@tyto4tme4l
I appreciate the encouragement, but I’ve never seen any real usage for AI when it comes to making artwork, aside from things like upscaling or to quickly add a simple background to a piece. There isn’t anywhere in the artistic process where automation really fits, at least not with my chosen mediums. I know when it comes to animation people have been wanting to automate the colouring process, but as far as standalone pictures are concerned, there’s very little a generator can do that a real artist can’t do better.
As for understanding the process, there’s no overlap between what a generator does and what a pen/mouse does, so I’m not sure what benefit my knowledge would have on an automatic process. Every generator seems to aim for the highest possible quality, so even if your prompt is lacking you’ll always be getting something with highly complex shading and whatnot, regardless.
@Parallel Black
It was about three hours, according to my files. I agree that there are still some issues with this picture, but I deemed it good enough to share. I’m still learning how to make better pictures, but the result is always affected by some randomness.
I encourage you to try using Pony Diffusion yourself. I’m sure, as a proper artist, you would be able to make pictures of much higher quality than mine. You would also have better understanding of a whole process.
@tyto4tme4l
By “nothing” I meant the fact that it’s AI in general; not a real image.
How long, exactly, did it take you end up with this Rarity? I can still see a lot of imperfections that wouldn’t have been there were this a real artist’s work, both small-scale and larger stuff. Melting jewellery, randomised details on the furniture and windows, inconsistent clothing and disconnected piece of mane. There’s also the fact that Rarity’s back legs aren’t affecting the shape of her dress, implying she doesn’t have any.
Any artist would have drawn Rarity’s whole body first and avoided that last issue, avoided the awkwardness with the horn, mane and the front of her dress, and etc etc etc you get the point.
@Parallel Black
I disagree. While using AI (specifically Stable Diffusion) to generate pictures usually requires much less effort than drawing them, it still requires a few hours of generating multiple pictures (usually around 100 in my case), and fixing the issues through inpainting, img2img and manual edits. It also requires some prior setup, knowledge and a decent GPU. I wouldn’t call it “nothing”.
It might be more accurate if applied to using Bing Image Creator (DALL-E 3), which has almost no barrier to entry, but it usually still requires writing a decent prompt as well as multiple generations to get something decent and the overzealous filter makes it very restrictive to use.
@tyto4tme4l
While I wouldn’t go so far as to call it a scam, I would call it a grift. You’re tempting people into giving you money in return for what is essentially nothing; anyone can create these images for free, of the same quality, of the same content, without ever even laying eyes on anything under your prompter tag. I’ve seen way worse examples, sure, but this is still the same situation.
@Background Pony #4153
Why would it be a scam? My pictures are clearly labelled as AI generated. Besides, I’m not selling them, I’m just giving an option to send me a tip.
I appreciate the encouragement, but I’ve never seen any real usage for AI when it comes to making artwork, aside from things like upscaling or to quickly add a simple background to a piece. There isn’t anywhere in the artistic process where automation really fits, at least not with my chosen mediums. I know when it comes to animation people have been wanting to automate the colouring process, but as far as standalone pictures are concerned, there’s very little a generator can do that a real artist can’t do better.
It was about three hours, according to my files. I agree that there are still some issues with this picture, but I deemed it good enough to share. I’m still learning how to make better pictures, but the result is always affected by some randomness.
I encourage you to try using Pony Diffusion yourself. I’m sure, as a proper artist, you would be able to make pictures of much higher quality than mine. You would also have better understanding of a whole process.
By “nothing” I meant the fact that it’s AI in general; not a real image.
I disagree. While using AI (specifically Stable Diffusion) to generate pictures usually requires much less effort than drawing them, it still requires a few hours of generating multiple pictures (usually around 100 in my case), and fixing the issues through inpainting, img2img and manual edits. It also requires some prior setup, knowledge and a decent GPU. I wouldn’t call it “nothing”.
It might be more accurate if applied to using Bing Image Creator (DALL-E 3), which has almost no barrier to entry, but it usually still requires writing a decent prompt as well as multiple generations to get something decent and the overzealous filter makes it very restrictive to use.
While I wouldn’t go so far as to call it a scam, I would call it a grift. You’re tempting people into giving you money in return for what is essentially nothing; anyone can create these images for free, of the same quality, of the same content, without ever even laying eyes on anything under your prompter tag. I’ve seen way worse examples, sure, but this is still the same situation.
Why would it be a scam? My pictures are clearly labelled as AI generated. Besides, I’m not selling them, I’m just giving an option to send me a tip.
Edited