After days of fun with stable diffusion I figured it was time to slow down start in believing in an output enough to put in some effort. This was one of my favorites after thousands of images.

Changes made to the original output:
- Removed a cutoff moon that was grabbing too much attention
- Removed a deformed cutie mark
- Fixed some weird ai windows and lanterns on the roof she's standing on
- General touchup
- Upscaling
- Canvas filter

I'm not sure what the final "ethics" and community of ai and prompt sharing will turn out to be. Personally, I think it's important to share prompts to give a baseline to go off of and show what works. Once you start making a piece your own, I feel you're more entitled to keep the seed and finetuned parts to yourself. It takes away a little of the magic when you're able to make an exact comparison to the original output. There's also the problem of admitting to including living artists in your prompts which has lead to a lot of discussion on the ethics of ai art in general.

A happy medium to settle on for these situations would be to share the prompt with the living artist tags removed and exclude the seed.

*derpibooru, (scootaloo), suggestive, (solo), [[[[mare]]]], [[[filly]]], [[[[young]]]], [[adorable]], pony, adorable, [[[[[[chibi]]]]]], sweat, (hnnng), heartwarming, dutch golden age, Flickr, gothic
Negative prompt: [anthro]*

And that's just talking about the ai community. It's an even bigger can of worms when you get into posting ai generated images on a site not dedicated to it. At the end of the day, we're the wedding DJs performing in a talent show at the school for the blind.