Posts

@Zealousmagician
I get what you’re saying, but that seems like a pretty vague definition for an artist.
I tend to generate a ton of images and then pick my favorites and post them, and I suppose whether someone else thinks that gives them more artistic value is up to them. But at least I don’t think that makes me any more or less of an artist.

Like if picking the favorite ones is what gives them artistic value or makes me the images’ artist, then what if I just generate 20 pics and dump them on a site without even looking at the results? Would they be less worthy of being called art, and would I even be considered as the artist?
Or if I send them to someone else and they pick their favorites and post them online, would they be considered as the artist, even if they had nothing to do with the actual image generation?

I’m sorry in case I’m just misinterpreting your point, but that just seems like a really vague way to define what is art and who’s an artist. Though I know those topics have been debated long before AI art was even a thing, so I’m not expecting you or anyone else to have a concrete answer to what counts as an artist/art.

I just think that the current prompter/artist/ai assisted artist labels are distinctive and descriptive enough for there to not be any need to start trying to redefine them.

@Clopxie
The mere act of selection based on criteria, of critique, is itself an art. If you generate many images and select a few because they are appealing to you in some way or inspire you, then when you share that, those images become a reflection of your judgement, of your personality- you have assigned them artistic value. Unless you are to say that they already had artistic value and that the machine is the artist, you must be the artist in this scenario because the only artistic meaning held in those images comes from you personally. An example of what exactly can qualify as art can be seen in the famous Duchamp’s Fountain and other readymades.

In your hypothetical, you aren’t an artist until you’ve applied judgement to the image generated and decided on if you now find it to have value.

I won’t argue for more than having significant edits be considered for the artist tag, because I know I would only be wasting my time. I only wished to add my thoughts on this part of the discussion.

That is such a weird argument. That’s like saying that a traditional artist isn’t actually an artist because the pencil does the work and the human only provides it with the inputs. What makes a 3D artist an artist is the fact that they use physical actions to move the mouse/pen to design or sculpt something.
If you make a humanoid robot that recognizes speech, and tell it to draw Mona Lisa, it doesn’t make you an artist. So why would it be any different with a web UI that understands text?

Whether it’s drawing, photoshopping, 3D modeling, or something else, I think getting to the final result by using your own brain and body to give the tools real-time input based on the progress is what makes someone an artist.

If you tell another person to draw something based on your inputs, that person is the artist.
If you put your hand on theirs and you’re the one controlling what goes on the paper, you’re the artist.
And if you let the other person finish the drawing, and then you remove the 7 extra fingers the other person drew, you’re not the artist of that drawing, you just edited art made by someone else.
Now, if you were to redo a significant chunk of the drawing, you could perhaps argue that you are the artist and the other person assisted you by giving you a rough idea to work with. Now that I think about it, I think there might be a tag for the AI version of that?

Sure, knowing how the prompts work does help, but it’s something you can learn in a day or two by scrolling through some guides and using a tiny bit of imagination when prompting. But you can also get some really good results with prompts that you put absolutely no thought or effort in.
A while ago I tried generating images with nothing but random verses of this song as the prompts. No quality prompts, negatives, poses, anything. Just random lyrics from that song. And it gave me some really good looking results that had next to nothing to do with the prompts.

As a bonus question, if I’m trying to prompt for a very specific image, but the AI interprets it wrong and comes up with something different, but way better than what I had in mind, does it make it make me a good artist for getting good looking results, or a bad artist for not getting the results I wanted?

Then the term is meaningless and you only want it as a little badge.

As an imageboard, the site uses different things to better classify things. prompter for raw prompting, editor for people that edit prompted works, and artist for people that create a significant amount of the image outside of AI.

Stop being so emotional about the label.

@LemonDrop
I doubt that would make everyone flock to this site right afterwards. Rather would try to push enabling style LoRAs or something useful. Twibooru’s trending is far too cursed usually.

@Delly
Sorry I thought I was using the definition the world uses:

Also there’s already alternatives like Twibooru, and people are already preferring them over this site. If you wanna kill the site though be my guest. I am just trying to give some advice.

@LemonDrop
Cool.

Create your own platform with your own redefinitions and interpretation of terms. The project this site runs on is FOSS, so noone’s stopping you.

The rules will not be changing here.

@Clopxie
That doesn’t make much sense, in any digital art the computer is doing all the generation of the information. The human simply provides input as all these things are tools. For digital art the tool is something like Photoshop and the inputs are a series of tablet motion inputs, for 3D art the tool is Blender and the inputs are bit more complex but similar, and for AI art the tool is some AI model and the inputs are a series of prompts and etc combined with potentially other digital art methods.

The computer is always the direct “creator” of digital art as far as what entity is actually putting the values into memory, but in terms of creator we mean which human artist was the one driving the process. As I said before, the purpose of an artist tag is to find more similar stuff to such work, the AI model should be listed too to find more art that was created with such a model in a similar fashion, but the creator or artist is referring to the human behind the process as that’s what people are interested in finding more of in that case.

