|
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 ... 29 30 31 32 33 [34] 35 36 37 38 39 ... 170
826
« on: October 07, 2022, 13:11 »
Therefore, you can not spend any time on tags, their analysis, you can load without tags at all.
It would be extremely interesting if you are willing to do this experiment: always add a minimum amount of keywords (7) to your artworks and report you sales (or the absence of those) in a year.
and the keywords should have nothing to do with the image, otherwise results would be contaminated
827
« on: October 07, 2022, 13:10 »
....I think sites use artificial intelligence, they determine what it is from the picture and offer it to the buyer. And even if there are 0 tags, the system will create them for itself.
again - actual evidence before making this claim?? did you do any research? like finding images that shouldn't be found from title & description? a quick search on both SS and pinterest showed the tag was part of the title or description. search for 'Himalayan mountain" - how would AI be accurate enough to distinguish Himalaya from Andes or alps? yet only Himalayan are shown -- again part of the description. For example, Pinterest has not been able to enter tags for a long time, the search is only based on artificial intelligence. And this is not only the future, this is already the present. Therefore, you can not spend any time on tags, their analysis, you can load without tags at all.
you can't extrapolate from pinterest to agencies - pinterest is not an agency & their primary goal is not to sell images.
- pinterest CANNOT use tags because most(?) pins are images w/o tags
- n evidece they usde AI - they cvould just
- even if it's the future, that's not a reason to self-destruct today by not usinf keywords
828
« on: October 07, 2022, 12:51 »
.. dupe..
829
« on: October 07, 2022, 12:40 »
Some stocks publish the words that buyers most often search for. So there are not only words, but also phrases. If you take these words and insert them into your work, then you will be found more often.
search algorithms likely see no difference between tags "red", "flowers" & "red flowers" main reason to include tag phrases is to disambiguate: "red river" & "river" rather than "red", "river"
830
« on: October 06, 2022, 17:56 »
...
If the extended keywords, or complex description are important and the image shows something specific, they are necessary and good. If someone just adds close words and similar meaning words, you're wasting time with words that buyers don't use to find image. If you were searching for your image, what words would you use to find it? If you were describing your image to someone else, what words would you use. That's what good keywords are.
good keywords allow someone who DOESN'T think like you to find & buy your work. (that's why agency claims of LCV were absurd) a common case would be someone who doesn't care whether a spider is an insect or not - they just want a scary or dangerous or venomous 'insect'
831
« on: October 06, 2022, 17:46 »
Interestingly, it obviously copies quite a bit as they were also including watermarks with the images they produce.
Might risk sounding like a broken record, but: The AIs sometimes generated images that have something resembling microstock agency watermarks, because they have been trained with so many watermarked (unlicensed!) images that they wrongly learned that the watermark was part of whatever it was supposed to generate. When an AI generates a watermark, it "thinks" it belongs in the picture like a suit to a businessman or the sun to a picture of a sunny sky. It's an issue of wrong learning, not an issue of copying. It recreates the watermark, just like it re-creates the sun or a suit. It cannot understand that the watermark is not part of whatever it is supposed to depict. If an AI was capable of thinking/realizing that whatever it is creating in images was actually something that exists in the offline world, then it would think that people walk around with floating watermarks in front of them.
I start to think that many people do not really understand what an AI is. Artificial intelligence. It's not a computer programm that copy & pastes stuff. It is a program that has learning abilities. It gets input and it learns from it. Give it the wrong input and it will learn to create wrong results.
exactly - it's much easier to make absurd claims rather than actually doing a bit of research to see how these AI actually work
832
« on: October 06, 2022, 17:42 »
..
