt results in "Your task failed. Sorry about that! Please try again." Old model still works though

[balmy finch]

anyone know how to view all previous generated images in bing image creator? It only goes back so far on the side panel.

hoping if theres no way other than manually doing it in my browser history that they will implement it soon on the website

[fleet wyvern]

Wondering the same

[remote mountain]

Some of us have noticed that Experimental is currently down. That's all that is known for now.

[summer rain]

I just said that it doesn’t handle “a cat wearing socks” well. I thought it would be simple for the AI, so I’m wondering why it isn’t good at it.

[little walrus]

Thanks. It's a weird error all around -- even when v2 was working, I would get it for certain prompts that didn't violate the content policy but seemed to trip some other filter

[prisma oyster]

Greetings!
What commands do I use to generate an image?

[remote mountain]

Dall-E 2 Experimental is working again.

[fair crescent]

This is not the place to share, even if the intent is good.

We cannot know your intent, or easily verify the files on malware. Even if your intent is good, someone else can spike the files.

Lastly, this is DALL·E discussion and not a learning site.

[safe bobcat]

What? It's just Pdfs , are there PDF malwares?

[fair crescent]

I can give arguments, but it does not matter. I appreciate you trying to help people (I assume), but this is not the place.

[tall mason]

Yes, malware can be placed in a number of seemingly innocuous file types like pdfs and pngs.

[sweet bison]

if I signed up for dall-e today, do I need to buy credits to use it?

[eager tundra]

Can anyone please get me a prompt to generate a hammerhead shark? 🥲

[lime quiver]

how did you this pictures

[dim cradle]

I need to creat a company logo and i am not beeing able to place the company's name in the picture

How can i do it?

[molten oxide]

Do it with DALL-e or with anothar ia, only give the instructions clear an precise

[dim cradle]

I am giving the precise description but it is not inserting the text in te picture, it gives me the kind of logo that i need but just not with the name in it

[molten oxide]

Maybe you can say the ai to make the space and then you edit the image with canva or something similar

[dim cradle]

Maybe that is going to be the solution. thanks

[remote mountain]

Most of the time Dall-E 2 will fail to produce coherent text. Occasionally you will get lucky. But that's the exception to the rule.

[tawdry plover]

How to make art using dall e

[remote mountain]

To generate an image with OpenAI's DALL∙E 2 AI model, please proceed to this link or click the button below.

[remote mountain]

You can't make images from this Discord.

[heady wing]

Could someone help me with a prompt? I always liked the attempts at semi-realistic units, buildings and background in old RTS games, mainly Tiberian Sun. I’ve been trying to recreate it but at best I get the more cartoon style isometric voxel art. Could anyone help me find the wording to make it?

[sick spear]

Hi! It's worth asking around in our #❌┃prompt-help channel for tips on this 🙂

[heady wing]

Thanks I missed that channel

[burnt acorn]

I stopped receiving free credits to my dalle account. Anybody know why ?

[frail skiff]

Is Dall-e getting better? I haven’t seen any new features or advancements recently. All the focus seems to be on ChatGPT.

[wispy gale]

Hello, are the images generated by DALL-E free of rights?

[vapid elk]

the rights of the image are attributed to the person who generated it

one cannot use the images without permission from the person who generated it

[wispy gale]

thanks

[near wasp]

Does anyone know the prompt for these kind of images?

[teal timber]

Like the art style?

[green birch]

hi why I don't have free credits for new accounts on DALL E 2

[green birch]

everywhere on the internet they say you receive 50 of them at start or smth like that

oh i see now

[little walrus]

Low-poly

[near wasp]

Thank you very much.

[lunar wind]

"It looks like this request may not follow our content policy." would be nice to know precisely what text is flagging the system.
lol, I am not allowed to "a mouse pouring milk" ? that is flagging the system?!

[fair crescent]

There are a select few able to test the experimental branch, which is also available on the bing picture creation. These advances are immense. Most for the best, a marginal few arguably not (which might be the reason for testing). For example, it is able to understand much longer or 'chained' prompts without artefacts. It is better at faces. It is more stylised, but people in generally like it more.

It is unknown when these changes will be put on DALL·E proper.

[past ginkgo]

Hi. Lets suppose that 4 images are generated of a dog and i want to ask for a change on the 3rd one. Is that possivle? For example: Put glasses on this man

[robust pewter]

Dall-e or Dall-e mini. Which is used more?

[teal timber]

It is not. But you can add to the prompt and hope it makes the change with a different generation of everything

[teal timber]

de mini is not from OpenAI so please keep that off the server or in #ai-discussions but I think DE2 is clearly used more.

[dim cradle]

how

[fair crescent]

What is your question?

[errant minnow]

Dalle seems to be infinitely loading when trying to generate an image for me

[inland nexus]

Same

[rugged cove]

Hello , can any one help me .. i can not upload a photo for chat gpt 4 for analysis

[gloomy mauve]

Are DallE2 API is error ?

[solid orbit]

is dall-e working right now? All my tries are canceled.

[kindred flare]

yes same here. dalle cant generate it is cancelled after 10+ mins of waiting

[fair crescent]

DALL·E has an outage. You can check the status here: https://status.openai.com/

DALL·E seems to be up and running again. Check https://status.openai.com/ if anything is still amis.

[proud bobcat]

hiiiiiiiiiiiiiiiiiiiiiiiiiii

[dim cradle]

I AM A AI SENT FROM NUKETOWN

[full cloud]

It's an AI, not a AI. 🙂

[crisp oyster]

hello, is there a kind of API for DALL-E that I can download to use it?

[teal timber]

DE2 has an API yes. It costs money instead of labs credits to use. Can find the documentation on the openai website and some examples.

[lunar wind]

will dall-e recognize engineering scales, if provided?

example: if I created a drawing at a particular scale and submitted it to image generation to facilitate additional content would dall-e recognize 1/16" = 1' if provided that information?

or tell dall-e not to manipulate a particular area/region of the user-created content

[teal timber]

Yes if you edit an image you can use alpha masks for where DE2 edits and the rest is left alone

[lunar wind]

i'm not familiar with that. how will dall-e create content particular to the user-created drawing?

[teal timber]

The generation frame in the editor says where the image will be edited. You can use the erase tool to create alpha channels where de2 fills in the blanks. Then you use the prompt bar as usual at the top to tell it what to put there. It will use your image that it sees within the frame and the empty space alongside the prompt to meld together a final image

[finite shadow]

Hello DALL-E fans. I have an extremely basic, n00b question. I'd like to use DALL-E to simply create a Pixar-like version of a real human photo. Is this possible with DALL-E, or is it more for creation of images that never existed in the real world?

[teal timber]

DE2 doesn’t do style transfer but you can ask for a Pixar style in a fresh generation

[unreal orbit]

bro why is it so slow now

[fair crescent]

If we're being pedantic, DALL·E is actually ML, Machine Learning 😉

[prime quarry]

artificial intelligence does not write letters correctly any prompt tip?

[lyric bane]

tanjiro

[dim cradle]

Hello you cuonpiuter has vairius 💀

[neon yew]

/help

[dim cradle]

Could someone tell me what the benefit of dall-e website is if I can use it for free on Bing?

[rapid needle]

Is there a way I can still get some free credits? I really love OpenAI's chatgpt 4 and didn't know DALL-e was made by you guys as well or I would have checked it out sooner.
I am extremely creative and would have a million and one uses for this app. Sadly I don't have much income, so I can't afford to get monthly subscription plans...

[full cloud]

Lol. But would that be a ML or an ML? 😛

(It's the latter, I think)

[fair crescent]

After some research I think it's the latter too, but without that research I wouldn't have known at all.

[full cloud]

I'm an annoying stickler for grammar. 🙃

[midnight nest]

#images-discussions Digital- "Future of Technology"

[teal comet]

are images people generate free to use, or does someone need to ask for permission to use the images someone else has generated? giving credit to the prompter would be a good idea

[tropic brook]

I think free to use would be better since the user isnt really doing much other than entering a simple prompt

[teal comet]

thats for the common user typing down one sentence. but some people seem to understand what works better than others and sometimes it can take hours to generate the image you have in mind. if its specific, if you take care of lighting, pose, background elements. then its no longer a simple prompt. thats why im wondering about this

[tropic brook]

You do have a point.I guess if the user is actually trying to creat art of some form and is being creative in the specifics of the prompt , I guess he could get credit.But then again , it could be the decision of the prompter itself whether or not he wants credit or not.

[fair crescent]

A difficult question. For DALL·E the ownership is with the prompter. However, copyright cannot be given in many countries. This makes it complicated. You probably have quite a bot of leverage if they try anything commercial, but nothing foolproof.

[full cloud]

As I've come to understand it, all AI generated images are now considered public domain in the eyes of the law, no matter what rights the company behind the images state in their terms of usage, and no matter how much work and creativity we've put into it. And public domain, as far as I understand it, means that it's free to use commercially (someone please correct me if I'm wrong). So we might not have any leg to stand on even if they do use it commercially.

However, I do believe that this will have to change soon. AI artists won't be silent if we can't have copyright on the images that have been generated from our own prompts and sometimes hours of work.

When I publish anything online outside of this servers, I always do a lot of editing in Photoshop (color correction, cleaning up artifacts etc). Enough so you can tell the difference between it and the original. And I make it clear that I've edited the image. In that way, if someone steals it, they're stealing my work and if that's the case, then I would have a leg to stand.

[placid kraken]

bot? ¯_(ツ)_/¯

[umbral spruce]

why is midjourney so much better at fingers than dall-e?

[full cloud]

I don't know how good MJ is now, but in the previous versions, Dalle was always better at hands than MJ. I got perfect hands in some pictures last summer when MJ was still making 10-finger clumps.

