Raw LLM Responses

I am of two minds with all the AI Art discourse. I think AI Art is interesting, and it is ever evolving, I remember the beginnings of AI art, and how it essentially came from running Image Recognition models in reverse. Instead of feeding it an image and having it categorize what it "sees" they fed it was it saw, and had it spit out an image, and the beginning of all that was a strange nightmare of colors and shapes, and almost pictures of whatever it was described. When it was all for training image recognition no one really paid any attention to the petabytes of data AI companies were scraping from the public space. Now that that same data is being trained and run in reverse to create rather than categorize, it's become problematic, because it's good at it. It's problematic because of how it works, it can only create from what it has been fed, it has no notion of actual creation. And when trained on and fed data without the consent of the creators of that data, it's not the same as another person spending time to study the whole of your works to emulate your style. Because as you said here, you can still see the bit of themself that's in there. Where I have the issue, is not in the existence in AI art, or the mass datasets used to train it. My issue is more the issue of private companies sidelining actual artists, and just using AI art for their products. Wacom, a leader in digital art tools creation refinement and sales, whose name is synonymous with digital drawing tablets decided it was OK to use AI art on the front of their packaging, and not even well done AI art. It was riddled with the obvious AI art problems of misalignment and blending artifacts, that make it painfully obvious that it was AI generated. I would be all for a company paying artists, for them to make art to be used to train a model, and that model being sold to allow for AI generated works in a particular style and things like this, or Models trained entirely on Public Domain pieces freely available and freely useable by all. I'm honestly not all that bothered by platforms like DeviantArt having their terms of service include a clause that art uploaded to the platform will be used to train Models. It's their platform, if you disagree, you don't use their platform, and your art doesn't go into their training. I do have an issue with platforms like deviant art, who already exist using an opt out system, and pushing that on their existing users, rather than an opt in system, or at least a well telegraphed plan of action that gives artists the ability to offload from the platform and relocate to a new platform, or start their own. Going forward though, artists who are strongly against AI art, now have options to take the time and 'glaze' their works before publishing, making a poisoned piece that cannot be used in model training without damaging the model, the more glazed work gets used. I highly recommend people to do that. I enjoy tooling around with AI art generation, and seeing what I can "create" with it. Prompt crafting takes some time to really understand, and it's somewhat different depending on how the model you are using is trained. And what kind of interpreter is being used to break down your prompt into the tokens to feed the model. The balance of your prompt, vs the negative prompt you use to fine tune, tags, lora, post processing, inpainting, etc. While I can never compare the level of training and self discipline and constant effort that goes into being an artist, and improving as an artist, for someone like me, with issues taking what I see in my mind, and getting my hands to make that happen on a canvas in front of me, and events in my life that has made that process even more difficult, being able to do something that to some extent emulates the end result, has been really fulfilling. I may generate 1000 images, before what I was seeing in my mind becomes partially realized, and may generate a 1000 more fine tuning, and inpainting to remove obvious problems and shift the design closer to what I intended. It can take a few hours, in the end, I sit there with something I took time working on, that fits my vision, and I feel a bit of pride in taking this still very much a work in progress tool, and using it to create something. I don't expect anyone to fully understand my view and experience. That's why I understand and support artists and their fight against being stolen from. It's why I would champion artists being actively paid in the creation of powerful AI Generation models, and why I would be willing to spend money on a well-built capable model that gave me license to commercially use the final output of my time and effort. Welp that was an absolute wall of text, if anyone actually read through it, thanks for your time, if anyone has had a similar experience as I have in life where doing art is hard, and AI art has helped in getting things from your brain into the real world in some meaningful way, I'd love to hear your experiences too.