...You yourself mentioned the concept of human-machine people but never mentioned the fact that after the singularity the next purely logical evolution doesnt bode well for humanity. Remove art, expression, individuality and perhaps most importantly trust from the equation and you are hastening the process.
the next steps will see AI for stock buyers, then they'll replace graphic designers. AI will read & post to social media and decide what their humans (a la 'Mr Peabody's boy sherman') 'want' to buy
i mentioned earlier a thoughtful take:
https://www.cold-takes.com/ai-could-defeat-all-of-us-combined/
What I find interesting is what will people do to earn money once AI is encouraged to replace all the jobs. Yes I know new jobs will recreated but nowhere near as many. AI is writing books, news articles and the like and that's before things like automated cars etc become the norm. Thousands, millions of jobs gone. If a large % of the population is no longer earning money (or as much money) who is going to buy and use the services. Given the amount of price cutting in all areas in order to get an edge, a reduction in sales is the last thing that is needed..../.
it looks bleak only if we continue the current robber baron capitalist paradigm with obscene inequality of income and huge corporate profits with few taxes it's a political problem - not economic or technological. a paradigm shift would see a much more progressive tax system, balancing of income ranges, etc. people, not corporation oriented. this would provide a basic livable income for all. people would have the choice to accept that and pursue non-profit areas that couldn't provide an adequate income. this is already seen from techies who achieve their monetary goals & retire early to work for np foundations et al . or they could continue to follow professions that haven't been overtaken by AI (yet) a shift from Hobbesian dynamics would allow more folk to have the options of the super-rich.
833
« on: October 06, 2022, 17:30 »
incorporating copyrighted elements, parts of someone else's artwork is inevitable
AI doesn't incorporate anything.
AI learn what is and how to recreate any object (or human faces, animals... everything)
Of course there are legal problems because images used to train are copyrighted; but there is nothing that will be "incorporated" in new images
It's quite new scheme, and it cannot be managed with "classic" discussion, it's completely new issue to solve.
Interestingly, it obviously copies quite a bit as they were also including watermarks with the images they produce. Of course, the programmers will write a bit of code to remove them in the future but it's obvious it's basing images on real content.
again, no - doesnt need to copy anything - a matrix analysis creates completely different information. the watermarks aren't stored per se - instead the ML thinks watermarks are part of the object. using a larger training set would eliminate some of that problem. of course, they shouldn't be using watermarked images for training min the first
834
« on: October 05, 2022, 13:54 »
..
...You yourself mentioned the concept of human-machine people but never mentioned the fact that after the singularity the next purely logical evolution doesnt bode well for humanity. Remove art, expression, individuality and perhaps most importantly trust from the equation and you are hastening the process.
the next steps will see AI for stock buyers, then they'll replace graphic designers. AI will read & post to social media and decide what their humans (a la 'Mr Peabody's boy sherman') 'want' to buy i mentioned earlier a thoughtful take: https://www.cold-takes.com/ai-could-defeat-all-of-us-combined/
835
« on: October 04, 2022, 16:51 »
...
Besides, It is always a good idea to read the terms of service. You can't even be the copyright owner.
did YOU read DALL E's TOS? it's been mentioned here before that they explicitly say the copyright belongs to the artist
836
« on: October 04, 2022, 16:47 »
...
What it's really bad at is named locations. It doesn't seem to be able to produce decent realistic images of Big Ben or Mt Everest or the Taj Mahal because those images require a single viewpoint and you can't combine images without it looking very odd.
these folk are likely screwed as well - just not as quickly as your other examples. when i asked for images of crowds and the yeni camii near the golden horn, the results showed several different angles only images got the minarets correct, but it's just a matter of time before that''s improved 3 of the 4 'sherpas on everest' pictures had a reasonable image of everest with a recognizable west ridge & summit pyramid
837
« on: October 03, 2022, 15:43 »
great article.
really surprised by how many images have 45+ keywords. maybe people are accepting all suggested keywords w/o discrimination. how relevant are 50 keywords for a sliced tomato? on a quick check, my images average 15-25 tags, mostly on the lower end
when you refer to page placement, what tags are being used?
keyword count shows about 90% have 40 to 50 tags, but the page 1 boxplot shows 25% have between 30 & 40 tags? interesting that first 6 pages have much bigger spread than lower pages. is keyword spamming reducing placement?
"It does not take to be a data scientist to notice an interesting pattern: low-competition keywords mostly have 2 words while high-competition keywords mostly have 1"
more specifically the high comp words are extremely generic, and, re Shannon, contain little information.
otoh, your example for high ranking shows
workteam 35s, make menopause, age gray, facial treatment, insurance afro, girl count, online transfer, older caucasian, workshop training, office benefit, african domestic, enjoy vat, businesswoman profit
again, information theory would predict this high ranking, but how does that ranking convert to actual views (much less sales)? how many times are these tags actually searched?
