Quote from: alexandersr on January 18, 2023, 23:09... Also i have used a app called craiyon.com, but the image don't have size for microstock. Do you know some program that could be use for free for microstock image?

i use topaz gigaAI to upsize, which outperforms PS resizing

Quote from: cascoly on January 18, 2023, 20:07

Quote from: ADH on January 18, 2023, 17:37
Adobe, over 500k AI images for sale, 99% will never be sold because there is little or no use for them. By the end of the year, following the actual trend of uploads, it will be 5 millions for sale. When do you think Adobe will start charging for uploading AI generated garbage into their portal?

your source for any of those claims?

guess i'm finally part of the 1% since my AI images have sold on AS, SS and DT

I'm not a 1%er yet, but I expect I will be soon enough. Just sold on SS "Ancient Paleozoic era clams. Brachiopods, with a symmetrical ray pattern" Macro, which isn't something I'd expect that most people have heard of, let alone in an sort of demand? I think a couple of my AI creations have a pretty good chance, as they are much more useful. I admit, not as pretty as some that people have shown, but still, useful.

Quote from: alexandersr on January 18, 2023, 23:09

I have not used DALL-E2, because it ask for a phone number and in my country is not listed (i live in Venezuela). I have used MidJourney but in the Discord chat, because i don't know if that app is in another place. Also i have used a app called craiyon.com, but the image don't have size for microstock. Do you know some program that could be use for free for microstock image?

You might just try, if you haven't already, it just sends a text code to the phone number. Unless your entire country can't use the free site? The phone is only to verify the account and one account per person, or one per phone number that is. Write them, maybe there's another way?

DALL-E / Open AI, 1024, resize the image to 3300 SQ, as they are square. You can work on sharpen or use Giga AI or something else, and reduce to 5MP and try that. Everything needs to be edited and cleaned and corrected anyway. None of this is, give a prompt and get a useful image. You will have to edit and work on them.

As usual, if it's FREE it's on the top of my list, and Open AI is that. 

Quote from: Uncle Pete on January 19, 2023, 19:11

Quote from: cascoly on January 18, 2023, 20:07

Quote from: ADH on January 18, 2023, 17:37
Adobe, over 500k AI images for sale, 99% will never be sold because there is little or no use for them. By the end of the year, following the actual trend of uploads, it will be 5 millions for sale. When do you think Adobe will start charging for uploading AI generated garbage into their portal?

your source for any of those claims?

guess i'm finally part of the 1% since my AI images have sold on AS, SS and DT

I'm not a 1%er yet, but I expect I will be soon enough. Just sold on SS "Ancient Paleozoic era clams. Brachiopods, with a symmetrical ray pattern" Macro, which isn't something I'd expect that most people have heard of, let alone in an sort of demand? I think a couple of my AI creations have a pretty good chance, as they are much more useful. I admit, not as pretty as some that people have shown, but still, useful.

Quote from: alexandersr on January 18, 2023, 23:09

I have not used DALL-E2, because it ask for a phone number and in my country is not listed (i live in Venezuela). I have used MidJourney but in the Discord chat, because i don't know if that app is in another place. Also i have used a app called craiyon.com, but the image don't have size for microstock. Do you know some program that could be use for free for microstock image?

You might just try, if you haven't already, it just sends a text code to the phone number. Unless your entire country can't use the free site? The phone is only to verify the account and one account per person, or one per phone number that is. Write them, maybe there's another way?

DALL-E / Open AI, 1024, resize the image to 3300 SQ, as they are square. You can work on sharpen or use Giga AI or something else, and reduce to 5MP and try that. Everything needs to be edited and cleaned and corrected anyway. None of this is, give a prompt and get a useful image. You will have to edit and work on them.

As usual, if it's FREE it's on the top of my list, and Open AI is that. 

No, DALL-E  looks like is not avaible in my country (Venezuela) , look.
Something went wrong
OpenAI's services are not available in your country.

Quote from: alexandersr on January 19, 2023, 23:43
No, DALL-E  looks like is not avaible in my country (Venezuela) , look.
Something went wrong
OpenAI's services are not available in your country.

