This article is a response to a fascinating article by musician David Rovics, in which he discusses how he crafts quite brilliant music using the AI music generator Suno:
For context, I recommend that you read this article and also listen to Rovics’ AI generated music. To my ear, he is absolutely correct that he’s able to create excellent music in this manner, and also the fact that his (human written) lyrics are often scathing attacks on the very technology he’s using to create the track is witty and ironic, and makes his critique much more effective, if only by demonstrating the reality that this technology is formidable. In fact, many of his lyrics are not just witty, they’re brilliant. By the time you get his point, you’ll be wishing you didn’t. (The truth always hurts when it hits this close to home.) I get the sense that it is precisely this irony which is Rovics’ motivation for this project. As an artist, he seems quite interested in our complex and changing relationship to technology, and the effect (especially the economic effect) this is having on artists and the arts. This project seems to be his way of exploring these questions by entering the belly of the beast, by fighting fire with fire.
The article “prompted” me to think about why I, personally, have zero attraction to working in this way, either to create music or imagery for my films.
The first question I had is technical: in Rovics’ description of his process of laboriously crafting and re-crafting verbal prompts for the music, it sounded like, for me, it would simply be faster and easier to play a musical idea myself than to use a generator like Suno. Generally speaking, if I have a musical idea in mind, I can immediately play what I’m thinking of, or I can get pretty close to it pretty quickly, so it just seems as if I can get better results faster by playing it myself than I could be writing and re-writing AI prompts. Even if I wanted to work with a singer to sing my music for me, rather than singing a song myself, it seems like it would be faster (as well as more interesting) to tell a singer how I would like her to change the emphasis and phrasing of a melody, rather than trying to goad Suno into doing the same thing through tricky prompts. It would interesting to hear Rovics perspective, and, as a very experienced musician, how he compares the speed and convenience of realizing his musical ideas by simply playing them himself, as opposed to using Suno. (This is leaving aside the question of how to get the polished, professional sound of a studio-produced track. It is obviously cheaper and faster to do this with AI than with human musicians.)
Looking more deeply into this question, however, I began to think about the more fundamental reasons I’m not interested in working with AI tools to create music or imagery. Writing and performing good music is notoriously difficult. It is difficult to craft a composition so that it works in just the way you want it to, and it is equally difficult to develop the physical, technical and mental skill to sing or play an instrument that fully realizes your idea. And it turns out that it is precisely this difficulty that I find most valuable about making music (or film). It is actually the central reason why I make music to begin with.
For me, being an artist is an opportunity for self-exploration. As readers of this SubStack know, I am attracted to using improvisation because it is a way of diving into the vast and fascinating pool of the unknown which resides inside of me, and inside of my collaborators, and gaining the skill and finesse to express the process of self-discovery in ways that make engaging, exciting music and film. Certainly, as Rovics describes his process of using AI generated music, there is a certain amount of self-exploration possible in that process, but for me the most direct and powerful way of getting right into the depth of my self-exploration will still be putting my hands directly on the piano keys. There is no more direct form of self-engagement than this. It turns out that the challenge, the difficulty, of making the effort to find “the good stuff” within me to generate my music, is the entire reason I’m making music to begin with. Without the opportunity to make that effort and learn about myself, I wouldn’t get anything out of composing. As John Barth wrote, “The key to the treasure is the treasure.”
I also do not discount the difficulty posed by physical performance: the difficulty of the body, of pianistic technique. I have always considered myself, in fact, to have pretty bad technique, as a pianist, probably stemming from the fact that I began playing at age 11, which is pretty late for piano. Throughout the 40 plus years that I have supported myself solely as a pianist, I’ve been continually amazed that people actually pay money to listen to me play, considering my limitations. Certainly, my technical limitations as a pianist are part of what has drawn me to improvised music. Of the two sides of my musical personality, the composer and the performer, my skills as composer are so much stronger than as a finger-crunching pianist that I have been able to make a living playing improvised music precisely because this is the form of music in which my musical ideas, my (spontaneous) compositions, count for much more that my ability to run up and down scales with perfect control.
For this very reason, I deeply value the difficulty of learning to play my own music well. I have known for a long time that finger technique is my weakest area as a musician, the area I need to work on the most assiduously. For this reason alone, the fact that it is quite hard for me to play my own ideas myself is what makes me more interested in doing it that way, and completely uninterested in generating beautiful renditions with AI. If I used AI tools, I wouldn’t have any opportunity to practice and improve my playing, and so it has very little value to me.
This is not simply a matter of looking for opportunities to practice and improve my piano playing. Because of my perspective as a person who doesn’t have the most fluent, easy piano technique in the world, I have a deep respect for the resistance posed by the body, for the challenge of trying to get my fingers to respond instantaneously and flawlessly to my musical ideas and impulses. To practice playing the piano is, for me, to forge deeper and stronger communion between my mind and my body. It is to come closer to being a fully integrated human being. As an embodied creature, a person who will spend my entire life living in my body, it is of (literally) vital importance that I strengthen and deepen my mind/body connection every day of my life. Among other things, I believe health depends on this. There are very few illnesses and injuries that can’t be effectively addressed by increasing physical awareness, increasing consciousness-in-the-body. Once again, the key to the treasure is the treasure. Every opportunity to experience one’s body in more nuanced, more subtle ways, is an opportunity to discover health and to become a fully integrated person. Learning to play an instrument well is a powerful tool to accomplish this.
My articles on improvisation technique are freely available to all, but are supported by monthly and annual donations from readers. Please consider becoming a paid subscriber to support my work. Thank you.