|
MicrostockGroup Sponsors
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
Messages - cascoly
Pages: 1 ... 11 12 13 14 15 [16] 17 18 19 20 21 ... 170
376
« on: October 24, 2023, 15:14 »
...
So it is SUPER easy to simply add "contributor-id" (which can be the name/URL/etc or an actual number that contains all that information). And then SUPER easy to associate WHICH contributors file(s) were used in creating a "composite" image (i.e., an "ai" generated image).
SUPER SUPER EASY. Just a matter of doing it, then fairly compensating contributors with the SAME RECURRING PERPETUAL INCOME REVENUE model that the agencies so desperately and greedily want for themselves, and trying to convince contributors that anything else is "fair" (which of course, it's not). Sharing the recurrnig revenue model, with opt-in/opt-out features so at ANY time the contributor can opt-out if they don't like the terms - and assets going forward do NOT reference the input items - is fair.
i agree, starting from scratch, adding contributor id is the easy part, but given the number of possible contributors to each piece of an ai-gen image, the book-keeping becomes expensive - many thousands of entries for each image. but the much larger problem you don't address is actually finding out who the author is and how to pay them as scraping only works if the image has the artist's verifiable info (and how to verify they're who they claim to be when opting out). true, when agencies create datasets thery have info for payment, when they license these datasets who tracks payments? and will agencies actually sell data sets with the private info of their contributors? your examples are great for explaining your proposal, but they don't scale up - the amount payable to any artist is 2-3 orders of magnitude less
377
« on: October 24, 2023, 15:00 »
putting AI aside,Microstock is not dead but it is simply no longer for amateurs who want to make easy money,if you want to make money in this game you have to put money into it,travel,buy expensive equipment,models etc.....
what's really happening is the tougher chance to do stock full time. and putting money into travel is hardly the best approach to full time as it's already saturated. i've been retired for awhile & i do manage to pay for 2 intl trips each year from stock income, but i wouldnt try to raise a family on it of my travel pix the best sellers are generic people eating and casinos on cruise ships, not the destinations themselves
378
« on: October 24, 2023, 14:50 »
...
I made a matchbox player, (AI or computer?) that could compete and not lose, at the game, in about 1960. Much like your example, the boxes didn't know the rules or the game and they didn't reason or think, they were just sliding drawers, where the decisions were learned and played.
...
was that the one from Martin Gardner's SciAM column? I still have mine, saving it for display in my memorial library.
379
« on: October 23, 2023, 13:21 »
the problem isnt just the long review times - it's random chance of review - many here reporting reviews of a week while others have images stewing for months.
i recently had a handful of images rejected (for the usual vague 'quality' reasons) that were submitted about a week ago, while many others (including AIM) have been there for 2 months
380
« on: October 23, 2023, 13:13 »
besides the fact that there is no, way to trace which images were used & worse even if your 'easy to do' way of marking were possible, for most images (maybe close to 0%) there is no way to identify who the artist is for the billions of images used - many have no names assoc'd and those that do lack verification and an address to pay to). how would your revised training know who (& how) to make payments
We're talking about Adobe generative fill. Adobe knows very well where to find the artists who's photos were used and how to pay them as they used images from their own database.
yes, it's about AS specifically but i was responding to the comments that were not limited to AS. again, how does AS know whose images were used for each creation since ML eliminates any way to track that, even if identifiers were attached initially. each image is translated into thousands of datapoints and millions/billions of operations are performed to generate each new image.
381
« on: October 23, 2023, 13:09 »
....
b) When an AI "image" generates an image - it does reference essentially computer models. However - it is totally easy to say for say a "car" - 55 contributors were "tagged" in that computer representation of a "car".
c) When the asset is generated, contributors get micropayments (i.e., say the "ai" image was "worth" $0.10 to the company, and the revenue split was 50-50. So 55 contributors each get $0.05 / 55 ~ 0.0005 cents. May not seem like a "lot" - but with the millions of images generated daily - quickly adds up. (So say each image was like that, 55 contributors for different 'models' - and say 100000 images were generated referencing their input in a month, then 0.0005 * 100000 = $50)...