[fair crescent]

Have you tried the experimental? You can try the experimental version via the bing create

[bitter glacier]

nepali handsome youth in mrr ultra realistic

[civic sorrel]

I just signed up for some credits, and tried to create images of kids to use for clipart. They look absolutely terrifying!

photo-realistic picture of a group of diverse children reading on computers with different emotions on their faces

[civic sorrel]

And faces and emotions, comparing my tries over the weekend with this morning.

[teal comet]

ive read something about only the changes you've made to the images being protected

[full cloud]

Yes, that is true. As I have understood it, that is the case with all images that are public domain.

But if you don't post the original image anywhere, no one can use the edited image, because there is no way that they could use it without the edited parts.

So if you edit the entire image, that entire image would belong to you, if it has been edited enough for it to be copyrighted (small changes aren't enough).

[teal comet]

haha big brain

[full cloud]

You think that MJ is better than Dalle at portraying emotions? It has always been the opposite. Where both MJ and SD often give emotionless faces, or over-the-top faces, Dalle gives natural emotions.

[teal comet]

change the prompt a little. make it more cleaner. for example. "...computers showing different emotions" instead of with. also dont write everything in one line.

[civic sorrel]

I've got a small sample size, but here's a Dall-E image:

[teal comet]

got this by replacing the word with to showing in bing image creator

[civic sorrel]

Do you see the distortion in the faces though?

[teal comet]

was worse before tho

[civic sorrel]

True, but I wouldn't use it - the distortion is the first thing that I see.

[teal comet]

ye you would need more tweaking, it would take time

[full cloud]

You should try Bing. The quality is much better. Or write longer photography prompts, then you will get much better images.

[civic sorrel]

Here's a MJ picture from over the weekend - different prompt, but still -

OK, maybe I'll try more. I'm a researcher (ML/AI in education), and these are just replacements for clipart / eye candy, so I don't have a lot of patience/time that I shoudl put into it. Just found MJ to be way better.

I did end up going artistic and "leaning in", creating an image using the style of a mexican muralist known for. distortion - orozco.

(used Bing for it, now regret I spent the $15 for Dall-e credits this morning!)

[full cloud]

Consider checking out my Instagram, link in my profile. I haven't posted anything since new years, but you can see some emotions there. Or you can look in #photography and you can see a lot of images that people have posted.

[teal comet]

this could be due to some stock photos thingy. its only generating 1 image instead of 4, flagging other three i suppose?

[full cloud]

That happens a lot for me when I try to generate humans. I also get a lot of people without clothing, so I think it might be because it doesn't always automatically generate clothing, so a lot of it gets flagged. If I specify a type of shirt, I get much less flagged images. It's strange.

[teal comet]

that didnt happen to me a lot or i might not have noticed. i thought it was due to some legal issues, of images seeming too similar to the images in the dataset they were trained from

[full cloud]

I don't think that they censor that at all, unless I've missed something. I think that they only censor inappropriate images.

[teal comet]

or it could be because of the environment u're specifying in your images. eg. if you use beaches

[teal comet]

no, they also clearly block prompts, and also images i think. because i tried to make nikola tesla, and i got a warning

some content policy note

[full cloud]

No, it doesn't seem to matter what kind of setting I use. But I often get the dog message a few times, and then I add clothing, and it works.

[teal comet]

hmm ig i might not have tried a wide variety of things

[subtle plinth]

Question about Dall-e.
I'm worldbuilding for a dnd campaign. I'm currently trying to map out my world. kingdoms, cities, natural formations, distance, etc.
If i put into dall-e descriptive prompts of areas will it generate a top down type map view of the area

[fathom jolt]

Is there any way to get free credit in dall e

I logged in for the first time .but it says i have 0 credit

[subtle ferry]

New account?

[chrome ocean]

Anyone here use dalle2 to inpaint picture ? Dalle2 always generates some ghost text/number on the generated result , can I use my prompt to remove it in the generation time ?

[teal timber]

I inpaint from time to time. Never had many issues with that. What do you mean?

[chrome ocean]

something like this

[chrome ocean]

I wonder if I can use the prompt magic to prevent it

[teal timber]

What prompt are you using now? If you enter a prompt it should follow it and merge with your image

lol

[chrome ocean]

I don't provide accurate prompt now , just say like 'fill the empty space' .

[teal timber]

I'd say just prompt what you want to be in the space

[chrome ocean]

Thank u , I will change the prompt and make it related to the image and see the results 🫡

[drifting vale]

Can I use DALL∙E 2 AI for free?

[teal timber]

Generally no. Users from the past got free monthly credits but that went away. Bing can use DE2 and is free I believe.

[pearl apex]

.

[fair crescent]

You can ask for a top down view. It is sometimes that easy 🙂
You can add to it by asking for a grid, but with more variable results.

[fathom jolt]

New to dall-e yes. but old to chat gpt

[wispy raptor]

k

[reef steppe]

a painting of a fox in the style of Starry Night

[native frost]

Can anyone get Dalle to generate a 6-legged Horse, i tried enough but it just won't do it

[fair crescent]

Interesting proposition. I'll see what I can do

[fair crescent]

I can't get it to work either. There's probably so much reference material to four legged horses that the six or eight legged horse will not work. I've tried the 8 legged horse Sleipnir as well, but to no avail. DALL·E knows its a horse, but more legs is too difficult

My best attempt

[native frost]

So is it one of those prompts, which might seem easy, but are nearly impossible for the AI
Maybe it needs more temperature, or some fine tuned settings to be able to apply proper creativity

[heady wing]

I tried as well, but most of my pictures aren’t shown due to DallE thinking them sexual/violent

[elfin cliff]

(worst quality, low quality, large head, extra digits:1.4) , out of frame, duplicate, watermark, signature, text, oldfashioned, pale colors, grayish, borring, dull, lowres, text, error, cropped, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, out of frame, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, incorrect anatomy, chibi,hat, horns, crown head gear, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck (searchin this good negate for any prompt)

[heady wing]

My closest attempt so far

[dusty fossil]

when i run out of credits do I have to pay for dalle?

like does it reset every month, or can I never get credits again when I run out

[subtle plinth]

Cheers

[dim cradle]

how to make a prompt

[rose charm]

a poster with a butterfly and mushrooms on it, poster art by Esther Blaikie MacKinnon, featured on pinterest, psychedelic art, half of mirrors, tarot card, poster art

[north minnow]

Can i add diferrent styles to my own logo in Dall-E ?

[meager trellis]

Hello?

[full cloud]

I also tried...

[quartz vale]

Running exp gave me those . They're closer I think

[fair crescent]

What you can try is grabbing a picture of a six legged horse and put them through a picture to prompt service. I cannot share links, but apparently IMG2PROMPT and CLIP INTERROGATOR work well. I can't get them to work on mobile though. With that prompt you might be more lucky. If it works, please share!

[light pelican]

.....

[pure flint]

I am trying out Dall-E for the first time and trying to get it to generate an image off of an image it generated earlier. I found this link (https://help.openai.com/en/articles/7437313-image-generation-beta), which explains how, but I am not an expert Python coder. How do I change the Python code so Dall-E does not generate another sunlit lounge?

[simple flame]

hi, can you help me someone? i am searching for prompt that make for me a bussines card. Is it possible? I am iphone android notebook repair shop

[fair crescent]

You can check out inpainting/outpainting yourself. It'll be easy to erase the icons and regenerate some other version. Note that it'll be square and 1024x1024. So you need to get part of it, erase the icons, regenerate 4 options, choose one, paste it back on the original and save it.

[scarlet dagger]

Thanks, is it free?

[modest viper]

Is there any news or rumors about a possible optimized 8GB vram version of DALL-E for local use?

[quartz vale]

There is no news that Dall-E will be open source in the future

[prime quarry]

lol

[hasty coyote]

yes

[fair crescent]

If you've got an active account from before the 6th of April you can use free credits. Otherwise no.

[crisp fulcrum]

DrawLove shaped, Chinese painter SanYu's style:: --ar 1:1 --v 5.1

[quartz vale]

you need to click the bot's "generate Image" link

[crisp fulcrum]

l get it！Thanks a lot ❤️

[quartz vale]

that's great to hear!

[zenith walrus]

Hello, I am trying to use the dall-e API for something maybe too unconventional. My professor's lecture slides are very messy and hard to read, and I tried uploading his slides via the API edit endpoint: (https://platform.openai.com/docs/api-reference/images/create-edit) with the following prompt:

"This slide from a computer graphics presentation is extremely messy and hard to read. Reformat the entire image with readable text and better diagrams. If the slide is dense, include more whitespace."

However, the response image I got back was the exact same image that I uploaded. My working theory right now is that dall-e had no idea how to fulfill my request, so just gave me nothing. However, I was hoping for at least a badly done revision, not nothing. Is this a bug? Has anyone experienced something similar?

[quartz vale]

Dall-E 2 cannot alter your image like you're wanting it to. It can only create, not change pre-existing parts of an image

[zenith walrus]

https://platform.openai.com/docs/api-reference/images/create-edit There's a beta version editing images here, so that is not it. However, I just realized what I did wrong. I need to upload a mask for what I want edited.

[quartz vale]

Dall-E 2 isn't capable of altering your image like you want it to though. Its "edit" feature allows you to recreate parts of an image

Additionally, Dall-E cannot generate accurate text

[zenith walrus]

But wouldn't I be able to have it recreate the whole image in a much neater way? Ok - the accurate text detail is certainly an issue. Thanks.

[quartz vale]

Dall-E 2 can only generate based on what's in the generation frame. If you delete part of the image, it no longer has context of what used to be there

[zenith walrus]

Theoretically though, if let's say dall-e was better at text, I could upload a lecture slide along with a 1024x1024 transparent mask, which would effectively say redo the entire lecture slide. The endpoint allows for the image and mask to be uploaded separately.

