1458 here. Not bad

c) Companies that "license" the dataset should be paying in perpetuity (i.e., monthly, regular income) for ANY "data training", because obviously they want to make "perpetual revenue"- and by stealing/using other people's data with the current "one-time peanuts payment" - essentially you are making extremely low cost competition for yourself for peanuts. You should be compensated EVERY single time someone makes an image/video/etc based off ANY kind of model derived from your works.

You go to a mechanic and he replaces your rotors and brake pads. Then you should pay him a fee, monthly, every time you use the brakes on your car? 

The images are used one time, to create a model and that machine learning model, creates the new images. Your images, and mine, are never USED again.

When you license a stock image, the contract often says in perpetuity now. Same for a training use for AI. It has been licensed for use, forever, not for a single use. If you prefer that to a single use, which is what the actual use was, then fine. We still get nothing more than what the training has paid for.

Another complication, if there are 10,000 images of a pickle, used to train the software, then every time the software makes a pickle, 10,000 people should be paid again, for having their image used in training the machine? How do you know which images were used for every new image? There is no direct connection back to any specific image, when the machine creates a new image, from what it has learned.

I'd really like someone to answer that last question.

You go to a mechanic and he replaces your rotors and brake pads. Then you should pay him a fee, monthly, every time you use the brakes on your car? 

The images are used one time, to create a model and that machine learning model, creates the new images. Your images, and mine, are never USED again.

When you license a stock image, the contract often says in perpetuity now. Same for a training use for AI. It has been licensed for use, forever, not for a single use. If you prefer that to a single use, which is what the actual use was, then fine. We still get nothing more than what the training has paid for.

Another complication, if there are 10,000 images of a pickle, used to train the software, then every time the software makes a pickle, 10,000 people should be paid again, for having their image used in training the machine? How do you know which images were used for every new image? There is no direct connection back to any specific image, when the machine creates a new image, from what it has learned.

I'd really like someone to answer that last question.

You are comparing apples to oranges, actually more like apples to palm trees on a beach. Pretty much no relation whatsoever.

a) The mechanic didn't steal the rotors, pads, etc to sell them to you.
b) You aren't in the business of making rotors, pads, etc that someone else steals from you to sell forever in direct competition, and doing whatever they can to simultaneously try to prevent you from future sales while they make a lot of sales.

The people behind the "AI" (which is not a thinking machine, and it is for the most part a jewish/"identify"asjewish consoratorium of people) are engaged in a combination of massive theft, and then trickery/deception/manipulation/etc to try to get you 'out' of business for a teensy tiny payment - because they want a "in-perpetuity" model where people use THEIR stuff, NOT yours.

In other words - they are trying to trick you into selling your "asset" 1x (to them) for peanuts (i.e., $0.10) - so they can then resell it 10,000+ times without ANY compensation  to you (i.e., to them $100,000+ for your single asset).

Single use is one thing, because it is used by a single entity. Not for derivative works to be resold (i.e., you don't sell your license 1x so then the company that 'bought' it can sell it 10000 times).

Re: 10,000 - since this is a SOFTWARE program - VERY VERY VERY EXTREMELY easy to do via micropayments. And yes - there IS a connection. You tag & tokenize the models - and then that is how you get paid. I actually DO know how to do that. It is VERY easy to do. I guess you don't have the technical background/expertise to understand that - which is fine - I do. And I am telling you - it is VERY easy to implement - it is simply a matter of doing it. Of course - the people running the "ai" companies want you to think you "can't" - but that is a load of crap - they want you to think that because then it makes it easier for them to try and steal your stuff from you.

The people behind the "AI" (which is not a thinking machine, and it is for the most part a jewish/"identify"asjewish consoratorium of people) are engaged in a combination of massive theft, and then trickery/deception/manipulation/etc to try to get you 'out' of business for a teensy tiny payment - because they want a "in-perpetuity" model where people use THEIR stuff, NOT yours.

Really? No discussion from you, with attacks like that, are worth answering.

I'll just quote you:

And the psychos LOVE playing "the victim"  to make them appear above reproach. They are prolific liars, and when/if confronted - tend to double down on their lies to try and confuse the person challenging them - because most (normal) people are good people, and find it really difficult to believe there could be such psychotic sociopathic liars. Sadly though, they do exist - and a good portion of them pushed this total con (surprisingly) on most of the world.

---