besides the fact that there is no, way to trace which images were used & worse even if your 'easy to do' way of marking were possible, for most images (maybe close to 0%) there is no way to identify who the artist is for the billions of images used - many have no names assoc'd and those that do lack verification and an address to pay to). how would your revised training know who (& how) to make payments
but you've stacked the odds in your favor - you claim there would be millions of images generated using only 50 originals, when when many thousands of images would be likely ( a conservative estimate), so your estimate of payment dueis off by several orders of magnitude and certainly not millions of images generated daily
You obviously don't have any computer background, or very little. Do you have ANY programming experience whatsoever, let alone large datasets? Yes, it is VERY possible, and VERY doable.
The illustration/example is designed to keep it simple, so you can understand. Obviously programming would be a little more sophisticated than that. But it is VERY VERY easy to do - it is simply a matter of DOING it.
you know NOTHING about my computer experience. you don't address the biggest problem - how do you identify who made the image & how to contact & pay them? that information isnt available in most cases and now you've changed the goal posts, saying funds should be distributed to everyone, not based on where their images were used - you know little/nothing about ML it seems and that is relevant as the major fallacy in your proposal assumes you can track where an image is used
382
« on: October 22, 2023, 13:03 »
Mat - is Adobe going to do anything about the useless rejections reasons when AI might be wrongly marked? with the current system there's no way to tell if an image has been rejected when it was falsely identified as AI-gen since it allows it may have been rejected for other reasons. so there's no way to know whether we should appeal/resubmit or not
here's what we get:
Possible reasons:
- Non compliant use of another artists name.
- Undeclared Generative AI Content.
- Content not compliant with overall guidelines:
383
« on: October 22, 2023, 12:42 »
....
b) When an AI "image" generates an image - it does reference essentially computer models. However - it is totally easy to say for say a "car" - 55 contributors were "tagged" in that computer representation of a "car".
c) When the asset is generated, contributors get micropayments (i.e., say the "ai" image was "worth" $0.10 to the company, and the revenue split was 50-50. So 55 contributors each get $0.05 / 55 ~ 0.0005 cents. May not seem like a "lot" - but with the millions of images generated daily - quickly adds up. (So say each image was like that, 55 contributors for different 'models' - and say 100000 images were generated referencing their input in a month, then 0.0005 * 100000 = $50)...
besides the fact that there is no, way to trace which images were used & worse even if your 'easy to do' way of marking were possible, for most images (maybe close to 0%) there is no way to identify who the artist is for the billions of images used - many have no names assoc'd and those that do lack verification and an address to pay to). how would your revised training know who (& how) to make payments but you've stacked the odds in your favor - you claim there would be millions of images generated using only 50 originals, when when many thousands of images would be likely ( a conservative estimate), so your estimate of payment dueis off by several orders of magnitude and certainly not millions of images generated daily
384
« on: October 22, 2023, 12:30 »
in the end, only agencies that value human work and not artificial intelligence will survive. This will happen very soon. Man is not a robot and never will be. Humans have emotions, not machines. I hope some agency realizes this soon, otherwise it will fail!
interesting - most of the comments here claim ai is going to run them out of business (even though few agencies actually accept ai)
385
« on: October 22, 2023, 12:24 »
The annoying part is the batch rejection, its not like just one of 2 images are rejected like the good days, now a whole batch of 25 will be rejected. It is as if the reviewer rejects one image so rejects all 25, no other logical reason.
it's like they've adopted alamy's silly rules
386
« on: October 20, 2023, 13:07 »
As a high volume seller with a rather small portfolio I put a lot of work into each image, and Adobe (or rather Fotolia) used to appreciate that, but nowadays it's getting ridiculous.
I'm in a completely similar situation to yours! In one of the topics I wrote that they began to ignore me 100% after I refused to give my works for free download.
Have you noticed this trend in yourself?
...
Wow, that's an interesting viewpoint. I had never considered that. I didn't opt in for the free downloads either, but I can't remember when this became a thing. Did they start it recently or years ago?
no need for conspiracy theories - i've always been opted in and also get the batch rejects & long (2 month+) wait times. correlation is not causation, especally when it's anecdotal
387
« on: October 20, 2023, 13:02 »
Canva always had pretty random (to me) rejections. I have given up with uploading there.