[quartz vale]

that would not work. The mask erases parts of the image. If you set it to be 1024 x 1024, it would make an entirely new image. Dall-E 2 can only create new parts of an image

[zenith walrus]

ahh, I see. So it has no knowledge of what existed in a section of the original image if that section overlaps with the transparency mask correct?

[quartz vale]

that's correct 🙂

[zenith walrus]

Maybe it's just me, but I think that could be clearer in the documentation. But thanks for clearing that up for me!

[queen oar]

how i make An image of a modern workspace with comfortable seating, large windows, and a kitchen area stocked with coffee and snacks.

[candid steppe]

how to describe an overweight and girly lady

how i make An image of a modern workspace with comfortable seating, large windows, and a kitchen area stocked with coffee and snacks.,,

[robust anchor]

bot

[full cloud]

The time limit was changed to 30 minutes in #daily-theme

It's terrible!
That is waaay too long and will probably quench our creativity. A lot of people post lots of great images for half an hour to an hour or so, and it's a lot of fun to see people's creativity flow, and be inspired by each other.

It is probably going to cause me to stop posting there altogether. Or even looking there, because people won't be able to post freely and there will be nothing to see. 🙃 Way to ruin a great channel!

[balmy folio]

hi guys i use dall e for free and i just made an acc

[eternal kindle]

can we use dall - e model to generate images for free?

[vagrant zodiac]

The change has been in response to a large uptick in how active the channel is. We've noticed that some art is now getting buried due to very frequent posting by individual users, so really the opposite is happening and not everyone is getting a fair chance to show off their work. It might be a good idea to save your best generations and use them for #daily-theme , and post other ones in the various galleries 🙂

[full cloud]

I understand your logic. But I don't think that any art is getting buried at all. I believe that most active people are going through all of it and leaving stars on the images that they like. If someone isn't getting noticed, it's not because the images aren't being seen.

We are only 10 or so truly active members, and many of us enjoy uploading many images and seeing each other's work. This is a tiny community, so it is a good thing that people are posting a lot. I would understand these restrictions if there were more active people, but not when there are so few of us. These restrictions are only in the way.

I'm not gonna have any images to post in the other galleries, because I simply won't participate in the daily theme. You're taking the fun and creativity out this server. It's just getting worse and worse. Soon this server will be a ghost town. You'll see a tumbleweed every now and then, just a few dried up uncreative and uninspired images passing through. And scammers and spam. And then silence.

[full cloud]

The daily theme was the only fun and inspiring thing left in the Dalle channels and the only place where people have been really active. The uptick you've seen is because people have been enjoying themselves, being creative together, letting their creative juices flow in an inspiring channel. And now you are trying to remove that. Guess what will happen with those people who feel like their work isn't getting attention, if people stop participating in the daily theme?

🦗🦗🦗

[vagrant zodiac]

While I appreciate you want to share more images, as I've stated this is why we have the galleries. The daily theme is meant to be a point of inspiration and I'd encourage everyone to share their images more widely outside of that channel. I'm sorry if you do end up choosing to not participate due to an increase in the slowmode, however there's nothing preventing you from sharing your daily theme creations elsewhere. The uptick in buried posts has actually been brought to our attention multiple times, so it is clearly a growing issue.

a more proactive way to work with the slowmode is to utilise the other channels more actively, but we will certainly monitor the effects of this slowmode and make adjustments as and when needed 🙂

[full cloud]

Would you really want people flooding the gallery channels with tons of images from the daily theme? That would bury all image that does not come from the daily theme. The daily theme images should be contained in one channel, don't you think?

Would it be possible to have two daily theme channels then? One were people can be creative and have fun, and one where people can post their favorite images? That's how they do it in the daily challenge in the AI art community server, and it is working.

[vagrant zodiac]

Not necessarily. They'd also be more dispersed over the relevant gallery styles 🙂

also, you can now create threads in #daily-theme

if you want to use a particular piece as a launching point for more discussion/iteration, that could be useful

[full cloud]

That isn't what people are doing though. It isn't how the daily theme is being used by the active members.

The way I perceive the daily theme as it has been working for the last month or so is a place where people are being creative with one theme each day, and pouring out their creativity together around one subject without being restricted. That's why people have been so active recently. It's a community of creative minds and not just a gallery. This could not exist if we have to go looking in lots of different galleries.

[full cloud]

By putting restrictions you are hurting the people who are posting a lot and who love art and creativity and are enjoying the process of creating (the people who use Dalle/Bing a lot) and helping those who are only posting a little (the people who don't use Dalle/Bing as much). So a few people who only are looking to get likes on selected images that they personally like (but obviously other people don't like as much) are getting more space and less competition. That is elitist and you are only encouraging selfishness. It is the opposite of what AI art should be all about.

[full cloud]

I will stop participating completely on this server if the 30 minute time limit on the daily theme is still there when this day is over (in Sweden).

Just letting everyone know in case anyone wonders why I won't be posting any longer.

It's because I don't have the energy for all these oppressive rules and restrictions any longer. It's really been getting out of hand and no one is benefitting from it except some people who are only looking to get likes on their images. I love DallE but this server is only getting worse and worse.

The only really great thing about this server lately has been the daily theme and seeing everyone's images and creativity. I wish people would be posting lots and lots more images in that way. But instead we're getting restricted. First they removed the emojis for some time, and this 30 minute rules is just the last straw for me. Just letting everyone know.

[compact vigil]

How do I get free credits

[full cloud]

You didn't do anything wrong. A lot of people have been posting lots of images. And it's been great. A lot of masterpieces have come from that process.

The issue here is that some people are upset because they aren't getting the attention that they want, so they want to limit everyone else. But the harsh truth is: people aren't giving them stars because their images aren't that good. It's jealousy, and entitlement! That's all there is to this! Some people are feeling insecure, and instead of experimenting and creating better art with the rest of us, they want to limit the rest of us, so they can be in the spotlight.

The mods are running this server as if it's a daycare, even though we're all adults.

[stray latch]

Dall-E cannot make Stella from Spiritfarer. Has anyone by any chance done some game art?

[full cloud]

Yes, exactly! It's fun just to spend time together. I love that time of the day when lots of people are just posting image after image, and being creative and sharing with each other. Sometimes we all go crazy and weird and have a lot of fun together. It has been so great!

[acoustic pewter]

Hello everyone.

Any of you have made CHATGPT API call on flutterflow ? I need help on issue

[quartz vale]

#off-topic is probably the best place for your question. This channel is for Dall-E 2

[full cloud]

I think that they need 11 stars, because the first starboard star doesn't count. At least that's what I've heard from others. But that could be wrong.

But yeah, I agree. The Hall of Fame doesn't really matter. What matters is how the art affects ourselves and other people. If my image inspires another person in a positive way, then I don't care if I get likes or not. And no one else should care about likes either. Our goal should be to have fun together. 🙂

[near elbow]

Hi everyone, using Dalle for the first time. I uploaded a photo of my face and gave a simple prompt "Age up face" but nothing happened, it gave me just 5 variations of the same photo. What did I do wrong?

[quartz vale]

I recommend looking through #📚┃resources for guides. Dall-E 2 cannot alter existing parts of an image. Instead, it can only generate new parts of an image

[near elbow]

Hmm thank you, But i've seen people use their own uploaded images and changed parts of it (like the face)?

[quartz vale]

you can erase parts of the image and generate new parts over them. The model won't be able to take your face in an image and make it look older though

[near elbow]

Ok, awesome, thanks!

[stuck zenith]

I agree 100%. For example, I usually only have a relatively short window of time a day in which I can generate and post images. During this time I post pictures, sometimes more, sometimes less. Half an hour is totally overkill.

[full cloud]

2 minutes was overkill. Half an hour is insanity.

[full cloud]

Yesterday at this time there were 40 posts in #daily-theme and it was bustling with activity. Now there are 8 posts... 🤦🏼‍♀️ What a royal mess!

[dense juniper]

It was just so nice to meet each other online at a fixed time. Now it is very quiet, pictures has been posted 8 times by now, which 2 x by me... Not really fun like this.

[full cloud]

It's terrible! They ruined the fun for everyone for the sake of a few people who were dwelling in self-pity over not getting likes. So now those people can have the channel to themselves and remain entitled.

[verbal sequoia]

<:book_icon:1046291503158399107> `` Rule 1 `` Be respectful.

Treat others the way you would like to be treated, and assume best intentions. Don’t harass or attack others, and don’t engage in hateful or generally malicious behavior (e.g. sexism, racism, homophobia, etc.). Keep the negativity to a minimum.

<:book_icon:1046291503158399107> `` Rule 3 `` Stay on topic.

Be mindful of what other users in a channel might find helpful or interesting when posting. Stay on topic in order to keep conversations focused and productive.

[full cloud]

Probably not active enough. They are probably the kind of people who posts only once or a few times each day. So now the channel is dead instead.

[verbal sequoia]

<:book_icon:1046291503158399107> `` Rule 1 `` Be respectful.

Treat others the way you would like to be treated, and assume best intentions. Don’t harass or attack others, and don’t engage in hateful or generally malicious behavior (e.g. sexism, racism, homophobia, etc.). Keep the negativity to a minimum.

[velvet karma]

Your message is off topic for this channel. Please use #chatgpt-discussions. This is an art channel.

[robust compass]

You do raise some good points, but the negativity is hurtful. I thought of the daily theme channel as similar to aiartcommunity where you post your best/favorite and that stars are like votes. I can understand what you are saying, and maybe there could be a general daily theme channel for just free-posting. Personally I avoided putting a star on a multi-image post because I wanted to vote for a single image, not a whole bunch of them at once and honestly I can't be bothered to browse through all the images on some days. (though I am the same way on insta.. I tend to forget or not see that there are multiple images on the post.)

