Home Artists Posts Import Register

Poll

A note (and poll) on AI images.

  • I like the AI images. 6
  • I am neutral on the AI images. 11
  • I absolutely hate the AI images 10
  • 2023-12-28
  • 27 votes
{'title': 'A note (and poll) on AI images.', 'choices': [{'text': 'I like the AI images.', 'votes': 6}, {'text': 'I am neutral on the AI images.', 'votes': 11}, {'text': 'I absolutely hate the AI images', 'votes': 10}], 'closes_at': None, 'created_at': datetime.datetime(2023, 12, 28, 11, 56, 4, tzinfo=datetime.timezone.utc), 'description': None, 'allows_multiple': False, 'total_votes': 27}

Content

Hey guys, just a friendly reminder that the AI image posts are NOT a replacement for audio posts, they are merely a way to visualize the themes and concepts of recent audio releases. They're basically bonus content.

If you re-read my intro page you'll note it has said "2 audio files per month" for about four years now and that has been fairly consistent, though you all know sometimes I do more files than that.

I had a couple folks drop off patreon over the recent Christmas AI images, one of them was particularly angry over it. So I want to get your opinions on the AI image posts.

I understand that AI images aren't everyone's cup of tea, I didn't like them originally either. But as someone who is not a visual artist I thought they might be a neat way to enhance my audio releases. (I certainly don't make anywhere near enough to actually afford a flesh and blood artist).

If you guys really hate AI images I can just got back to not generating them.. which would mean you'd just get the two audio posts a month on average.. if that's really what you'd prefer.

Comments

Adria Reed

I’m personally here just for the drone audio and don’t engage with the bimbo stuff. But that’s just me. That being said, generative AI hurts artists. I understand that it allows people who wouldn’t otherwise be able to create visual art to realize ideas without having to pay tons of money to an artist. But AI learns from images usually without the creators consent. And AI music is coming - this isn’t just visual artists being hurt by AI. Lastly, AI art, whether visual, written, or audio, is soulless and uncanny.

Anonymous

I’m going to elaborate on my neutral vote. (And some of this may sound like playing the devil’s advocate.) To me it is a hard debate and something in the eye of the beholder. Since all current AI is based on its training there is a lot of debate around copyrights and prior works. For things like ChatGPT, they use LLMs which are usually trained on things like websites, dictionaries, scientific papers, etc.; for things like DALL-E, Stable Diffusion, they use existing pictures, images, etc. So there are some people that say that this is just theft of the existing works. But what is your brain actually doing when it is creating works? For landscape art, for some it is setting up an easel and painting the scenery around them. But what is that? It is taking the input of your eyes and converting it into something your hands can draw. Computers don’t have eyes, they just have either cameras or being fed other imagery. For new written works, there are only so many different literary styles out there. Is it historical fiction? Current fiction? Science fiction? Dystopia? Romance? Documentary? Etc? Where did you get the ideas that you are coming up with? What existing knowledge do you have that you are using? What are your influences (and in some cases prejudices and biases)? In the same way LLMs are just a compilation of existing knowledge and a prompt of how to turn that information into a new output. So if you compare the work of a 6-yr old, to the work of a teen, to the work of a renaissance master or Shakespeare; what do you see? That as the person gathers more information the more skilled and detailed their output is; similar to using larger LLMs or image sets. So to summarize the past few thoughts, I’m going to quote Mark Twain: “There is no such thing as a new idea. It is impossible. We simply take a lot of old ideas and put them into a sort of mental kaleidoscope.” So how does that tie into the question of the poll? I don’t hate AI works from the moral or ethical perspectives that some may use to discredit AI. But if the person running the AI doesn’t put enough effort in, you can have images that look like the person has too many fingers, too short of legs compared to the rest of the body, and other traits that negatively impact the work. So it is all in the execution of the image.

Chaos Doll

Oh indeed, a lot of AI music tools are already here. For the most part it's still just buzzwords and marketing, but the creative industry is making a concerted effort to push in that direction. Music folks have just been trying to ignore it as long as we can. Honestly if you look at the music production side of youtube, it's all about selling things to musicians these days, new gear, new plugins, new production tools.. because end listeners stopped buying as many albums a lot of these companies decided to make the artist into the consumer, etc. Like I said above, I was originally fairly opposed to generative AI, especially when its being sold as original. Whereas I've been upfront that these are just goofy things I generated to enhance the vibe of my audio files.. so they are really just things I wouldn't have otherwise made. If people are fine with less posts, I can go back to that.

Chaos Doll

This is all pretty similar to my own thoughts on the matter and pretty much the same reasoning I had when I started generating the AI images in the first place. Everything is iterative, even human creativity, but some people really don't like to acknowledge that. Ultimately I'll abide the results of the poll, if people hate it.. they hate it.. I just hope they are prepared for me to go back to posting nothing but the two monthly audio files.