I have a fairly high percentage of images accepted on Canva, but even to me acceptance seems largely random. I'm not trying to find logic in it anymore, just upload and forget is the best strategy.
agreed - i stopped uploading for awhile when the rejection rate changed so drastically, and now it's my lowest priority. it may just be that they've become more selective than other agencies for topics they already have 'sufficient' coverage. their audience is very different from the other agencies & that may affect their decisions.
388
« on: October 18, 2023, 12:56 »
yes, pretty pathetic - $14 will see how monthly payments accrue
389
« on: October 17, 2023, 13:12 »
Has anyone already received their payment?
i received the payment at the same level as previous months recently and a day or 2 before the stated 15th - the later payment will be for dataset usage
390
« on: October 16, 2023, 12:42 »
..
If you look at the image numbers, this isn't from ages ago, but is recent. The rules are just ignored and the reviewers don't catch it.
we should have more sympathy for overworked reviewers - they're too busy tagging ai-gen in standard images
391
« on: October 16, 2023, 12:36 »
I would like to add to Pete's contribution that for about 10 years, at least in Germany, sales of vinyl records and players have been growing steadily.
Jazz goes for me only on vinyl and that seems to be the trend 
Not everything is lost. 
and after 150 years of photography we still have artists who smear colored liquids on various surfaces, though few in the photorealistric style of David, Coubier, Daumier et al - instead it led to the expanding conciousness of impressionism, expressionism etc
392
« on: October 16, 2023, 12:31 »
...
A computer can do some things, that humans have taught or programmed, but it can't reason or know function, or why. That's why 3 arms, 7 toes, or mechanical things, just get mashed into flawed and often impossible combinations.
a bit too optimistic - we already have ai translators which do not understand the languages they render - world class chess & go & Jeopardy apps that don't even know the rules of their games and many more. for now these work for specific problems but that will expand to include tasks now considered to require human consciousness. remember that once humans were defied by being the only species who used language earle (1999) summarized his Chinese Room Argument concisely: Imagine a native English speaker who knows no Chinese locked in a room full of boxes of Chinese symbols (a data base) together with a book of instructions for manipulating the symbols (the program). Imagine that people outside the room send in other Chinese symbols which, unknown to the person in the room, are questions in Chinese (the input). And imagine that by following the instructions in the program the man in the room is able to pass out Chinese symbols which are correct answers to the questions (the output). The program enables the person in the room to pass the Turing Test for understanding Chinese but he does not understand a word of Chinese.https://plato.stanford.edu/entries/chinese-room/
393
« on: October 13, 2023, 13:48 »
Topaz actually DOES inject metadata into the files. They seem to do it randomly, so some files might have it, others might not - a bit of a dishonest tactic. You need to select your files (under windows), right click 'properties', then remove the "title" (which they usually put a compression method directly linked to topaz), and I think it is 'program name' as well (two different fields).
yes, program name shows "topaz photo ai' which is not ai-gen but mis-informed reviews may think it is.
394
« on: October 12, 2023, 13:02 »
Is anyone else finding that the selection of categories for photos is a bit limited on AS? Two extra categories that I could think of right now (that could be very useful) are Health and Products.
The reality is that the category has very little influence in search and/or sales results. My recommendation is to focus your attention on strong metadata with your title and keywords.
...
i've often thought categories weren't much used by buyers one strong sugvgestion is to change the rejection report: >>
possible reasons:
- Non compliant use of another artists name.
- Undeclared Generative AI Content.
- Content not compliant with overall guidelines:artist name is easy to detect, but including 'not compliant' means we don't know if the image was falsely rejected as ai-Gen, so impossible to know whether we should re-submit
395
« on: October 11, 2023, 13:23 »
Just got the reply from Shutterstock, the problem seems to have been the use of Topaz:
"..... After checking the Topaz Sharpen AI and Topaz Gigapixel software that was designed to remove noise of images and after confirming that you are the copyright owner of the content that was rejected on suspicion of being AI-generated, allow me to provide the follow steps:...."
