I'd say if someone chooses to do pictures/footage for any one of the fake narratives going on being pushed by the media tend to be popular, i.e., "climate change"/"sustainabilty/zero carbon" crap/garbage, "ai/security "threats"", "convid", the "gay/trans" agenda to try and discourage caucasian/western people from having children (have you watched any t.v. recently - is excessive on some channels), etc... if you want to see what narrative/storyline is being pushed - just read any of the stories published on the world economic forum "wef"/un/who/etc...

But aside from those narratives... I'd say it depends/goes in cycles, depends on the type of footage you have/etc...

pps, instead of the movie line "release the kraken!" (forget which movie), it could be "release the army of eastindian spammers!" lol. oi vay. once they get word of that... you are probably going to have 10,000 "videos" of different angles of the same person eating an orange lol... and then posts asking what happened to their port because 10,000 videos of a person eating an orange didn't get approved...

Hi Matt -

Would you please discuss perhaps some kind of contributor pool for the new '4k subscriptions', to balance payouts? Basically - now contributors (estimated) get on average get about 1/5th to 1/10th of the sale - since the 'purchase' now becomes a 'subscription' purchase.

Basically - splitting any unused 'subscriptions' between contributors & adobe @ the 67/33% split?

So if a customer (using this #'s for simplicity) gets a $100 subscription that say had 10 credits, and only uses 1 credit - and say had 9 unused credits - of which the contributor would get $3 (and adobe $6.7 + the remaining $90 balance for a total of $97) - could you please discuss instead sharing unused credits from subscriptions with contributors at the same 33%-67% split? (So instead of $3/$97 = 3% to contributor + 97% to adobe, it would be 33% to contributor + 67% to adobe)?

Thanks!

Believable. Need 3rd party auditing to ensure they do things fairly. They've already demonstrated a number of times they are dishonest (i.e., "selling ai" data first, "then" saying here's some peanuts and 'now' you can choose not to participate). Jon originer cashed out & bailed a couple years ago too. The "exciting news" that they restructured payments to make it poorer for contributors, etc. They are dishonest. Wouldn't be surprised if that was the case of what was going on.

This type of auditing by contributors would actually be a very good thing. Share results/youtube videos/etc if indeed that is what is happening. First raise awareness of the dishonest tactic, and then get pressure to get accountability and things 'fixed'.

Saw a few posts about the "4k" videos going 'subscription' model - and it's even a bit worse than one things - so you should push back to have "on demand" clips - or request that adobe allow you to decide which clips are "subscription" based and which are "on-demand". (Of course, the unlimited models are very poor too - just commenting right now on the "shift" from on-demand to credit subscriptions for companies doing credit subscriptions, and right now Adobe is one of the ones being discussed).

It's basically a sleight of hand method of stealing revenue from the contributor, without them really realizing it - with the 'promise' of 'maybe' they can make more, when they don't (in most cases).

Adobe in both models (for simplicity) keeps 67%, gives contributor 33%. But in reality - in the 2nd model, if a buyer only downloads 1 clip - Adobe is actually keeping 97%, and only giving the contributor 3%. Many times buyers don't actually use 'all the credits' in a subscription pack, and may only use 1-2. And in fact - many times they may be purchasing the "clip pack" simply because they want "your" clip.

a) In model (a) if you sold a $100 clip, you get $33. Adobe gets $67.

b) If you have "$100 plan" instead - where someone gets "10 clips". The buyer still pays $100. However, since most buyers don't use the full "clip" - and as I mentioned they may have only signed up to get "your" clip (with the "promise" of more if they wish, which many don't use)  - now you only get 33% of "$10" aka $3, and adobe keeps the rest ($97). (As an aside - obviously the subscription model has an added bonus for Adobe & other companeis - in that people forget to cancel/don't cancel/etc - as well as not using their membership - so in reality that $100 sale became a $300 sale, and you get 1% ($3) while Adobe (& other similar companies) keep the remaining $297).

It is a 'sleight of hand' way of moving more revenue into Adobe's (and other companies that do the same) into their pockets. Effectively stealing revenue from you.