[velvet karma]

You received a timeout because you were engaged in negative comments about other users. Doing so is a violation of Rule 1. Please don't continue with attributing motives to other users.

[full cloud]

I also view it as posting your best images. But when I get creative I make hundreds of images, and post the best of them, and that's true for others too. The multi-image posts are there because of the two minute time limit. Because some images go together as a collection (this is art after all, and that isn't uncommon), and can't be posted with other people's images in between them. Without the time limits, we could post them one by one.

[velvet karma]

The decision to try an increased lag for Daily Theme was made by server staff based on overall assessment of channel activity over the past month or so.

You comment is off topic for this art channel. Please use #chatgpt-discussions.

[velvet karma]

Regarding collections, we offer the #1036148240434860033 feature for that purpose.

[full cloud]

Why did you not put this to a vote first, to see how those of us who are posting would react? You could have easily asked for our opinions and also put up an actual vote by using reactions. Because in the end, we're the ones who would be most affected by it. Does that not matter?

[full cloud]

That's not what I meant. I meant that a few images go together as a cohesive art piece and should be posted together.

[dim cradle]

whats dalle and how do i use it online is there a online version of it

is it not free

[quartz vale]

you can use Bing's AI that's built on Dall-E 2.exp for free. The base model of Dall-E 2 is paid, though

[dim cradle]

i did not get free credits upon signing up. what do i do?

[clear lion]

So, DALL-E works on a credit system only?

[dim cradle]

yes

[clear lion]

It's not like CHATGPT where you can get a membership?

It says 115 credits for $15 which seems pricy but I guess I don't know how much a credit is worth.

[dim cradle]

i don't fully know about that. it has options to buy credits

[clear lion]

Is one credit an image or something? I might have to do some research. Their explanations on the site were lacking.

[verbal sequoia]

[❓] — How do DALL·E Credits work?

What’s a DALL·E Credit?

• You can use a DALL·E credit for a single request at labs.openai.com: generating images through a text prompt, an edit request, or a variation request.

What are free credits?

• Free credits are granted to early adopters of DALL·E who signed up before April 6th, 2023.
• They expire one month after they are granted.
• Free credits replenish monthly.
— For example, if you received credits on August 3rd, your free credits will refill on September 3rd.
— If you joined on the 29th, 30th, or 31st of any month, your free credits will refill on the 28th of every month.

[quartz vale]

please change your profile picture. The current one breaks Rule #6

[dim cradle]

yeah my prev email account has some free credits due to signing up for the waitlist but i was gonna delete that email account. signed up with new email and got 0 credits. shame cause the service is so good.

[quartz vale]

If you made your OpenAI account before April 6, you'll receive 15 monthly credits for free. 115 credits costs $15. Each credit can be used on one prompt that generates 4 images

[dim cradle]

i don't even have a bank account so there is no hope for me?

[quartz vale]

not unless OpenAI decides to re-enact free credits. You can use Bing's AI image generator that's built on Dall-E 2.exp for free, though 🙂

[dim cradle]

okay thanks for letting me know. guess i gotta wait for dall e 3 waitlist then smh...

[main merlin]

Are payments on DALL-E safe?

[robust compass]

I've paid for credits and have not had any issues with it

[naive flame]

Oh, wasn't aware Bing has an image creator powered by DALL·E. Will check it but will stick to OpenAI's DALL-E to avoid MSFT emails. Thx.

[dim cradle]

How can I make Dalle more accurate? Sometimes the drawings are bad or super off, i need a few tries for a decent result

[terse forum]

How can I make Dalle more accurate? Sometimes the drawings are bad or super off, i need a few tries for a decent result

[wicked elbow]

help me draw a cute kitten

[fair crescent]

Go to https://labs.openai.com/, write "a drawn cute kitten", click generate, marvel at your creation.

[fair crescent]

It all depends on the prompt. What are you teying to do, minding what kind of training data is available. You cannot generate a six legged horse, however hard you try. But you can generate a corgi spider hybrid. So it really depends.

[ionic totem]

Hello dall-e is payant no ? but the credit are necessary !

[dim cradle]

Is there a free version of Dalle?

[ionic totem]

No for me it's a problem ?

[stone quest]

Hey guys.

[robust compass]

Morning!

[heady wing]

So, DallE for short movies when? 😉

[quartz vale]

There is a short film that's made with the help of Dall-E 2. It's called Critterz I believe

[quartz vale]

You can use Bing's image generator that's built on Dall-E 2.exp for free

[bold heart]

how do I downsize a picture?

[quartz vale]

you can find image resizers online for free

[bold heart]

can you send me one?

[ionic totem]

I research image in dall-e but the cready are necessary please helpe

[quartz vale]

I'm not supposed to share links outside OpenAI's domain. You can look "image resizer" up, though and find loads

[bold heart]

i mean for outpainting

[quartz vale]

there are outpainting guides in #📚┃resources

[paper minnow]

;/'

[cold hedge]

howto download aI

[robust compass]

Dall-E 2 image generation is web-based- https://labs.openai.com/

[lunar wind]

You're really limited with Dall-E and Dall-E 2 text input. 😦

[sick spear]

DALL·E is also available via OpenAI's API
https://platform.openai.com/docs/guides/images

[mental roost]

Anyone know how to make a poster with ai dall e doesn’t work?

[dim cradle]

Why i cant use dall-e

I never used it before

[raw tree]

#community-help

[bright holly]

https://www.bing.com/images/create?FORM=GENILP

[dim cradle]

Im trying to use it but the images it generates are in a different language, not english. Ive asked it so many times to change it to english and told it that its not english and its not doing it lol. whats wrong?

[verbal sequoia]

[❓] — Generating texts

DALL·E 2 is not engineered to properly generate text. While it may get a few simple words right now and then, that is the exception rather than the rule. If you need to add text to your image, consider using an external image-editing software.

[quartz vale]

this explains the issue with trying to generate text ⬆️

[dim cradle]

very new to dalle how can i generate for free?

[fair crescent]

How "doesn't it work"? Have you tried ending the prompt with "poster", or something like "movie poster", "album cover" or the like?

[prisma mason]

same here

[fair crescent]

Checkout the Bing pictures creator. A single search engine away 🙂
It uses DALL·E experimental. This is mostly regarded as superior, though the pictures are mere KB instead of MB.
@dim cradle

[true junco]

jopa

[atomic patrol]

hello

[dim cradle]

Gpt be liek

[jovial hollow]

Hello, So am creating a discord bot mainly for chemistry and it can reply to some personal questions. What am trying to do is to connect it to google search or bing search (Any search engine or website) to upgrade its character (So it can reply to nearly everything) i tried using API (I dont understand it at all) but it didn't work, any help would be appreciated. I use Replit for programing the language is Node.js

[fair crescent]

I think you're better served at the #chatgpt-discussions

[last minnow]

Picture of a lofty royal palace, with detailed meaningful patterns on the walls, project management rules carefully written on the walls, the palace is lit, with a mountain, slightly dark cloudy sky, 4k, high quality, the soil is golden sand and crisp flowers .

Picture of a lofty royal palace, with detailed meaningful patterns on the walls, project management rules carefully written on the walls, the palace is lit, with a mountain, slightly dark cloudy sky, 4k, high quality, the soil is golden sand and crisp flowers .

[fair crescent]

See the automated message below 🙂
Otherwise check this link: https://labs.openai.com/

[woven lava]

its free, with monthly point limits

[proper gust]

Cosmonaut herman titov, neon world, cartoon style, 4:5

[hard mirage]

so in order to use dalle you need to buy $15 USD?

[midnight burrow]

is dall-e able to do proper pixelart?

for now I've seen some tries, but everything had some smooth parts and/or changes in size of pixels

[sick spear]

Hi, you can use DALL·E on the web with https://labs.openai.com or Bing Image Creator. DALL·E is also available on OpenAI's API at $0.020 per 1024x1024 image (lower resolutions available). More information is available at https://openai.com/pricing 🙂

[real pond]

Hi. I want to see the prompt from my history. (I've forgotten the prompt I've used.) Is there some way to find that? I'm clicking around but don't see it.

[quartz vale]

at the top of the page: https://labs.openai.com/ , click on history and then click on All generations and scroll

[inland pivot]

So, i was attempting to use Dall-e to make a drizzt model for a manga i am making of Forgotten realms. I keep getting the weirdest crap.
the prompt I used was, Drizzt: lithe, muscular drow with obsidian skin, flowing white hair, and intense lavender eyes. He wears black leather armor studded with silver, a purple cloak, and wields two deadly, jewel-adorned scimitars. His striking appearance is marked by a scar that runs from his forehead to his left cheek, a testament to the battles he has fought. LooK at the weird crap you get when you try this.

[dim cradle]

how can i get dall e free

[quartz vale]

you can use Bing's image generator that's built on Dall-E 2.exp for free

[dim cradle]

ok

is it on discord

[quartz vale]

it is not

[dim cradle]

oh is there stuff like that on discord

[quartz vale]

there are, but that's for #ai-discussions as it doesn't pertain to OpenAI

[dim cradle]

ok ill ask there

[teal comet]

i was thinking as we keep prompting, and saving images that we like, we are also training the ai to become good at prompting

[hushed egret]

is there any plans for dall e 3? The competition is absurd. I'd love to see a MidJourney-like levels of out/inpainting

[elfin trail]

i dont understand where can i generate my pics

[merry crescent]

Help

[robust compass]

unknown, but Bing Create offers an advanced version of Dall-E 2

[robust compass]

replied to you in another channel on this question #❌┃prompt-help message

[robust compass]

what do you need help with?

[robust compass]

I don't know if that is how it works, but that would be interesting

[dim cradle]

buss!