That's unfortunate. It's entertaining and fun.

Things like Avatars and logos.

on my blog the last part:
AI: friend or foe? (5 of 5)
A very brief history of art: where will it end?

https://luisafumi-digitalart.com/blog/2023/01/19/ai-friend-or-foe-5-of-5/



enjoy

Artists behind a lawsuit against AI: https://www.buzzfeednews.com/article/pranavdixit/ai-art-generators-lawsuit-stable-diffusion-midjourney

Interesting article explaining how AI works: https://www.washingtonpost.com/technology/interactive/2022/ai-image-generator/

Quote from: Lina on January 20, 2023, 18:32
Interesting article explaining how AI works: https://www.washingtonpost.com/technology/interactive/2022/ai-image-generator/

That is a great explainer.

Quote from: fotoVoyager on January 20, 2023, 18:48

Quote from: Lina on January 20, 2023, 18:32
Interesting article explaining how AI works: https://www.washingtonpost.com/technology/interactive/2022/ai-image-generator/

That is a great explainer.

decent beginner's explanation, but it elides the actual link between dataset & image creation. HOW are individual data stored? HOW is the resulting dataset (w no images) accessed to create the image?

and some info is just wrong
"You have selected a dachshund. After ingesting the images and corresponding text, the AI model begins to connect the words a dachshund to the visual representation of a dachshund."

individual images are not used (if they were, there's no way an image could be formed from such disparate results. ) and the odds of having JFK in the image are infinitesimal! besidees which, JFK was not part of the prompt, and 3 of the 4 elements of the prompt were ignored (besides which DALL-E will not create images of actual people)

a better description is https://nanonets.com/blog/machine-learning-image-processing/#working-of-machine-learning-image-processing

Quote from: gameover on January 20, 2023, 16:56
on my blog the last part:
AI: friend or foe? (5 of 5)
A very brief history of art: where will it end?

https://luisafumi-digitalart.com/blog/2023/01/19/ai-friend-or-foe-5-of-5/



enjoy

interesting follow-up  on Turing test in today's NY Times
https://www.nytimes.com/2023/01/20/technology/chatbots-turing-test.html

Quote from: Lina on January 20, 2023, 18:32
Artists behind a lawsuit against AI: https://www.buzzfeednews.com/article/pranavdixit/ai-art-generators-lawsuit-stable-diffusion-midjourney

Interesting article explaining how AI works: https://www.washingtonpost.com/technology/interactive/2022/ai-image-generator/

It will be of course a long legal battle, but the sad side is that nothing is really done to protect artist.
I mean, it's all about money: Getty sue for AI? They will find an agreement, Stable diffusion, Midjourney and Deviantart will pay some money to them. Original authors will receive a kind mail saying that they are the winner, and they will receive... 0,01 cent for every image used.

It's not really an artist war against AI. It's only an economic war to gain some money for agencies, and give cents to authors

Adobe now has over 605 000 generative ai images. It also has a lot more that were uploaded without marking them as gen ai.

Once the Midjourney app comes out, this will be the fastest growing collection ever. With endless duplicates.

But then...if there is no limit on uploading a plate with spaghetti, why should there be a limit to genai? let the market sort it out.

But it does sell. My stuff from dalle is much simpler than the stunning Midjourney files, but I already have a few sales, which I find really suprising, because I am limited by the square format.

This will be the greatest incoming flood of content in stock history.

Quote from: derby on January 20, 2023, 21:14... Original authors will receive a kind mail saying that they are the winner, and they will receive... 0,01 cent for every image used.

It's not really an artist war against AI. It's only an economic war to gain some money for agencies, and give cents to authors

that's pretty much true of most class action suits, like the visa credits, tmobile etc- awards were in the $5-20 range

what sort of payment would you like to see for use of your image in a training set with millions (billions now) of images used?    for image creation, there's no link to those training images, so payment would have t be linked to the entre dataset - if agencies are mostly rec'ing < $5 per image, what amount would be given to those millions of images?