"It is clear that keyword competitiveness has direct effect into where you will rank. Although it does not guarantee sales, it does get you eyeballs from the customers."
again, page placement does not mean views, so those keywords actually need to be searched for? how many people will be looking for single keywords like 'background, white,nature,beautiful'? and subjective values like 'beautiful' really give little information (compared to 'ugly') what would more interesting (but explosive) would be 2-3 words combos: eg "white backround", "green nature"
838
« on: October 02, 2022, 13:38 »
again, unsupported assumptions tending towards conspiracy theories- that's not how machine learning operates,
Then please enlighten us how it does work.
...
i & others have supplied multiple links about ML on the 2 threads about AI -- or just google it
839
« on: September 30, 2022, 18:58 »
I see AI as a new means of adding to my creative inputs. You (we) still need the traditional skills as we always will but like all of technology from the first camera obscure to the digital ones we have today. They are tools to capture a moment in time or to capture a thought or inspiration. I see AI as a new tool to help us generate something that could have been impossible or out of reach before it was invented. My thoughts on using AI is exactly like this video... https://www.youtube.com/watch?v=yAy0NlGesOE
I think combining both traditional skills with AI is mindblowing and such a time saver.
In The Third Man, Orson Welles character Harry Lime says, In Italy for thirty years under the Borgias, they had warfare, terror, murder, and bloodshed, but they produced Michelangelo, Leonardo da Vinci, and the Renaissance. In Switzerland, they had brotherly love, they had five hundred years of democracy and peace, and what did that produce? The cuckoo clock.
840
« on: September 30, 2022, 18:54 »
I don't want to be the guy thinking of smart words of creating an great AI image.
I want to be the guy that that takes a photo of an actual landscape, person or happening. I want to be thinking about composition, light and technique or maybe even having the lucky opportuntiy to take a once in a lifetime shot when a rare situation occurs.
Sorry, that I am not embracing the future with all of it's new capabillities, like DALL-E, but I think it will be a total waste of human's capabillities and good things will truly get lost by it.
It will be really sad if AI programmes could reach the point that they would replace us (by using our work to begin with).
sad or not, you keep using the same arguments painters used against photography - yet painting hasn't disappeared. creative destruction means you either adopt the new technology or find a way to make photography relevant in an AI age. you describe making 'art' but we're talking about producing stock images. agribusiness forced out many smaller farmers. those who survived found new approaches - organics, buy-local, heritage veg, etc
841
« on: September 30, 2022, 18:48 »
and you continue to pose the false narrative that these AI are copying images to create new ones. Even those agencies banning AI do not make that unsupportable claim. instead, they are concerned about the training of AI which is an entirely different issue
Then how do you think this works?
Let's take for example the picture that was supposed to be a 19th century naval battle that you created some some ago with the AI. How do you think the AI created this? Clearly it did not learn what a naval battle is, because the picture did not show one. So either it looked on the spot for pictures with a 19th century naval battle in description or keywords and created a similar image, or what it thought is similar or else during the training most of the information from the images the AI was treined must have been written into a massive database and then the AI created the image from that information. Either way, it used more or less directly existing images to create your images. What other way is there?
Clearly the Ai did not make the jump to understand the images it was trained to understand that a naval battle requires ships fighting with each other in water, so it could not for example create a photo realistic image with water and sailing ships, but the style of the image clearly showed that it was created with old oil paintings a a basis. The AI was not able to abstract from that.
again, unsupported assumptions tending towards conspiracy theories- that's not how machine learning operates, and your claim DALL-E copies other images is specifically denied by open-ai. you may choose not to believe them, but that doesn't justify your claiming to know how the image is created. the point is that the AI doesn't know anything Of shoes and ships and sealing-wax
Of cabbages and kings
And why the sea is boiling hot
And whether pigs have wings.'like the other examples i mentioned, the AI is not mimicking a human
842
« on: September 30, 2022, 18:12 »
How is the training issue (not seeking permission or giving compensation to the copyright holder) different from using samples in music? There've been lots of lawsuits over this and I don't think the notion that the sample is short gets you off the hook.