[teal comet]

same idk, but i think there's no way its not happening on some level. as we keep repeating certain prompts or terms, ai must be learning what the user wants, and possibly even forming connections between them. and then again, you're saving some images, looking at them for a longer time while paying no attention to some.

[merry crescent]

How can i use the ai

[dim cradle]

Is dall-e 2 available to use? I follow the link from https://openai.com/product/dall-e-2 but it takes me to dall-e.

[quartz vale]

That takes you to Dall-E 2. The first version was never public

[robust compass]

go to https://labs.openai.com/ and make an account, or look up Bing Image Creator which runs on an advanced version of Dall-E 2

[glass blade]

Hi. I've used dall-e before to generate image but seem to be now having a brain fart, and figure out where to go to do that.

[teal mountain]

open ai = discord

[robust compass]

Hi! there is a link at the bottom of this chat to access the DALL-E 2 page

[serene ridge]

nobody uses Bing. lol

garbage

[fair crescent]

For free ai picture creation, everybody uses it.

With a search powered by chatGPT, their share of searches conpared to google have skyrocketed. I have no more love for bing than you do, but I'm not ignoring leaps in technology because of it.

[floral lantern]

BHAIII BOARD EXAMS E 95% ESECCHE BHAIII!!!

[dim cradle]

is chatGPT openai going to be free forever ?

[heady wing]

It speaks volumes about how well Microsoft did with Explorer and Bing that I still feel… uncomfortable? Just using them.

[fair crescent]

No doubt. I've so far only used the picture generation. But the new functionality in searching is still a major advancement I cannot overlook, and keep thinking I should start using.

I think it is a testament how well bing is doing that I'm even considering it.

[wintry bay]

Hello fam!

I am looking for a good resource (a long post, a course, ...something that more or less covers everything in an structured way about AI image generation (schedulers, samplers, learning rate, upscalers, hacks...etc).

Is there any worth mentioning?

Thank you!

[summer beacon]

is it free?

oh nvm it isnt

[robust compass]

@summer beacon you can generate images on Bing Image Create for free. It uses an advanced model of Dall-e 2

[summer beacon]

thanks there

[kind eagle]

I found a big bug in chat gpt which allows you too use the bot for illegal activity

[fair crescent]

Best said in the #chatgpt-discussions or other chatgpt channels 🙂

This is DALL·E, the picture generator.

[kind eagle]

Ok

[fallow solstice]

hi gyz

[fallow solstice]

hi
what is happening madly badly with DALLE-2
i just create my geniune account yesterday and then whoot i did not even use a bit but from start my dalle-2 credits shown 0 ??? :(( why ><

[brittle hawk]

/char f

[echo berry]

how to generate art using Dall-e?

[robust compass]

You can check here: https://help.openai.com/en/articles/6399305-how-dall-e-credits-work which says
What are free credits?
"Free credits are granted to early adopters of DALL·E who signed up before April 6th, 2023."

you could also try the Bing Image Creator (which uses an advanced version of Dall-E 2)

[robust compass]

There is a link at the bottom of this chat channel to be taken to the Dall-E 2 site

[fallow solstice]

my problem
see this link

[fallow solstice]

this is my dalle-2 problem shown i did use my local number and yahoo mail then i create an account and i signup from where for example "i search google DALLE-3 signup" then i directly go on dalle-2 not else where after that from there i create my dalle or openai whatever account directly ok then after that immediately when dalle-2 wellcomed me then credits shown 0$
why i believe i believe its their website error bcz openai website from start are always buggy 😦
plz i need my credits dalle-2 team T_T i did all thing systemically...

help me

[quartz vale]

When did you make your account?

[fallow solstice]

last night!?

new account

and and

[quartz vale]

That's the issue, then. Accounts made after April 6th, 2023 don't receive free credits

[fallow solstice]

i heard by someone that DALLE-2 closed free trial 50 credits system after end of april ??

[fallow solstice]

WT... huhh what is running in their mind why they made cool things less shiny thts very bad decision openai not everyone has rich family now its mean from now new account always got 0 credits if they made genuine even >?? huhh tell me plzz explain and there is no way to use 50 credit free tiral option anymore just need to BUY credits from start when create and finish any new account ??? hmmm

😦

M i right guess ?

😮

[quartz vale]

There is not a way for new accounts to receive free Dall-E 2 credits. You'll have to pay. You can use Bing's image generator that's built in Dall-E 2.exp for free, though

[fallow solstice]

haha wow thts awkward you just remove my authentic and reality base comment of future hmmm is that so scary well anyhow -_-

[quartz vale]

You discussed products that aren't made by OpenAI which broke Rule 8. We can talk about them in #ai-discussions, though

[wispy badger]

I have signed up for Dall-e before April 6 but the problem is I used my Institute Id and in 1 month i will be graduating and i will loose access to my email id so is there a way to transfer those credits or to link my personal account together

[worldly gale]

era bait

[hexed basin]

Login to your account and from account settings change your account email

[wispy badger]

The problem is that I used google signin so the only options i get are these

[next scarab]

I have a question, they said that I had some free credits but when i tried for the first time they said that I have none, does someone know how to activate them?

[frigid perch]

Da

[mellow aspen]

I don't think you're allowed to post nsfw images here

[subtle ferry]

#rules Rule 2.

[mellow aspen]

It was a response to someone who had posted a nsfw image

[subtle ferry]

Ah, gotcha.

mb.

[dim cradle]

midjourney is no longer free to use?

[subtle ferry]

No, also, ask in #ai-discussions. This is DALL-E.

[edgy pine]

Why do I feel that Bing Images has lowered the quality of its images?

[feral furnace]

huh

[undone stream]

Is it possible to use dalle to build out product shots for a shoe / website? Is it capable yet?

[remote mountain]

It can make images of different types of shoes. But you won't get images of actual existing models of shoes.

[undone stream]

So I can use my image of our shoe as a base image and create artistic shots with it?

[remote mountain]

Something like Photoshop would be better suited for adding things via filters.

Dall-E 2 could in-paint your image, but not do a style transfer to it.

[undone stream]

Ya like use my shoes in an art scene you think?

[remote mountain]

In-paint means to change certain aspects of the image. Like change the mouth or eyes in an uploaded image. Or add in a tree where there wasn't one.

You should be able to something like that by out-painting from your image. I haven't experiment much with either in or out painting

[undone stream]

I’m about to go down the rabbit hole… need to figure out how to make beautiful product based artwork for website / marketing

Shhhh

[remote mountain]

Taking your own pictures wouldn't be a bad idea. Most cell phones have pretty good cameras nowadays. With good lighting it would be acceptable imo.

[limber granite]

HII

[torpid sigil]

Use bing.com it uses dall-e 2.5 and its free

[clear lion]

Are there any tips for this thing? I don't know whether to use all 150 characters to describe an image or make it shorter. Most seem to be short.

I am not quite clear on how to interact with Dall-E in that manner

[dim cradle]

Awesome

[minor crest]

Anyone if I can use generate image on my website or I have no rights?

[neat ferry]

a cat

[sick spear]

Hi! You have commercial rights to the images you generate with DALL·E 2 🙂

[undone stream]

I have very hi res pics

[remote mountain]

Personally I'd take those pics of yours into either Photoshop or a free offering like Gimp and go from there.

[subtle ferry]

Ask @shadow bane in DMs my guy.

[celest gate]

help

whats dall e lol

[errant vortex]

[subtle ferry]

https://labs.openai.com

AI image generation tool, it's awesome.

[celest gate]

I'm going to have a bunch of fun with this.

In Design tech.

No more hours of manual drawing.

[subtle ferry]

Using it for design tech is genius dude, it's not gonna be perfect gens but great for mockups I'd say.

#3d-renders and #things-that-should-exist are great examples of that.

[celest gate]

I have access to university detection software

So I can make all ChatGPT text undetectable.

Same with DALL-E.

Therefore making life so easy.

[subtle ferry]

I wouldn't really recommend passing off ChatGPT and DALL-E stuff as your own work. You'd just suck at the skills when you go to get a job, and besides it breaks the TOS.

[celest gate]

Darn it.

I don't wanna break the ToS.

But it really does provide a helpful image when studying, especially in the exam season.

Sums everything up into one small piece of text.

[undone stream]

what do you mean and go from there? the point is i want to upload a pic of my shoe and have really cool pictures made with it

[subtle ferry]

Best use it as a reference or baseline.

I get my key points from it, ask it for a structure, and then follow that structure. Best to ask about that sort of thing in #chatgpt-discussions though.

@wispy storm check it, some of my gens which I didn't want to put in the DALL-E channels.

[wispy storm]

This is epic

I don't know if it can do character like dnd book style

[frigid depot]

need prompt very badly

[jovial junco]

Howdy

I'm looking for some old friend.

[dim cradle]

Hello all, glad to join this community! 😉

[teal timber]

Hi

[signal thistle]

R u ?

Fine

[lapis stag]

hey uhhhh

is Dall e a TAD bit racist?

[fast vector]

Any news on when they will release the next version of Dalle that is currently on closed beta? The current version is absolute garbage (sorry) next to other models including stable diffusion

[fair crescent]

Have you checked the experimental version on bing? It is DALL·E and free. There's a few differences with the DALL·E closed beta, but for most that will not matter.

[fast vector]

I didn't check, I tried it already, but didn't know it was the new or a newer version. I asked because I want to use it as an API, which I don't think is possible with the Bing version, but it already helps, Thanks

Oh, just tested a prompt that I used in the labs.dalle on the bing dalle, it's night and day difference and it's almost as good as Midjoruney

Much better indeed

[fair crescent]

Currently it is unknown when the experimental version will be rolled out. We still see many changes happening, most for the best. Any deployment date is known only to the team, which might still be in true experimental fase without a deployment time in sight.