I am now to resubmit the image with a case number.
Somehow everything is getting more and more complicated.
one of the early 'experts' said i should add "not AI generated" when submitting images! i escalated & got an intelligent person who set up a case file & the false positive were accepted
Yes great, titles will have to look like this in the future:
Beach - the image is not AI generated and I own all rights to the image (proof attached), focus on the sixth parasol from the left
may be i was foolishly poking the bear, but when the asked for details on how i 'generated' the falsely labeled images, i replied 1. Point camera 2. click shutter button 3. copy image to computer 4. add metadata 5. upload to SS
396
« on: October 11, 2023, 13:16 »
Maybe sticking with the old non CC photoshop tools will be a selling point eventually. I'm surprised that the so called AI makes up nonsense text when it probably should be able to recognize what language it is and actually recreate the text even if it doesn't get the font or every word correct. It certainly makes any image that has been uprezzed a bit suspect now since it is actually making up details. ...
terrible text is a direct result of ML - training doesn't recognize the content it's transmogrifying, much less that there is text & even less what language is there. if anything, just that is text (a failed example of the Chinese room metaphor). so when using the trained dataset (which has none of the actual images - much less the text on a t-shirt - it ends up creating hieroglyphs because many t-shirts have text on them.
397
« on: October 11, 2023, 13:07 »
...
Tools like Topaz Sharpen and Topaz Gigapixel use so-called GAN networks and rebuild a sharpened or enlarged image completely from our photo - i.e. from the original photo there is not a single pixel left in the final image.
The image is reassembled in small blocks from a GAN AI network similar to MidJourney, Firefly, etc. and the automatic recognition software now recognizes that certain features from a generative AI are to be recognized here.
...
So from a technical point of view, an image sharpened or scaled with GAN networks is a generated AI image and that's why these systems strike.
In short, the systems are not good at deciding between "Generative AI for complete image generation" and "Generative AI for small image elements for image enhancements" -...
thanks for the details, but the results aren't consistent - i submit hundreds of images to SS and, luckily, only a few have been tagged as ai-gen (largest was 10 out of 100, almost all run thru Topaz apps including upscaling - I'm working thru niche subjects from scanned slides, and older images - Topaz has resulted in acceptance, and sale, of images rejected as out of focus or noisy tears ago.
398
« on: October 11, 2023, 12:55 »
Just got the reply from Shutterstock, the problem seems to have been the use of Topaz:
"..... After checking the Topaz Sharpen AI and Topaz Gigapixel software that was designed to remove noise of images and after confirming that you are the copyright owner of the content that was rejected on suspicion of being AI-generated, allow me to provide the follow steps:...."
I am now to resubmit the image with a case number.
Somehow everything is getting more and more complicated.
one of the early 'experts' said i should add "not AI generated" when submitting images! i escalated & got an intelligent person who set up a case file & the false positive were accepted
399
« on: October 10, 2023, 13:16 »
I've had similar problems of false positives claiming ai-gen - sometimes 1 of a series is rejected as AI, all others accepted. all used Topaz, no ai-gen
in each case i contacted support & they gave me a case number to send to the reviewer and resub those images which were then accepted
my concern is we might get suspended because of false positives
400
« on: October 08, 2023, 15:24 »
... The same will go for the AI image creators that just let their AI crawl the whole internet for image training. How would you even make a list with ALL images of the internet? That's hardly possible. But the point of this law and why I would welcome it is that it would show these companies that they screwed up and used copyrighted material they had no right to use. If they just had bought licenses from for example microstock agencies they could make a complete list with image IDs and artists' names. But they did not do that, so they can't.
yes, the EU proposal would be interesting, but it also shows why the continual demand for payments for images used can't/won't happen - the deed is done. and most images don't have an artist's real name attached. and if they do, there's no way to contact them (slightly different when agencies use only their own collections to train)
Pages: 1 ... 11 12 13 14 15 [16] 17 18 19 20 21 ... 170
|
Sponsors
Microstock Poll Results
Sponsors
|
Show Posts