The fact that you can't create these images without a large database to "train" with is not at issue, as far as I know. The fact that there are lots of people's copyrighted work that you're only stealing a very little bit from doesn't really change the basics of the transaction. Even images lifted from social media have copyright - the person who snapped the image holds it.....
thanks for those links! Sampling music takes one work & adds/modifies it; similar to extracting sky, background from an image & pasting it into your image. but there some relevant issues: ========== C. Fair Use Defense
Exceptions exist to the exclusive rights granted to copyright owners. The
fair use doctrine allows someone other than the copyright owner to use the
copyrighted work in a reasonable manner without permission.
...
If the trier of fact determines that the level
of unauthorized use does not rise above the threshold of substantial similarity,
then the trier should find the unauthorized use to be de minimis. [too small to matter]
...
Problems arise when the
appropriation is quantitatively small and not the heart of the work.
the difference again is that these sampling cases address specific, detectable copying. in ML the individual images are so, if the data used is only one of thousands tested, it isnt the heart of the work? (very different than transforming one image) for ML training, the images are analyzed and classified. the result is not copying or storing the image per se, but storing a digital analysis of the image retrievable by its tags. such a process is one way, not commutative -- you can't reconstruct the image from the new collection [for some a wonky details: https://nanonets.com/blog/machine-learning-image-processing/#working-of-machine-learning-image-processing ] analogous to netflix scraping copyrighted works and producing recommendations? In Sandoval v. New Line Cinema,
artist Sandoval alleged that New Line
Cinemas theatrical release of Seven infringed his copyrighted work when his
photographs appeared in a movie scene without his permission. The court
reiterated the de minimis analysis in Ringgold, but held as a matter of law, that
the use of the photographs was de minimis because the photographs were not
quantitatively observable.
similarly, no individual image is observable in images created de Novo? To create a new image, the generative phrase is compared with those non-image tables. the training results are a novel expression. quotes from https://cpb-us-e1.wpmucdn.com/sites.suffolk.edu/dist/5/1153/files/2018/01/SRWILSONV1N1N-rp2cve.pdf (hopefully fair use!) the copyright questions arise in the collection & use of the imagers in the training phase, not in the later creative phase. should infringement in the first area contaminate the later use of that resource? if you use a thesaurus which was prepared using copyrighted collections/compilations, are you liable if you use that book to write the next great novel? similarly, the results of training are usable in creating new images. IANAL
843
« on: September 30, 2022, 12:31 »
there are 2 aspects of concern here and many others are conflating them in the other thread - the creation of new, unique images isn't violating any copyrights, and Getty et al aren't making that claim.
And the saddest thing to me is that they did not even bother to pay for the images used to train the AI, as they were clearly watermarked.
Yes, they give copyright to the "describers" of the AI generated images, but they used images to train it where they did not own copyright themselves. They can probably get away with this legally, but moraly I find all of this highly repulsive. They basically used our own images without paying for them to create something that one day in the future will most likely destroy our line of work.
this is an important, but separate issue. given that millions of images are being used in training, the effect of any particular image on the new creation is minuscule. So, the question is whether the owners of the training images should be compensated, and if so, how? and how much? What are the possible ways to address this concern?
844
« on: September 30, 2022, 12:01 »
..
I get how it works. It isn't alive though. What you call "learning" I call copying. Though what we call it is irrelevant really. The fact is they appropriated our intellectual property to make their product. Without our keywords and images they have no product regardless.
you're setting strawman arguments: No one is claiming AI is 'alive'. it's 'artificial' intelligence, not a human neural system. 'calling' it copying doesnt make it so. your implied definition of AI doesnt conform with any objective definition. the question of copyright appropriation is a separate problem from the question of AI AI results already are impressive: https://www.forbes.com/sites/bernardmarr/2018/04/30/27-incredible-examples-of-ai-and-machine-learning-in-practice/?sh=4f56ffdd7502other examples include Deep Blue, netflix recommendations Google's AlphaGo beats non-masters without knowing the rules of the game AlphaFold has excellent results better than any other approach in solving the problem of protein folding
845
« on: September 30, 2022, 11:40 »
The AIs don't "scrap" the internet to create images. Images from the internet were used to train the AI to create images on its own. So the "scrapping" process already took part when the AI was programmed.