[raw tinsel]

Hey guys what bot generates images

[grave pulsar]

Do not type in the prompt "A logo that sells goods from Nigeria", worst mistake of my life

[dire summit]

Rapid Research

[dim cradle]

@vivid stag yo bro

[heady wing]

The daily theme seems to have taken a big hit since the 30 min timer has been introduced. The goal was to have people post their best there which would mean a higher average star rating. The amount of stars given out seems to have gone down, showing less people engaging with the daily. Perhaps some experiments should be done to try and find the sweet spot for where people still engage with it but do send their best? Assuming that is even achievable and stifling people’s creativity in the daily is a good idea at all?

[heady wing]

Only people who wear hoodies or are darker skinned go with technological progress and use AI?

[wispy storm]

Dalle

[digital radish]

hhi

[tidal lagoon]

hello everyone the ChatGPTV4 api is free or i need to buy?

[stark rapids]

can anyone please tell how to prompt a image cuz I am a new guy

[supple tide]

Hi, will you still get 50 credits on login in dall-e?

[subtle ferry]

Nope

[haughty mirage]

hiyaa..

[steady yacht]

Hi, I have subscribed to dall e since it published to general public. I used to have credits and now I see there are no credits

[proud ivy]

hello ¯_(ツ)_/¯

[distant barn]

openai API key doesn't work at this time.
Is it stopped?
Somebody help me.

how to use api to generate image?

[dim cradle]

How good is Dall-E? It's quite expensive but could be worthwhile if it gets it right on the first try. Is there any kind of trial?

[spark bear]

a

[fair crescent]

Try the bing version first. It is DALL·E experimental version and generally agreed that it's better. It is also free.

[rare sundial]

https://platform.openai.com/docs/guides/images/usage
This explains the API calls

[coral topaz]

Create a jail environment where Facebook, Instagram and YouTube icons are as prisons inside.

[sonic jacinth]

I Agree With You, racial and religious. Very bad!😶

[sick spear]

Hi @sonic jacinth, our gallery channels are only for images created with DALL·E, available at https://labs.openai.com or https://bing.com/create. Other models can be discussed in #ai-discussions, as per Rule 8. Thanks!

[flat mantle]

Hey guys, can DALL-E also generate bird's eye views of sceneries for my next RPG campaign?

[shell trout]

Is there a tip on getting dall-e to generate text as I have it written?

[teal timber]

I find it can do okay with short words but otherwise it usually doesn't work ever

[verbal sequoia]

[❓] — Generating texts

DALL·E 2 is not engineered to properly generate text. While it may get a few simple words right now and then, that is the exception rather than the rule. If you need to add text to your image, consider using an external image-editing software.

[quartz vale]

text FAQ ⬆️

[fair crescent]

It really depends on what you want. If you want a true top down view, you can try some things with the "top down" keywords. Otherwise "birds eye view" will suffice

[flat mantle]

thanks a lot, that'll be helpful!

[median mango]

draw me a red lion in the deset of africa with 4k quality

[quartz vale]

you need to follow the bot's "Generate Image" link

[fierce flicker]

I am piecing together an immersive exhibit installation that I will need large photographs for. These photographs show first person perspectives of violently charged social encounters from domestic violence to hun violence and more. The immersive experience is meant to heighten empathy in groups that feel these situations are not a huge issue, or doesn’t get the attention it needs for change. Can DALL-E2 create images based on these topics? Or it against the standards in its programming?

Gun violence***

[quartz vale]

Those topics will likely trigger the safety filters for Dall-E. I would advise not using the model for such images as you may be disappointed in the outputs

[modest spade]

i love among us

[remote mountain]

You can't create images on this Discord.
Go to https://labs.openai.com

[still valve]

is it possible to only improve a specific thing of the image without changing everything else if you give it a reference image?

[teal timber]

Yes you can inpaint with the editor. Delete the part you want changed, prompt it to fix it

[still valve]

I only want to improve the chains and the padlock, wouldn't it be kinda hard to remove everything else?

[teal timber]

Nope. Upload it to the editor and erase those parts. Prompt it to fill in like "A portal with chains and a padlock covering it"

[still valve]

so I should use this as a prompt image?

[teal timber]

What's your goal?

[still valve]

as I said before, to only make the chains and the padlock look better while the portal looks exactly the same

because I have one portal without the chains already

my prompt was like "A portal with chains shaped like an X and a padlock in the middle"

[teal timber]

So like this you could delete the parts with the chain/padlock

it'll fill in the background tho

So give it a white background or something if you dont want that

[still valve]

OH

that's interesting 🤔

hmm not sure if that's doable in stable diffusion 🤔

[teal timber]

Well we discuss Dalle here not SD so not my field sorry 😉

[still valve]

aaw ok

but doesn't Dalle and SD use the same AI?

[teal timber]

Nope!

[still valve]

oh 🤔

[mild wave]

Hi, still no release date for exp2 ?

[sleek ibex]

Hello

[lone meteor]

hello! how can i conntact with dall-e support?

[upbeat storm]

Hey! ive been doing some digging and testing but without much success. So i have 2 images (a female top part and a dress) i want to use dall-e to generate full body and her wearing the dress but im not sure what would be the best way on how to achieve this.

[fair crescent]

Have you checked the inpainting/outpainting and how to use it in the #📚┃resources ?

[warm sand]

a

[floral lava]

hellow

[gloomy thicket]

hellp

[rapid carbon]

in real or fake if its fake can i crop the image to remove the watermark?

[fair crescent]

Yes!

[rapid carbon]

thanks

[fair crescent]

Anything to make it look more real is ok 🙂

[raven mango]

Sup designers

[plucky phoenix]

hey is dall-e down or is there a lot of stress on the site rn? it will not generate anything for me

[strange pulsar]

ello

[granite raft]

Just signed up, are free credits still a thing? I didn't get any

[pliant spindle]

No it is not a thing.

[quartz vale]

Accounts created after April 6 do not receive free credits

[rugged patio]

Hi, good afternoon, I receive a message, that my account was blocked. does any have been received this??

[harsh socket]

logo

[open sky]

is this server boosted?

[fluid cliff]

Is there a way to make specific objects a certain color or is it just random when u add object colors?

[quartz vale]

https://discord.com/channels/974519864045756446/1021929896185970708
This explains color allocation issues

[fierce flicker]

Thank you for responding! It’s pretty hard so far to find an AI image generator that can generate the content. Looks like I’m going to be going with the old fashion location scouting and staging for actual photos 😂🤣😂

[empty axle]

I'm sorry, but I'm confused. Is the open AI "dalle page" actually dalle 2? Or is dalle 2 not avail right now?

[sonic jacinth]

There are two links that take you to two different websites:
https://openai.com/product/dall-e-2 (DALL E 2 is an AI system that can create realistic images and art from a natural language description)
https://www.bing.com/images/create (Image_Creator with DALL E Technology)

[empty axle]

Ah, so do I understand then that even as a open AI plus member, I still only have so many credits with dalle?

[sonic jacinth]

!Of course! 10 credits are enough to generate more than 10 images times 4 (10 x 4=40) What a Gift Microsoft! you should thank him!

[empty axle]

Wait that's 40 images some how? If this info is listed somewhere I didn't keep up with, I apologize.

[quartz vale]

ChatGPT Plus doesn't have anything to do with Dall-E 2

[verbal sequoia]

[❓] — How do DALL·E Credits work?

What’s a DALL·E Credit?

• You can use a DALL·E credit for a single request at labs.openai.com: generating images through a text prompt, an edit request, or a variation request.

What are free credits?

• Free credits are granted to early adopters of DALL·E who signed up before April 6th, 2023.
• They expire one month after they are granted.
• Free credits replenish monthly.
— For example, if you received credits on August 3rd, your free credits will refill on September 3rd.
— If you joined on the 29th, 30th, or 31st of any month, your free credits will refill on the 28th of every month.

[quartz vale]

This explains credits ⬆️

[sonic jacinth]

In Dall·E there are 4 images per Credit. Dall·E 2 is also from Microsoft, you should apologize then!

[quartz vale]

it's from OpenAI, not Microsoft

[empty axle]

Ah! That's the 4x thing your talking about. Thank you for the clarification.
And as for Microsoft vs open AI, it's so hard to follow anymore, everything is moving so fast!

[rotund mortar]

cking

[meager thicket]

hallo everybody! I'd like to test DALL-E but I have no free credits. I registered on openai before 6th april, so I thougt I can have few monthly free credits for testing.... how can I understand why I don't have free credits? Anyone can help me?

[verbal sequoia]

[❓] — How do DALL·E Credits work?

What’s a DALL·E Credit?

• You can use a DALL·E credit for a single request at labs.openai.com: generating images through a text prompt, an edit request, or a variation request.

What are free credits?

• Free credits are granted to early adopters of DALL·E who signed up before April 6th, 2023.
• They expire one month after they are granted.
• Free credits replenish monthly.
— For example, if you received credits on August 3rd, your free credits will refill on September 3rd.
— If you joined on the 29th, 30th, or 31st of any month, your free credits will refill on the 28th of every month.

[fair crescent]

Open ai has multiple services. Did you get an account with only chatGPT for example, then nothing has been noted for DALL·E. Only the date when you started to use DALL·E with the account should count (or at least that's what I understand).

You can use the bing picture creation mode. They use DALL·E experimental, are free, but a lower pixel resolution. For most purposes this won't be a problem.

[meager thicket]

thanx anyway

[dense juniper]

Is Dall E planning to have a booth at the IBC? Saw that a new category AI could be filled in as an interest 🙌

[dim cradle]

ok

[orchid canyon]

i signed up before april 6th

[fair crescent]

For chatGPT or for DALL·E?

[dense juniper]

Ok, nice to have the free credits and see what the difference is with Bing Dall-E. I can see the picture when I download it but see a question marks on the website.