Of course - there are some people who will use the full amount of their pack - i.e., if they buy 10, they will make sure they download 10, but for most customer habits on a subscription plan, (educated guess based on experience) they 'maybe' use 20%.

Let's write and see how we can opt-out of their subscription plans, and make certain clips 'on-demand' only.

@Uncle Pete
Interesting, I didn't know about the token length restriction in Stable Diffusion. Thanks for the info!
I'll try it out with longer, detailed prompts.

Which online provider do you use?
Or do you generate locally on your own machine?

I currently use Mage.Space. Can't complain. For 15 USD / month unlimited generation with private mode and many pre-installed models / loras.

The cool thing is that in combination with precise prompts and special models, you can maintain character continuity quite well without too much deviation.
So it has potential for image series. I haven't seen sofar Midjourney users using this, but maybe nobody thinks about it.

I'm currently trying to generate colorful portraits with African-American models:
https://ibb.co/PhRSgNN
https://ibb.co/PTkRFZZ

And also the current body positivity, female sexual wellness trend:

https://ibb.co/f1sWkQR
https://ibb.co/1J0f0mM

Oh and yoga in a more "raw" style:
https://ibb.co/nLPSjnw

yes, the hands and minor artifacts are still really annoying at the moment ...

But as I said, I see much more potential in Stable Diffusion than Midjourney to stand out through individualization from the huge same-looking crowd. The buyers will be soon so pissed off because of millions of same looking stuff.

Just a commentary on the 'body posivitity' thing. It's funny how the tell-a-vision & media has convinced people being "fat" and "obese" is a "good thing" when it's actually very, very bad, by inventing new words & garbage phrases like "body positivity", when it's actually very, very bad to be that disgustingly overweight and fat. Being fat & overweight is not something to be proud of. It means you need to stop eating the potato chips, refined sugar, canola oil products, get off the sofa, get some exercise, stop taking meds & injecting yourself with crap. It means you are sick, because you have a bunch of parasites & crap inside of you that your body can't get rid of.

Pretty messed up how "they" (the people who own/run the wef/imf/UN/etc) appear to have confused so many people.

If I uploaded a picture like that, I'd be tempted to call it "Fat slob pretending to be happy by being slothful". It's a much more accurate description.

Having done it for a while,

a) Yes, you  can make some income.
b) If you get "lucky" (some of your images take off) then you can make what some would consider to be a good income.

I'm not quite yet sure what the magic formula is for (b) yet... I've seen some people that have images similar to mine really take off, and no idea why my nearly equivalent ones didn't... As far as I can tell, I have good keywording, good titling, good images... So I don't know - it seems to be a little bit of luck/right timing/etc to get an image to take off. I did personally have a couple that did relatively well, but others, not yet as well...

So... if you are playing around with it, sure, go for it. But... if you are looking to make more than a couple hundred $$/month - you probably have your work cut out for you.

I actually use ChatGPT currently to create prompts.
I usually get inspired by an interesting photo/image, write a brief description or some keywords of the content/topic and then rephrase it via ChatGPT.
This saves a lot of time and increases productivity.
Sometimes I even create new and better image compositions.

My perception is also that the more extensive the prompts, the more detailed images can be created with a higher prompt guidance scale in Stable Diffusion.
It's crap to use keywords like "higly detailed", "photorealistic", blablabla, etc. in prompts.

As Evaristo tenscadisto has already written, it makes sense to create templates with short, specific instructions.
Sometimes, for example, I only paste in 3 to 5 keywords and create a detailed prompt because ChatGPT knows from previous example descriptions and specifications how I would write the prompts myself.
As a power user, you could certainly try to automate the whole thing.

Interesting... what would be an example of a prompt you use? Playing around with it myself - I found ChatGPT starts to make very similar/redudant stuff very quickly, which almost removes any kind of utility it would have, because have to constantly be checking it to make sure it did things correctly...