Quote from: derby on January 20, 2023, 21:14

Quote from: Lina on January 20, 2023, 18:32
Artists behind a lawsuit against AI: https://www.buzzfeednews.com/article/pranavdixit/ai-art-generators-lawsuit-stable-diffusion-midjourney

Interesting article explaining how AI works: https://www.washingtonpost.com/technology/interactive/2022/ai-image-generator/

It will be of course a long legal battle, but the sad side is that nothing is really done to protect artist.
I mean, it's all about money: Getty sue for AI? They will find an agreement, Stable diffusion, Midjourney and Deviantart will pay some money to them. Original authors will receive a kind mail saying that they are the winner, and they will receive... 0,01 cent for every image used.

It's not really an artist war against AI. It's only an economic war to gain some money for agencies, and give cents to authors

True, I agree about agencies, except article I shared is about three artists hiring lawyer and suing AI companies. These artists might represent whole community and just maybe make some change.

Quote from: cobalt on January 20, 2023, 21:36
Adobe now has over 605 000 generative ai images. It also has a lot more that were uploaded without marking them as gen ai.

Once the Midjourney app comes out, this will be the fastest growing collection ever. With endless duplicates.

But then...if there is no limit on uploading a plate with spaghetti, why should there be a limit to genai? let the market sort it out.

But it does sell. My stuff from dalle is much simpler than the stunning Midjourney files, but I already have a few sales, which I find really suprising, because I am limited by the square format.

This will be the greatest incoming flood of content in stock history.

That's Microstock isn't that? A great flood of endless duplicate similar content.

Quote from: Lina on January 21, 2023, 03:53

Quote from: derby on January 20, 2023, 21:14

Quote from: Lina on January 20, 2023, 18:32
Artists behind a lawsuit against AI: https://www.buzzfeednews.com/article/pranavdixit/ai-art-generators-lawsuit-stable-diffusion-midjourney

Interesting article explaining how AI works: https://www.washingtonpost.com/technology/interactive/2022/ai-image-generator/

It will be of course a long legal battle, but the sad side is that nothing is really done to protect artist.
I mean, it's all about money: Getty sue for AI? They will find an agreement, Stable diffusion, Midjourney and Deviantart will pay some money to them. Original authors will receive a kind mail saying that they are the winner, and they will receive... 0,01 cent for every image used.

It's not really an artist war against AI. It's only an economic war to gain some money for agencies, and give cents to authors

True, I agree about agencies, except article I shared is about three artists hiring lawyer and suing AI companies. These artists might represent whole community and just maybe make some change.

they can't represent everyone unless they create a class action suit, which they haven't

Quote from: cascoly on January 21, 2023, 19:06

Quote from: Lina on January 21, 2023, 03:53

Quote from: derby on January 20, 2023, 21:14

Quote from: Lina on January 20, 2023, 18:32
Artists behind a lawsuit against AI: https://www.buzzfeednews.com/article/pranavdixit/ai-art-generators-lawsuit-stable-diffusion-midjourney

Interesting article explaining how AI works: https://www.washingtonpost.com/technology/interactive/2022/ai-image-generator/

It will be of course a long legal battle, but the sad side is that nothing is really done to protect artist.
I mean, it's all about money: Getty sue for AI? They will find an agreement, Stable diffusion, Midjourney and Deviantart will pay some money to them. Original authors will receive a kind mail saying that they are the winner, and they will receive... 0,01 cent for every image used.

It's not really an artist war against AI. It's only an economic war to gain some money for agencies, and give cents to authors

True, I agree about agencies, except article I shared is about three artists hiring lawyer and suing AI companies. These artists might represent whole community and just maybe make some change.

they can't represent everyone unless they create a class action suit, which they haven't

OK, they represent other artist's opinions and many artists are celebrating that lawsuit because they are hoping it will change something for everybody.

Generally I don't mind the existence nor using this technology myself but this is going too far. Every second prompt is using terms like "in style of ARTIST NAME" and as it if that wasn't enough now they started to create models based specifically on artist style to get even closer, and again...as that is not enough you can upload a reference image scraped from the internet and decide how close the generated image will look to the original and balance the prompt and reference photo.

So artist that don't want to be included have absolutely no choice, not even formally.

I have no doubt this will be stopped sooner or later and someone will get compensated big time at the end.