This is why everyone here is an artist fwiw, you are driving a creative process so you are projecting your vision into what the AI is creating just like any other digital art process. It is not like these are just totally random outputs without any artistry behind them, even a single word prompt is enough to start guiding something in an artist direction. All these methods involve iterative refinement of ideas to reach some “goal” you have in your mind, they just use different methods of doing so. Acting like art needs to be “hard” to be valid is just nonsense, the true ideal of art would be to beam people’s thoughts directly into some sharable medium such that people can express themselves creatively without that being “locked” behind a difficult amount of mechanical skill or something that some do not have the time or ability to achieve. AI is just one step in getting closer to that ideal. Of course the most impressive art will always reflect what has the most effort put into it or is the most creatively inspired which is just something that will always vary, but yeah it’s certainly wrong to discount art just because it’s “bad” or because it’s “trivial” to create.

@Bendy and Boney
Done. I did former into latter, can always be reversed if that name is preferred.

https://tantabus.ai/search?q=generator%3Aautismmix+confetti should be an alias for https://tantabus.ai/tags/generator-colon-autismmix_confetti
(and https://tantabus.ai/tags/generator-colon-autismmix+pony should be one for https://tantabus.ai/tags/generator-colon-autismmix_pony)

@Zerowinger
Done.

@Background Pony #4F98
And done, except the OCs that imply it, because I don’t want to look up each individual tag as I am unsure which ones are here or not, and I know I can look at Derpi for what it has, but what do we have?

Linking things directly helps save a lot of time and effort.

Should imply “twilight sparkle” and “alicorn.”

@LightningBolt
Well on derpi its

Aliases: equestria at war, eaw

Implied by
oc:river swirl, oc:starry night (eaw), oc:rosa maledicta, oc:grover vi, oc:ferdinand dawnclaw, oc:alexander kemerskai, oc:giulio beakolini, oc:archon eros vii, oc:gabriela eagleclaw, oc:wingfried von katerinburg, oc:auburn leaf, oc:lucent eclipse, oc:caramel marks, equestria rises still (equestria at war submod)

Detailed description:
https://steamcommunity.com/sharedfiles/filedetails/?id=1826643372

And it should have same color like “them’s fightin’ herds” tag aka pink

Personally I feel like “prompter” is the most fitting description. I definitely wouldn’t consider myself an artist, even though I often do some editing on my pics. Like removing extra fingers and adding missing ones, or removing extra legs wings or other issues like that. At least for me, that amount of drawing and editing isn’t enough to qualify myself as an artist.

Even “creator” sounds a bit odd to me, as I’m telling the computer to create something instead of creating it myself.
If I give the computer prompts and it does the work, I’m a prompter. If I were to draw a pony and generate the background, I’d consider myself an artist who creates AI assisted art, so I’d be happy with being called either an artist or a creator.

So at least in my mind the current tags make perfect sense and there’s no reason to start messing with them.

@Adorable Blue Fox
Spoiler images will likely not be carried over so those won’t be part of it, but that could be useful otherwise, thanks.

@SpikeyTum
Thank you for finding that! There’s so many little details in philomena that are baked into the code. Maybe because the idea that ‘other sites will have authors instead of artists’ is just so different than what it was originally created to do that making it ‘a thing that you decide and configure once’ seemed like a waste of work.

So, thanks to all of you for putting up the ‘still moving in’ gotchas here, and for letting us know when you find something weird or wrong.

IDK if this is useful or makes anything more manageable, but I’ve compiled a list of tags with missing details, such as the spoiler image, implies and the category.

I’ve also got a list of tags that also check for the descriptions and aliases.

The list is extensive, so it is sorted based on the derpibooru’s tag image count.

@Background Pony #4F98
It is known, low priority.
@Background Pony #4F98
Please provide more detail.

Derpibooru exclusive is for art uploaded solely to Derpibooru by its original creator. Images on Tantabus originally on Derpi were automatically imported and can’t be considered the prompter/artist’s official upload. Also, images here originally on Derpibooru will remain there as well until January 6, invalidating images that were imported by their creator.

Same as “equestria at war mod” dont have it short names and pink color

All (or well most of not sure about all) “Stupid sexy X” missing “X” character

@tyto4tme4l
Done.
@Zerowinger
Also done.

Backing up a bit from the deeper conversation here, I’d just like to please suggest that the upload page is updated, so that the primary prefix is used when using the ‘fetch’ button.

Right now, it defaults to “artist:”, so I have to change it to “prompter:”.

https://tantabus.ai/tags/alicornified
Should imply “alicorn” and “race swap.”

https://tantabus.ai/tags/starlicorn
Should imply “starlight glimmer” and “alicornified” and its associated tags.

https://tantabus.ai/tags/trixiecorn
Should imply “trixie” and “alicornified” and its associated tags.

https://tantabus.ai/tags/shimmercorn
Should imply “sunset shimmer” and “alicornified” and its associated tags.

“Drool” is missing aliases “saliva, slobber, drooling”
https://tantabus.ai/tags/drool