Notes

A Response to Adam Neely's "Suno, AI Music, and the Bad Future"

You can view the original video here. I would highly suggest it. Adam is one of my favorite creators in the music theory space, and his channel is filled with fantastic content.

https://www.youtube.com/watch?v=U8dcFhF0Dlk

Dear Adam,

I broadly agree with your high-level take in your video on Suno: most AI art is not very good. But I think this video has a lot of issues that make the specific arguments it makes not super compelling, or at least in need of work. For context, I'm speaking both as someone who works in the tech world, and also someone who studies as an art historian. Broadly, I despise AI art.

  1. On Iteration: I don't necessarily think there's anything wrong with speeding up the iteration process. You could argue that DAWs did this as well. There is obviously value in the iterative process but often taste (and refinement) can be good enough to create a song that's genuinely compelling. If the idea (chord progression, lyrics, vocals, etc.) come from someone but that person has limited technical talent, it's not entirely clear why their output would be of lesser value than someone who had the requisite talent to create and produce the song. Also, spending a bunch of time talking about how impatience isn't a virtue seems self-indulgent. It's obviously a play on words, but also, it's not difficult to make a compelling argument that impatience is moral (i.e. an impatience to right a past wrong would certainly be virtuous). Many creatives, like journalists or authors, often see patience as a virtue—it helps them avoid perfection and finalize their work. Plato's conception of virtue doesn't necessarily hold true all of the time. We might "delay pleasure" by working for charity instead of indulging in the act of music creation, but that doesn't mean that music creation is useless or has no societal value.
  2. On Voice: I generally agree with your argument, although I disagree that it's impossible to generate novel creations with the use of AI. Obviously if I came into Suno with the entire song written, it could be novel... so where do we draw the line? If I come into Suno with a novel or new chord progression, can it create novel and interesting music? To use your 1970s pop metaphor, if I asked Suno to introduce record scratches, could it? These examples are extremes, but there are certainly workflows where the artist is still producing novel, valuable music with the help of the AI. It's just that most users don't use it this way, which is the actual problem.
  3. On influence: This whole argument feels a bit unfair, no? First, you seem to have asked for AI creator influences and I would challenge you to name any AI artists who the public is broadly aware of. The whole point of AI is that it allows people who know nothing about music to produce music for the first time. However, these people could probably name the musical artists who they listen to, or more broadly, the music that influences their perception of taste. With AI tools, the job of the "artist" is reduced to curation or ideation. But just because they are using AI doesn't mean they strictly need to be influenced by other artists who use AI. This would be like asking someone "who are your favorite FL Studio artists" and then berating them for having no "role models." Also, more broadly, I would just say your characterization of these people is a bit... mean. Just because someone doesn't have AI role models doesn't mean they don't have role models?
  4. On Narcissism: First, citing Reddit comments to broadly characterize narcissism in a community is a bit of a tautology, no? But more broadly, I just disagree with your premise here. When I create music, or photographs, or poetry, the final result is something I am proud of, or that I would listen to. Playing around with my op-1 is not groundbreaking, but I generally try to create stuff that I enjoy... and I don't think you would have a problem with that process?
  5. On Deskilling: This is the same narrative that people have used to reject conceptual artworks for years (since Duchamp in 1918!). Trying to argue that "because something takes less skill, it is not valuable to the world" is fighting an uphill battle. Your response to the "new skills" argument is deeply unsatisfying—you take the notion of "craft" and appropriate it to mean "without innovation." I would argue that many artists can't predict the output before they begin a creative process... that's the point of a process. That doesn't mean their work is craft-less.
  6. On Taste: I don't feel like this point receives a sufficient response. The main thing you say is, "I didn't dream of prompting, I dreamt of shredding" or some variant of that argument. But that's not... that doesn't interact with the argument behind "tasteful creation."
  7. On I Run: I find this argument funny to include this in the overall video as it seems to undermine some of the points made. As a musician you are obviously inclined towards the creative process but a lot of people just like... listening to music. I would argue that a lot of radio pop is derivative or otherwise doesn't provide a ton of value to the world of music. But it also exists, and I don't know that people should be shamed for enjoying it or listening to it. If AI is truly just derivative, then there will be a marketplace, sooner or later, for human-created music.

47:58