Also, forgot to mention that I personally got signatures in the bottom corners. I really wonder if they are real, and if they are,  it pulls a lot of questions.

First one is , is the technology really just using training to create something new or its directly scrapping parts of existing artwork, because how to justify signatures if they turn out to be real. ( Which I can not claim now, but personally I think its a matter of time someone will start creating art based on his name a get a signature if that's the case)

Microsoft is preparing to add AI content creation to Bing.  Google will follow.  Free and easy public access to synthetic image creation tools could pose a serious challenge to stock agencies business models.  If Bing and Google add usable commercial licenses terms buyers wont need shutterstock subscriptions anymore.

https://www.cnbc.com/2023/01/23/microsoft-announces-multibillion-dollar-investment-in-chatgpt-maker-openai.html

Millions of images can be downloaded legally for free with a cc license.

I still have sales on stock agencies.

Mobile phones have become so good, everyone can take pretty pictures.

I still have sales on stock agencies.

Apps beautify mobile phone images so everyone can have even better pics.

I still have sales on stock agencies.

Now the general public can ask an ai to generate content for them...

....and I will still have sales.

The general public has no visual training, so the results they get from the ai will be worse than doing a quick search on a stock agency.

The reason we make money is because the customer has no time.

If billions of free images everywhere don't kill the agencies, how will the ai do it?

Quote from: Lizard on January 23, 2023, 19:42Every second prompt is using terms like "in style of ARTIST NAME" and as it if that wasn't enough now they started to create models based specifically on artist style to get even closer, and again...as that is not enough you can upload a reference image scraped from the internet and decide how close the generated image will look to the original and balance the prompt and reference photo.

But there is no need of AI to find everywhere any kind of creation "based on someone else previous work and style".
This has always been happening, even in ancient ages, it's not new

Quote from: Lizard on January 23, 2023, 19:42So artist that don't want to be included have absolutely no choice, not even formally.
I have no doubt this will be stopped sooner or later and someone will get compensated big time at the end.

Of course it will happen; Getty, Shutterstock and others will be compensated for giving their database for training; artists will be compensated only by database owner, as SS is doing with "contributor fund".

Artists will probably gain the option to mark their own work as "not usable for training".

Quote from: Lina on January 21, 2023, 19:41
OK, they represent other artist's opinions and many artists are celebrating that lawsuit because they are hoping it will change something for everybody.

I'm quite sure these artists will gain the right to refuse to be trained. That is absolutely due.
There are work in progress to technical solutions, like a universal signature in digital image that refuse to be trained. It's the same of the robots.txt that in a web site stop the search engine scraping.
But this will not stop AI development.

Quote from: trek on January 23, 2023, 21:06
Microsoft is preparing to add AI content creation to Bing.  Google will follow.  Free and easy public access to synthetic image creation tools could pose a serious challenge to stock agencies business models.  If Bing and Google add usable commercial licenses terms buyers wont need shutterstock subscriptions anymore.

https://www.cnbc.com/2023/01/23/microsoft-announces-multibillion-dollar-investment-in-chatgpt-maker-openai.html

SSTK and the rest will be forced to pay us more for "Real Photos" as most artists will drop working when most everything they can do, will be produced by free AI software? Stock images of real subjects will become harder to find and more valuable? (well I can always hope?)

The only problem I see with AI (or SS, as I call it - Splice Synthesis) is the terrible thing it does to human hands. Once they fix that, it's golden.

Oh, and the lack of SSS. Without the SubSurface Scattering all those portraits look like 3D renders from 2003-2004.

Quote from: Digital on January 24, 2023, 18:26
The only problem I see with AI (or SS, as I call it - Splice Synthesis) is the terrible thing it does to human hands. Once they fix that, it's golden.

Oh, and the lack of SSS. Without the SubSurface Scattering all those portraits look like 3D renders from 2003-2004.

And a list of others, clocks come to mind?

I was trying to get DALL-E to generate a photo of a Ukranian soldier firing a bazooka at a Russian tank in winter.  All efforts and variations in wording produced totally unusable results.  Even with a lot of fixing up, I can't see sending results to a stock site.