[rotund fractal]

i want to apply for discords moderator

[quartz vale]

#spotlight message There's the link to apply 🙂

[dense juniper]

No booth on the International Broadcasting Convention 2023, they will cover some AI I wonder which companies? Could also be ChatGPT as they possible can come with translations of video for the subtitles. Should I ask the question in ChatGPT?

[sick spear]

#ai-discussions might be best for this

[timber bobcat]

is the website down?

[urban gate]

hi \

2+2

[sonic jacinth]

Life is an Event, enjoy the road, at the end of the road you will realize that it is short!

[ashen sierra]

hello evrybody

[jovial jungle]

when my discord bot prints the request made to the DALL E-2 API, the link does not embed.

rather, it looks like this. I'm aware that this format is not a conventional format discord takes, so how could I convert it for the image to show? I am using discord.py.

[fluid cliff]

y i can no put images from the dalle app in the images area?

[spring badger]

wow that dall-e image generator is immensely disappointing

@fluid cliff I think it's only text to image

right?

why does that bot keep popping up and taking up so much real estate on my screen??

[robust compass]

You might try https://discord.com/channels/974519864045756446/1037561751362863144 for help? Not sure this is the proper channel for that.

[dim cradle]

@cinder pulsar

[regal elbow]

Hi,who can help me?I'd like to test DALL-E but I have no free credits. I registered on openai before March,I don't know why I have no credits.

[lyric tulip]

sample image featuring a construction worker dressed in full construction attire, standing near a dusty construction site. This image captures the essence of a bustling construction environment.

[regal elbow]

Please help me ,I wanna credits to creat my image,how can I do?

[fair crescent]

Subscribing to open ai is not enough. You need to subscribe for each of the services. Fir example, if you have been using chatgpt since March, but only started with DALL·E today, you'll be a member of DALL·E from today with all those consequences.

You can use /daily for 4 free credits each day. These credits will disappear after a day if not used.

[regal elbow]

Thank you my friend,how to subscribe for each of the services which I need,and how about be a member of DALL·E

I am Chinese,how to buy credits,there no way for me to pay

[fair crescent]

You've started using DALL·E, so you're 'subscribed'.

[fair crescent]

That I do not know.

[regal elbow]

oh,okay,thanks anyway

As you said,if I will have free credit from today but not now?

[fair crescent]

Try /daily, fill in your email, find 4 free credits for today on your DALL·E account.

[thick oar]

What is discord

Can somebody please help me what is discord i just signup to see what is this

[livid lantern]

Is dall-e getting as good as midjourney? I have gpt-4 and would pay for a dall-e that is on the level of midjourney

[timber shale]

I have a member of my org that lost their free credits when joining my org. Once they removed themselves from my org the free credits became available to them again. Is it possible for them to be part of my org (to use the API) while keeping access to their personal free credits or would they need to setup a separate account?

[teal timber]

I think they can make their own org and also join yours. Then they can select between orgs for different use cases

You can select an org in API call headers too

[timber shale]

Yes, that seems to work for the gpt api (there you can choose the org). For some reason in dall-e we didn’t find a setting to change the org an it seems that they get my orgs (non existent) credits and no longer access to their free credits. Let me know if there is a setting hidden anywhere in dall-e.

[teal timber]

If you go to the playground page you can select organization that you're currently using and I BELIEVE that carries over to labs

I could be wrong never tested

[dull sinew]

Hi all, noob here.. so excuse my ignorance, where can I access or create my API key for DALL-E?

[teal timber]

If you want to use the API, you can make keys https://platform.openai.com/account/api-keys but if you just want to use the website then its https://labs.openai.com/

[bleak vapor]

Hey how can i get my free Credits i got my account new but dont have Credits

[pliant spindle]

Those credits were for early testers. It doesn't exist anymore.

[bleak vapor]

Ah okay

[timber shale]

Hmm, wasn’t able to find an option to select the organization on the playground page. Just under “Manage account”/“API keys” but that setting doesn’t seem to have a y effect on dall-e

[teal timber]

When you click your email top right it drops down with a list of organizations. You can select the one you want active. it works to choose which org you use for playground API.

[void briar]

Anyone have any good prompts for game character designs?

[dull sinew]

Ah okay, I just wasent sure if there was a different API area for DALL-E specifically. thanks!!

[vast falcon]

anyone have experience with shap-e? I can't find any channels here for that yet :((

[craggy olive]

Thanks!

[teal timber]

I keep meaning to try shap-e

[vast falcon]

I'm having so much trouble trying to figure out how to leverage the api for usedcases ;-;

[rich star]

Help

[lavish anvil]

help

ok

help

[dusky gorge]

for some reason me ChatGPT says "Sorry, you have been blocked" anyone with the same issue? i did not do anything to be blocked.. Few days ago was working fine and now trying to use it, saying been blocked

[fair crescent]

I suggest you check at #chatgpt-discussions . Here we only discuss DALL·E.

[sharp widget]

new app can’t login with my mircosoft account（sad

[viral spoke]

Can you generate a photo where an athlete runs through with bears?

[timber jungle]

for DALL-E free credits does it require some explicit sign up? I've had my account with OpenAI for about 2+ years. not sure if that count as early adopter for DALL-E

[fair crescent]

As per the FAQ, the free credits are awarded to early adopters of DALL·E. An openai account isn't enough. Use of DALL·E before the 6th of April is required.

You can use /daily to get 4 free credits every day that expire 24h after requesting them.

Alternatively you can use bing picture creator for free. This uses DALL·E 2 experimental, which is generally considered better, but at lower resolution.

[undone fern]

what should I do to post my generated image to daily-theme using the 4 credit that I just received? I still dont get it, please help

[fair crescent]

If you've generated the image you can download it. There is a download button on the page. Afterwards it is just adding a picture as an attachment to your message. Message is optional.

[undone fern]

oh so we just attach the generated image to #daily-theme manually? I thought we can generate it through discord automatically

[fair crescent]

Not in this case. You might be familiar with some other services where this is very normal. Here you need to generate it via any means you have (api, website, other), download it and attach it manually to a message in discord.

Personally I now prefer this setup. I've tried some Stable Diffusion services, but they do not work nice in discord. I see a hundred messages, none are what I try to create. Even if I find a message for me, it is generally that people in trial periods cannot generate right now. So I haven't been able to verify the quality on my own, which prevents me from getting any subscription.

Besides, I've seen that people ask for what comes down to illigal images about way too young persons repeatedly and receiving images. I'll stick to openai for now.

[vocal elbow]

does anyone know when or if the dalle plugin is coming to chatgpt?

[vestal charm]

I am getting distorted facial images is that normal?

[quartz vale]

There is a Dall-E plugin currently, so I know it exists. However, it's in closed alpha, and there isn't a way to sign up for access to it. I don't know if/when it will go public

[vocal elbow]

Alright, thanks for letting me know!

[shadow onyx]

Where do I open a. Ticket.

[kindred coral]

Hi! I’ve seen a lot of references to this specific channel that shows [No Access]. I wonder which channel you’re actually referring to? Thanks!

[fair crescent]

It is the resources channel. I do not know why you have no access and am asking for assistance.

[kindred coral]

Huh, maybe the server bot glitched during setup. 🤔
I’ll reach out to the mods. Thanks!

[dim cradle]

come here

[worldly gulch]

ask chatgpt

[patent quail]

How to crash dall-e ?

[dim cradle]

👍🏻

[coarse star]

giv credit

this is why i dont like dallE

no credit

[teal timber]

You can get 4 free credits a day via the daily theme bot

[shell lion]

lol

HUH

[vestal bolt]

A profile picture that expresses realistic stories and mystery

[open pumice]

Test

[fair crescent]

Test successful.

If you want to make pictures, check the link in the bot message below.

[dim cradle]

https://labs.openai.com/sc/ieQk2CF8VYg7Eo89PteQ7yfq

this what is got

[vocal valley]

Making a gift for a phd defense in physical chemistry. Someone want to improve my prompt? 🙂 Should I use (way) less terms?
A coloring book page of a self-propelling colloidal finite state machine with huge googly eyes, fast microswimmers that imply rotational diffusion. Science. Lab-on-chip devices. Microgels that are responsive to light. Polystyrene dumbbells. Multi-microgel colloidal clusters. Molecular robotic systems. Formal logic. Temperature, light, electromagnetic signals.

[vocal valley]

Dall-e coloring pages are generating too much fuzzy detail. Major lines only needed. Will try to adjust the prompt accordingly.

[covert storm]

is there any "interface" to dall-e (besides the labs) where I can enter my API key and use it to generate images? Something that wraps the API, where I can use my API key and input the prompt and get it to generate images

[fair crescent]

Consider the training data. It is unlikely that the training data had "microswimmers" that you want. If they are in the training data, they are probably few in number. Self propelling collidal finite state might also be a hard one 🙂

[tacit fern]

Oh god the OpenAl ChatGPT app is coming

[quiet relic]

A majestic dragon soaring through a vibrant, colorful sky, in the style of John Howe and Alan Lee, with intricate details and a sense of motion

[timber jungle]

what is /daily?

[shell lion]

jhbjhb

g

[royal fulcrum]

Hey there. I'm wondering if there is a way (besides having registered before April) to get monthly free credits ?

[sick spear]

Hi, there are some free credits involved in #daily-theme 👀 Check the message there for more info.

[fair crescent]

In discord you can start a message with / and see a list of commands. The command /daily can be selected in most DALL·E channels. After selecting you'll be prompted for your DALL·E account (the email address). If you fill this in, you'll get 4 credits for free that expire every 24h. The daily can be accessed every midnight UTC iirc.

