I don't like AI-generated yiff. I want to feel the lust behind every stroke of the pen, the attention to detail necessary to add just the right amount of excessive_cum, and the passion and love in the characters' eyes as they fuck each other silly. Sure, there are artists whose style isn't as "refined" or whatever, but they spend so much time and put so much effort into drawing their naughty little pics, and that alone makes them so much better than AI-generated slob.
Sure, there are artists whose style isn't as "refined" or whatever, but they spend so much time and put so much effort into drawing their naughty little pics,
I’m not even a furry, but have occasionally gone down rabbit holes where I ran into a random artist’s collected works. It’s always interesting seeing how their style developed and refined over time. There are some artists who clearly started out by scanning pencil drawings, then later transitioned to digital. It’s always funny seeing the style have a hiccup when they make the transition, and suddenly have to figure out how to draw in a new medium that they’re not used to. Then they figure it out, and continue refining their style.
But I also work in an art gallery, so maybe I have a habit of looking at things that the average gooner would miss during their three hour long goon session.
Yes, that's one of my favourite things to observe! There are so many artists in my saved searches who I'm excited to see progress over the years, assuming they keep making art.
Yes, there are furry-trained stable diffusion checkpoints available, as well as at least one paid service that offers furry generation as a specific feature.
Likely yes, the hentai model can do some wild things, but not everything (whoever spams the "anal prolapse" tag with poorly generated AI slop on Pixiv: Fuck you!). The furry model would need to be trained to keep animal features consistent too.
We don't need any more furryGPT, we already have diffusion at home.
All the best researchers went to Anthropic and elsewhere. The difference between proprietary and open weights is meh now. SolarRoofGPT burns the Sol, and tech bro digi slave traders. Long live open weights Chinese AI.
i feel peeps will jus keep on investin - or the US gets scared by deepcent n funds the big evil ai company with taxes... cuz of warfare or whatevr - or "ownin the economy" somthn like that-...
They're trying, but i legit don't see how it would make a difference. Their only option is to try and lock down the domestic market by making ai work legally copyrightable by companies. By using US law to give companies the profits using stolen training data, they'd be basically subsidizing the industry. Investors will naturally subsidize the tech since they can now put it into big budget projects without fear of legal issues, and we'd see ai take over Hollywood blockbusters and shows.
Of course, such an approach would further undermine the US in the long run, as our rush to twist ip laws would not be embraced by everyone. Most countries aren't as hyper capitalist about their media, and flooding the market with low quality slop would only make the US entertainment industry more irrelevant globally.
It'd be yet another capitalist grift that destroys the empire, dooming our country to some truly dark times. There's always the possibility that those with power will see the wisdom in reformation, but I really doubt it. Our current generation of politicians are too brainrotted by neoliberalism to every compromise wealth for the sake of anything, including the long term survival of the country. It wouldn't take much to convince most of them to embrace breaking up the nation, which would mean mass death and war. I'm not optimistic that the end of the US would be anything more than a feudal shit show.
Just noting that image gen is done by s different architecture than an LLM, and the environmental impacts are not the same (in particular, much much less energy use than having a human make one image, and substantially less water use than going to the toilet).