Hi Pete,

a) I do agree with you when you say that images are not "directly" used to create new images. That is correct. You may have misunderstood or misread what I wrote. But the images are stolen to create the initial model - in order to create a "model" - from which new images are then created.
b) The courts may "decide". If they decide to say that theft is not theft, it does not mean its not theft. One does not need a 3rd party "authority" to decide if something is wrong, when they know it is in fact wrong.
c) I actually do understand how "machine learning" works, in fact I even have a background in it (among other things). "AI" (at least the "ai" that is used now) - can't "reason", doesn't re-write its "code", etc. It is a set of algorithms that computer programmers have put together, to create a "model" (i.e., mathematical representation) of what say an "apple" is. The reason the "ai" machines need keywords (and shutterstock,etc just "gave" them, then gave most contributors a paltry some for their data) - was so it knew what words/phrases to associate scraped (stolen) images with. But it is - in fact - stealing.

Basically - the companies have no idea what a "cat" is. So they scrape (stealing) a bunch of random images. But they need to associate keywords with the images they scrape. The algorithm then essentially determines a pattern (i.e., say a "cat" has "lines" representing "whiskers", has a "round head", has two "triangle ears"). It then "compresses" this data (so instead of storing 100,000's of images, it simply says "a cat is an object which has whiskers, round head, and triangle ears") - which it deterimined soley by keywords/etc associated with it, and the scraped stolen data.

It's kind of like using winzip to zip a text file. You could have a 100MB "text" file, and then a compressed 1mb "zip" file (10000 times smaller). But the original "data" is still there, encoded. The latent diffusion model essentially just "blends" multiple text files together and then when you 'unzip' a file its decoding the file.

What is going on in the "AI" space is kind of like...

Someone "knows" someone "else" stealing a bag of apples from a street vendor, and drops several apples on the street. The person (group/organization) "stealing" the bag of apples justifies their actions saying it was "research", and they are just "studying pyschological effects" blah blah blah, and have "funding" from various sources... the person stealing the apple from the street is the "ai" company knowing that they were stolen apples, but pretends to play innocent saying "oh gosh darn golly gee, I just saw the apple lying there"...

If the person makes it into apple jelly, or apple vinegar, or whatever - they still knowingly took it knowing it was theft...

And what makes things even more interesting, is if you find out the person that "found" the apple - was the one that "funded" the person doing the "research"...

What "they" (most of the "ai" companies) are doing is theft, plain & simple. When your major "problem" is "watermarks" - and how to get rid of those "pesky things" - where do you think "watermarks" come from? Certainly not properly licensed content where the creator of those works has given explicit consent...

Quote from: SuperPhoto on January 27, 2024, 08:01

Hi Pete,

(a) Actually, "ai training" is actually stealing images.

In your opinion. I tried the detailed explanation, now the shortest version possible. Images are not directly used to create new images.

Yes, the courts will decide this and if it is Fair Use or not.

Your reasoning for why hands are wrong is one of the most laughable things I've read today. You clearly are in denial or don't understand how AI/Machine Learning works. "not enough examples" 

LAION, the nonprofit that put the database together, used web crawling tools to create datasets with more than 5 billion links to online images, which companies can then pull from to use as training data for their own AI models.

AI does not understand function or purpose, it only understands training. AI doesn't know how many fingers or how many legs on a chair, or how many legs on a horse. Not because it hasn't seen enough in 5 billion images, but because AI lacks reasoning. AI doesn't understand anatomy, only what it has seen. And since AI creates new images from training, it might have more or less fingers, based on the random samples it uses to create the new hand.

Hi Pete,

(a) Actually, "ai training" is actually stealing images. One of the big "problems" they've had is figuring out how to remove "watermarks". Sure - they create "models" - but it is based off of theft.
(b) The algorithms (instructions) are not actually "learning". Just because the computer textbooks say it is "machine learning", doesn't actually mean the machine is "learning". It is not. For clarification - it is code a person wrote to look for "patterns" in order to create a "model representation" (i.e., compressed data structure), so instead of 100 gb of say pictures of "oranges", it could be say "represented" in only 0.5 gb, and then taking varying inputs (i.e., "red" orange, "blue" orange, etc) - to essentially 'blend' models to 'make' new stuff - based off of theft.
(c) The "courts" (i.e., like a 'tennis court') - simply (sadly right now) - basically do whatever someone with the most $$$ tells them to do. (Look @ the last 4 years - just follow the money to figure out what really went on).