Also relevant for @royal fulcrum

[ruby ginkgo]

hmmm so ive been using DallE on bing vs the open ai site
and for some odd reason my Bing creations are much more accurate ... is it because it considers my history (within bing dall-e)

[sonic jacinth]

butterfly-technique_🐶 🦋

[fallow egret]

[dim cradle]

Is there a Dalle trial? I only need to make one logo..

[restive hearth]

https://bing.com/create

[dim cradle]

Bing doesn’t do text/words.

[restive hearth]

I'm sorry to hear that. Though I recall some simple words showing up in my own experience

[sonic jacinth]

📚 💻 🎨

[fair crescent]

It is because bing uses an experimental version, which model has been trained further in many ways. It is better at larger prompts and extra additions and does many things (but not all) better than the other version, like faces. The styles are more and generally seen as better.

[dim cradle]

Do you have Dalle?

[fluid cliff]

There should b a way for dalle to use certain keys like -/+ so its able to pick up when u want something a certain color. And yes ive read the colors thing, im just saying like if i want like the ball red i could put -red ball- and cube wouldnt b affected. Wonder if thats possible

[wispy storm]

You can't gen picture here

[minor frigate]

this might be asked several times but will dall-e receive any update on how it generates image to increase how "realistic" and well it looks

[teal timber]

bing uses the exp model which has been pretty great in that regard. You can head to there.

[dim cradle]

@fair tendon

[cyan wing]

Hey, Dale-e is opensource?

[restive hearth]

I use Bing Image Creator "Powered By DALL-E" which is free. I sent you the link yesterday.

[fair crescent]

Nope

[vital mauve]

Where can I post animations made using images generated by Dall-e2?

[fair crescent]

It's better to ask your question first and see if the community can assist. Mods can get spammed easily, so they're less likely to react to a mention without anything else.

[hot idol]

1girl

[proper shale]

spiderman

[dim cradle]

It doesn’t do text on images though. Always a different language. I don’t understand how something so smart can be so dumb

[restive hearth]

It's probably a responsible decision by the developers to hold back that capability at this time.

[slender acorn]

Why would anyone pay for DALL. E credits when Bing is powered by it or are there differences in the outcomes?

[gilded osprey]

That's a good question as Bing is way better.

[frail egret]

Hello, is it possible to get free credits for testing ?

[fair crescent]

Try /daily and fill in your DALL·E email

[storm yoke]

Will there be a dalle mobile app?

[fair crescent]

Make sure the command comes up. After selecting the command, fill in your DALL·E email. Thrn hit send 🙂

[sonic jacinth]

Hi, friend? Your AI-generated image looks fuzzy!

[honest magnet]

3. DALL-E 2

[slender acorn]

How would someone create their own art style using this AI so when using different prompts it would come off like an artist replicated their own art style

[dim cradle]

So is bing image generator better then Dalle?

[slender acorn]

test it out yourself to see

https://www.bing.com/images/create?

[quick wigeon]

Any tips on adding blur effect on top of image so that text can be visible or any other way of doing it

[orchid current]

How do you get the daily command to show up?

[teal timber]

The free credits? Just start typing /daily and it should appear

[teal timber]

Looks like there may have been a bug in the moderator bot that kept images from being posted. Should be fixed to try again

[slender acorn]

What's the differences between Midjourney and Dalle/?

[terse ether]

Imagine cat

[potent oriole]

yhm guys why it's didin't work

why

[last merlin]

nature, sunset, mountain, lake, landscape

[fair crescent]

From the message in the top corner it seems you have no credits. Did you use HTML edit to change the credits in the center?

[shy flame]

does DALL E use VQGAN or GLIDE?

[quartz vale]

I believe it uses GLIDE

[fair crescent]

This reminds me too much of the '3D' engines of old games. Want incredible smoothness and graphical fidelity? Turn on GLIDE!

[lament veldt]

Hi, I'd like to try Dall E but it is asking for credits right away. Isn't there any kind of trial I could do before paying for the credits ? Thanks!

[candid fable]

dd

[sick spear]

Hi! If you're interested, there are some free credits involved in #daily-theme 👀 check the message there for more info.

If you'd like to experiment with an updated DALL·E model, you can try Bing's Image Creator preview for free too https://bing.com/create/

[fair crescent]

#images-discussions message

[restive hearth]

Anyone else tried describing their dreams into DALL-E? I described one and the result blew my mind. It was visually so similar to the real thing.

[lofty arch]

how do I talk to the Ai Bots?

[sick spear]

You might be looking for ChatGPT, which you can use at https://chat.openai.com and discuss in #chatgpt-discussions 🙂

[pliant spindle]

How to generate experimental images on official Dalle 2 website? I currently only have standard.

[quartz vale]

some people have access to the Dall-E 2.exp model on the labs website while others do not have that option. There isn't a way to sign up for it or enable it as it's in closed alpha testing 🙂

[rotund tide]

Hi. I noticed that in Bing's Image creator there is not possible to add following details/changes for generated images (ideas) like in Bing chat itself. Is it true or do I miss something important here? Thanks

[tepid hornet]

Can You Protect Your IP in the Age of ChatGPT?

[fringe orchid]

Hey guys what is the best Free image A.I generator that would be good for Instagram posts. Or do I have to pay?

[fringe orchid]

How much does Bings imagine genrationt service cost?

[rotund tide]

Zero

[fringe orchid]

Really? Do you recomend it for social media posts

[abstract apex]

Hello guys i‘m a fashiondesigner and want just based on my draft designs to make photorealistic runway photos of models wearing my designs. So that i get a better idea which design of mine i should realise. Is that possible? And how?

[alpine onyx]

Hi guys is dalle good with words ?

what i mean as text on the image not as prompt thank you

My question is answered all AI is bad at generating words xD

[teal timber]

It sometimes does okay but yeah, all AI art is bad at words

[devout jolt]

Can we use DALL E images for commercial purposes?

Like if we have to sell a product can we print a DALL E image on it?

[alpine onyx]

I don't believe there is a problem as long as you adhere to the requirement of not falsely claiming to use AI.

[devout jolt]

What do you mean by not falsely claiming to use AI?

[alpine onyx]

Like claiming you created it by hand

and i found the term of use it says "3. Content

(a) Your Content. You may provide input to the Services (“Input”), and receive output generated and returned by the Services based on the Input (“Output”). Input and Output are collectively “Content.” As between the parties and to the extent permitted by applicable law, you own all Input. Subject to your compliance with these Terms, OpenAI hereby assigns to you all its right, title and interest in and to Output. This means you can use Content for any purpose, including commercial purposes such as sale or publication, if you comply with these Terms. OpenAI may use Content to provide and maintain the Services, comply with applicable law, and enforce our policies. You are responsible for Content, including for ensuring that it does not violate any applicable law or these Terms."

https://openai.com/policies/terms-of-use

Dall-e policy: https://labs.openai.com/policies/content-policy

[devout jolt]

Okay Thanks!

[manic garden]

When does it become free

[quartz vale]

you can use Bing's Image generator that's built on Dall-E 2.exp for free 🙂

[oak halo]

Do a ChatGPT app exist?

[quartz vale]

I recommend checking out #chatgpt-ios-discussions

[gaunt sorrel]

Hi all. Very new here. is it possible to use Dall-E to create an image of a "3 Musketeers" candy wrapper but instead it says "Solutioneers"?

[fleet wyvern]

Can I add an image in outpainting that I used firefly from adobe to edit or is that against the Discord TOS?

[quartz vale]

Dall-E 2 cannot accurately generate text, so that will not work

[quartz vale]

in the #inpaint-outpaint channel? We would prefer to keep the Dall-E galleries to images that are 100% Dall-E generated. You can share it in #—post-dalle-edits , though

[raw wave]

Anyone else experience worse quality in image and prompt understanding using the API compared to Bing Image Creator or the DALL-E web app? Prompts are exactly the same and generated at 1024x1024

A photorealistic full-body, front view, product shot of a backpack with teddy bear design, front pockets, brown , professional design, 4k, white background

[teal timber]

Bing uses the exp model which is newer and not a public api. That’s the main difference so it’s hard to compare the two

[fleet wyvern]

Cool thx

Any idea when outpatient/inpaint will be experimental vs just v2

[fair crescent]

I think even the team doesn't know. The experimental version seems to be still in development. If it is known I'll be super excited.

[quartz vale]

I do not have any information about inpainting/outpainting being added to the exp model

[gentle stone]

How can generate new ai image? Need help? Anyone

[robust compass]

Use the Dall-E 2 link at the bottom of this channel (use /daily to get 4 free credits!) or you can search Bing Image Create which uses an advanced version of Dall-e 2

[minor crest]

anyone know how to hide text or add specific text on image creator ?

[tribal echo]

well these images are a bit curse

[wispy storm]

Look like my book in high school

[restive hearth]

Been experimenting with DALL-E capabilities and limitations. It does not understand: Aperiodic tilings (non-periodic tilings), Ammann Bars, quasicrystals, or knots. I can give specific examples of knots (running bowline, double hitch) and the machine clearly does not know what they are. Including scientific scans or images to prompts create interesting effects (LIDAR, petrology/petrography, electron microscopy etc), but one such as: "petrographic image of a carbonaceous chondrite with numerous chondrules in cross polarized light " do not yield a convincing image. The AI does not understand the relationship between between the grains or the relationship of shape and color to the composition and formation conditions of the "material". I wonder if the AI is simply untrained in these subjects or unable to understand them.

[fleet wyvern]

WYM use daily to get 4 free credits?

[fair crescent]

Discord has commands. If you go into the text bar and start with a / you can see which are available. There's commands that are global throughout discord. But it starts with specific channel and server commands if available. Many channels here have the /daily and /current_daily command. Type /daily without sending the message and you'll be prompted for your DALL·E email. Fill in the email and you'll have 4 free credits that expire after 24h.