Raw LLM Responses

TL;DR - I've used AI. I understand using AI. Ultimately, I don't think people should use AI, at least in the forms it exists in now. If it were more open-source and cut from a cleaner cloth, it could be fine as a minor tool but not as a replacement for anything. Also, Nightshade is good, it works, it's very difficult to even try to counter which is why they've been failing to, and everyone should use it. As someone who's paid for bills, groceries, and rent with my art: genAI can theoretically have a place in the workflow. I've used it for three things, for the most part. 1 - to generate noise, not something that most people would default to thinking of but the same noise generation that makes static or plasma filters is what gets used in two steps (minimum) of genAI processes. 2 - to generate dynamicism: sometimes, I get stuck. We all do. I've generated an image a total of three times over the past year specifically to see what my brain _didn't_ like about them, so I could kickstart actual inspiration. A sort of negative-actualization. 3 - to generate thumbnails: I use a program that adds detail to random doodles, low-quality things about 144 px square. I then take the re-contextualized doodle, now a bad little cottage in a landscape, and see if I can make a better version. A way of using writing genAI that I'm interested in is self-referential analysis: seeing if there's phrasing or syntax I use or misuse in ways I'm not comfortable with. This'd only work if I were able to cleanly train one exclusively on my own material. If visual genAI were able to be clearly trained, I could similarly use it to see if there's things I put too much or too little emphasis on - people can always tell if there's a part of the body or fluid type you focused on when learning to draw. None of this is to say that genAI creates art. The little wooden mannikins people use are bad tools, AI is a worse tool: AI can be replaced with a more clean setup. 1 - if I had my old setup, I could do this step with my analog tools, which would not only be more ethical it'd give a personalized result. Until then, I'll probably stay using this element. 2/3 - if I had space and money for things like "inspiration dice" that could just solve these. Both of these really boil down to decision fatigue. It should be noted that there's plenty of free generation, and even things like Photoshop have inbuilt AI now. Subscriptions aren't an issue when the programs can be free (like PS vs GIMP vs Krita, it's just a different service) and people have the computers and tablets anyway. I think people get caught up in the hype-trains of these "trendsetter" developments without considering how few trends are ever here to stay.