The "ai" actually is directly using your images, as well as the images of others, and essentially creating composite images. It does not create "composite images" on the fly (it is not simple alpha blending) - but basically a more sophisticated form of theft - basically "pre-blending" the images in a "model" - and then using that "model" to generate new images. Theft is still theft. One does not need a "court" to decide if it is theft - when it is in fact theft.

Lets start, the agencies aren't in the EU. I'm not either. But lets say, OK if they want to sell there, they will need to abide with those regulations.

Problems with the disclosure Publishing summaries of copyrighted data used for training.

1) Yes, we looked your image. So what?
2) We downloaded your image? But it wasn't used in the training
3) We used your image to train the machines.
4) There's no way to prove a one to one direct use of your image, in any generated results.
5) Images are not directly used to create new images, the machines learn and create new images from their training.

But aside from that nest of snags and questions.

Fair Use What is fair use. The courts need to decide that.

I look up a word in the dictionary and find the spelling and meaning and I use that knowledge to write a sentence. I publish the sentence. Did I steal the meaning or the word? I look at a photo and find the meaning and learn about how it's composed and the subject. I create my own image, using that knowledge, I didn't copy the image, I just learned from it.

AI does not copy images or re-use bits of them in the new images. It learns from what's in the images, the content and shapes, and creates an entirely new image.

Fair Use?

AI training is not stealing images. As much as I don't like AI and I'm not impressed by people with seven fingers and three arms or hands growing out of their ears and face, it's not directly using my images in the output product. AI / machine learning, is looking and learning.

The courts will decide that, not us or some political agreement.

How do you report that your image or idea was stolen?
I found about 50 contributors who redid my best selling image and uploaded it using my exact unique description and keywords.

Now some of them literally follow my portfolio and rip it as soon as my images get approved. It�s disheartening

I�m not going to post a link to my portfolio, it�s just asking for more copycats. And my portfolio is not generic, it�s illustrations and png.

a) People will copy anything. For the images, tends to be east indians who do that (steal/spam/copy cat to "get rich quick"), create fake profiles using stolen assets then uploading those, etc, etc... (of course not all, and of course not 'just' restricted to them, but does seem to be a rather high percentage that do that to other people...)
b) Depending on "how" it was generated - you may/may not be able to do something, which is why I think people are asking... if you have a sample image (not your portfolio) - and can show the similarities between your image & the copycat, it would make it a lot easier to determine if it was ripped off, or it was more of a conceptual thing...

If you aren't comfortable sharing your images here - that's fine... I might suggest contacting adobe then and explaining what it is you need assistance with.

Quote from: DiscreetDuck on January 24, 2024, 15:17

Hello.
- You can consider that the majority of the users of this forum are fans of AI. They may not really be sensitive to these issues. I even think that they wouldn't be put off doing the same, in secret of course
- For my part, I now prefer to ignore the way in which I am (highly) robbed, too much energy, too much time, and stock sites that don't care much because: It doesn't matter who uploads and how the image is produced, they collect royalties in any situation. And they have less possibility to detect fraudsters anymore, and fraud is massive.
- I Know myself that no idea is really owned. And since the AI machinery is fed by anyone's properties, our images photos are now very relatively our exclusive property. New world of stock imagery.

But... did you generate your images with AI yourself? (it seems to be). If the case, you don't own any copyright and you should not even complain. 

I don't stop anyone from generating A.I. images based on their own ideas. These guys literally used my photos and A.I. images as a reference to try to replicate my portfolio. The fact that you are OK with that is... I don't know what to say.

Even trying to replicate my portfolio by using my ideas would be wrong, but trying to create almost exactly the same photos is just a fraud to me.

a) What is your portfolio? I'd like to take a look - can you provide an example of your original image + their "modified" image?
b) Are you saying that they took your original photo, fed it into the midjourney engine, to generate a virtually identical "ai" image? (same people's faces, same poses, maybe just a different shirt color, etc?)