391–400 of 496 posts
Yeah, Adobe, release a competitor of your own clients, they'll love it.
Someone's going to have a slick and extremely powerful product where all the tools for creating and editing work seamlessly with all the tools for generating and other ML-assisted processes.
Yeah, Adobe, release a competitor of your own clients, they'll love it.
They'll have to buy it (or something like it), they won't have much of a choice. Artists that use this kind of tool will be so much more productive than ones that don't, that they simply won't be able to find employment otherwise.
Discovering an artist using AI generated images won’t be different.
Earlier quoted context omitted.
I am definitely glad to see attention being paid to ethical sourcing of training images but I am curious: did the people who made all those stock images get paid for their work being used for training? Did they check a box that explicitly said "Adobe can train an AI on my work"? Or is there a little clause lurking in the Adobe Stock agreement that says this can be done without even a single purchase happening?
We are working on a compensation model for Stock contributors for the training content, and will have more details by the time we release. The training is based on the licensing agreement for Adobe contributors for Adobe Stock. (I work for Adobe)
Earlier quoted context omitted.
Imagine a paying Adobe CC customer. They use Firefly to generate a poster, and unbeknownst to them, the image it generated is a reasonable facsimile of a copyrighted/trademark character. The person has inadvertently committed copyright infringement. So does Firefly need to come with a warning? The safer solution, to the chagrin of another commenter, is for Adobe to neuter the tool by only training on data in which Ad…
Surely with all our contemporary AI prowess we can train a model that identifies "reasonable facsimiles of copyrighted/trademark characters" after generating them and alert the user that it could be argued as such. Still, let the user decide. We do not need creative technology to regulate observance of copyright law. (By the way I think the chagrined other commenter was yours truly ;-))
Earlier quoted context omitted.
It would have trouble matching on trademarked styles, or individual artists / creators styles. One of the primary goals for Firefly is to provide a model that can generate output that respects creators and is safe for commercial use. (I work for Adobe)
I don't really understand the negative comments on this. Though a hacker by heart, I'm a designer first and foremost And I'm extremely eager to get my hands on AI tools that let me extend my capacities based on _my own_ styles & context, and that is focused enough on this scope to evade future legal obstacles when used in production Very excited to try this tool!
Earlier quoted context omitted.
I would consider both Stable Diffusion and MidJourney to be startups and both are better by far than the established companies like OpenAI, and there are dozens of LLMs soon catching up with GPT3/4. As you say, it will be a very interesting next 6-12 months..
I played around (till free account expired) with MidJourney and what one can produce out of blue is mind-boggling. Indie devs can with a bit of effort generate much of their art via this for peanuts (the only problem may be consistency from what I've seen, basically every image is like from another artist, even in same batch)
Earlier quoted context omitted.
>That is, these companies are largely not doing the hard part, which is creating and training these models in the first place. fyi, for Adobe Firefly, we are training or models. From the FAQ: "What was the training data for Firefly? Firefly was trained on Adobe Stock images, openly licensed content and public domain content, where copyright has expired." https://firefly.adobe.com/faq (I work for Adobe)
Sure but are you standing on the shoulders of Stable Diffusion or not? Fine tuning Stable Diffusion with your own images is way easier than creating Stable Diffusion in the first place. If you're creating your own I stand corrected and that's some serious investment.
(i work for Adobe)
dall-e for example never gets faces right and good luck trying to have the same character look the same between generated images. Feeding it a seed image sometimes got something similar but always with major uncanny valley vibes.
Anyone know how to get around this?
This is cute and all for creating one-off, in distinct, stylized images. But what if I want to illustrate say a children’s book with generated images? dall-e for example never gets faces right and good luck trying to have the same character look the same between generated images. Feeding it a seed image sometimes got something similar but always with major uncanny valley vibes. Anyone know how to get around this?
You'd use what's available as "good enough", and manually modify it when needed.
Then you'd wait a few years until it does get faces right (haven't seen much issue with faces to be frank. Hands, yes).
Earlier quoted context omitted.
They do not currently operate on the basis of fair use. Operating as a human by looking at images and learning how to draw or paint is not 'fair use', it's a right given to you by either God or Mother Nature, so the legal basis for neural nets learning from other art is that it's learning like a human and creating new art from just knowing what art human think is good and optimizing its creation of art to mimic if no…
As far as I know there are no religions or legal systems that posit that there are any rights inherently given to machines.