Comment by smilbandit
11 hours ago
I dabbled with AI music for a bit with Suno. Worked out well for the most part, only way I'm ever going to hear music with themes for some of niche things I like, like Shadowrun. I threw a bunch of music genres at it and some were good enough that I added them to my normal playlist but after about 30 completed songs I had a hard time coming up with new stuff. As someone who has never tried to create music myself it was fun to play with.
There are two arguments about AI art, one of them is trivially reducible to the “is sampling/collage art”. If you are spending time expressing something using AI produced components then you are producing art, and probably the amount of time you spend working on it (either developing your skills or creating the work) roughly will correlate to how much value others see in it. It’s no different than building a hip hop track out of drum loops.
The second question is more interesting, which is “does raw AI produced artwork have any artistic value” and I am going to punt on the “artistic” part of that equation and answer the “value” part with no, and not because people might not enjoy it, but it falls victim to the classic “my five year old could so that” critique of modern art, except in this case it is true. Anybody can go to an AI and produce some mediocre media.
Where this gets interesting again is _volume_. What AI unlocks is exactly that anybody can create songs, videos and images for _themselves_. The value of it is probably the pennies worth of time ajd expense they put into it, but it might he worth it for them to make something, be mildly amused by it and immediately dispose of it.
You wanted some shadowrun themed music, you got it and enjoyed it. You made something of value only to yourself, but that seems okay? Multiply that times billions of people probably eventually people might luck into something genuinely good and worth sharing from time to time.
> people might luck into something genuinely good and worth sharing from time to time.
A) it would be impossible to find in a sea of AI generated slop
B) even if it were to be recognized as good, it would be instantly copied by other AI’s such that it would be very shortly thereafter be also considered slop
For any work to gain traction with an audience, there needs to be scarcity. Art and artists are valued because they are unique in some way, something about it or them cannot be replicated by others. The ability to instantly produce a piece of “art” negates any artistic value, at least as far as audiences are concerned.
I think your starting definition of value is basically worthless. Value is not about what things cost to make, but what people are willing to pay for them. You reached this conclusion by the end of your comment, but I think it's important to emphasize. My friend group has created incredible value with suno, mostly making meme songs we forget about after a day, but every once in a while we create lasting memories that have real emotional impact. It doesnt matter that anyone can do that, I dont think that cheapens the output at all.
My comment was mostly that it lacks value to _others_. It was probably worth to your friend group roughly the time and money you spent on it. Nobody else is ever going to care.
1 reply →
> is sampling/collage art
Yes.
You will owe royalties.
The latter part is the actual problem.
[dead]
I played around with Suno a little too. It's actually kind of crazy what it can produce. I mean I don't think it's objectively good in any way but for many applications this (and a lot of other generative AIs) are what I'd call "sufficiently good".
When you move into an apartment or furnish a rental or whatever you might put stuff up on the wall. For many years that might just be some mass-produced prints from IKEA, for example. These might be photos or paintings but a lot of them are "abstract". For this kind of application, current generative AIs are probably sufficient to create what I'd call "wall fillers".
So if you were doing an indie game, it might not be large enough to pay for artists to come up with music or even some basic art assets but an AI can I think fill this role. You can use them as placeholders.
So I'm generally sympathetic to the plight of artists. There is certainly an issue with how these LLMs are trained and if that's "stealing". Legally and ethically we're still working this out because the issue is new.
But I also think there are some things you just don't need an artist for.
So this kind of music has a name, it’s called production music and it’s been long expected in the industry that AI-generated music will compete with the lower-end production music, basically elevator music or background music for corporate training videos etc. It is unlikely, however, that it will get much traction in scripted long form productions, partly because studios believe it’s a legal minefield, and partly due to resistance from creatives (whether justified or not).
My favorite example of resistance from creatives was the space shuttle landing gear button. The space shuttle orbiter was technically capable of performing an automated mission, with the exception of opening the landing gear doors. This was ostensibly so that there would be no risk of the heat shield being compromised, as the landing gear doors were in the heat shield. But it is widely acknowledged that this was an effort by the astronauts office to ensure the continued need of a human crew.
For what it's worth, I support manned spaceflight. But sometimes allowing "creatives" to impede progress has its costs.
2 replies →