At least DALL let's you upload your own image to modify it, but you can't have it re-trained based on your own images.
Again, so they say. Their code is not transparent and they will protect it as propriety. The images they output up till recently would occasionally had identifiable watermarks from the stock agencies in them. They were saying the same thing then. So the app was and is copy pasting to some extent (lets call it what it is as calling it AI is really meaningless; when is something become AI). It has just gotten better at hiding its sources. The rest is just buzz words.
Its not copy pasting, its creating new images again is pretty meaningless. Thats how copy pasting on a computer always works. You convert it to ones and zeros and output a new images at the other end. You can run as many filters and warps on it as you like, it is still using other peoples work to come up with the new images.
just because you don't trust a source is no justification to libel them with your unsupported claims, especially when those opinions are presented as 'facts' and you continue to pose the false narrative that these AI are copying images to create new ones. Even those agencies banning AI do not make that unsupportable claim. instead, they are concerned about the training of AI which is an entirely different issue
846
« on: September 29, 2022, 19:12 »
Mass production is part of the AI.
The computers don't need to rest, take a coffee break.
Even if AI doesn't prevail it will still be a competitor. Something that didn't exist a few year ago.
once again, that's not how these AI engines work! they generate a few images based on human input & the results need human post processing & editing metadata my latest submissions: DT 21 of 34 SS 34 of 34 AS 18 of 18 for AS 16 were rejected as needing a release, even though others in the batch were accepted!
847
« on: September 27, 2022, 19:54 »
Why would an "agency" need "contributors" to generate this stuff? Why not just have their own employees do it, on demand, in response to customers' requests? And eventually, of course, those former customers will learn how to generate their own images.
economics 101 - dont do it in-house when it's cheaper to out-source. for agencies, no employee can produce images as cheaply as available from contributors buyers have even more incentive to use the agency rather than pay their own designers (if they have them). now a designer can choose from 'thousands' of images - what are the chances the 1st phrase they give AI is going to be what they want? and, the raw images from AI will likely still need work for some time.
848
« on: September 27, 2022, 16:06 »
I used the following words,using an AI app, "Donald Trump astronaut" and i got this...
trying th at with DALL-E yields: "It looks like this request may not follow our content policy."
849
« on: September 27, 2022, 13:26 »
...
850
« on: September 27, 2022, 12:53 »
If there are buyers for it, if the agencies are accepting it, and it's proven to be Ok from a legal point of view, what's the problem?
Microstockers already took jobs from dedicated newspaper photographers (and others), so why are you still here?
Hiding your head in the sand, will not prevent progress to be "devastating" for you. You better embrace it.
This is why you should be thankful to them for generously sharing their experience, instead of playing their cards close to chest (as I would do), while "others" have their mouths full of sand. 
...
The problem is that AI generated photos have nothing to do anymore with the art of photography. ...
stock photography itself has little to do with art! and who says art has to be created by humans?
Nothing. However, with regards to submitting work to agencies, they require you to be the original copyright owner. If you use AI software, you're not! Same as submitting someone else's work as your own.
AI software is scraping images and metadata from the internet without the original copyright owners permission and using it to create the images in the AI software. So, submitting images from AI software under your name is similar to you downloading images from the internet, slicing and dicing them, merging them and then selling them as your own. How happy would you be if someone did that to your portfolio?!?!
no - copyright is owned by the creator of the work & DALL-E makes this clear. your description of how AI works perpetuates a false and misleading narrative - again RYFM - this is NOT how machine learning works and saying it 3 times doesnt make it true. no one making this argument has shown actual evidence that this is true. it's easy enough to do the experiment
Pages: 1 ... 29 30 31 32 33 [34] 35 36 37 38 39 ... 170
|
Sponsors
Microstock Poll Results
Sponsors
|
Show Posts
