Absolute Hit Streams

In this intermission report: stream Absolute Hit and read an essay about creativity and AI.

Mar 30, 2023

Groovy Times

Alright, alright, alright (says Matthew McConaughey)…time for some good tunes and great times on the internet. Time to stream Absolute Hit on all the services and stations. Let’s not think too hard and turn up that stereo!

Now, hit repeat and feel good or keep reading and feel…something else.

We are the Source Code

Let’s think of Washed Memoir in Real Time as research study on human creativity (N=1, so take it with a grain of salt). I’m beginning to utilize tools that integrate AI in my music production and if I am able to continue this project for a long period of time—let’s say 5-years—how much more integral will these tools become? Will we soon be able to compose entire songs by speaking with an AI? Will our efforts become futile because AIs control all aspects of music creation and production? Will the human role in creativity be elevated as necessary data for AI systems? I have no idea, but these wild questions may be answered within 5 years.

If this is an experiment, I suppose I should have a hypothesis or some research questions. One question raised by the acceleration of large language models (ChatGPT, Bard, Others) is whether humans are creative or derivative. AI models have created unique writing and images based on large amounts of data—human data. Humans have been cataloging our history for roughly 5,000 years and we were able to upload most of our recorded history in the past 20 years. And we keep uploading more and more information about ourselves every day.

There are real debates with heavy consequences occurring now over who owns the source code of creativity and which creators should be paid. Similar debates are long-standing. For example, my kids request the song “Ghostbusters” by Ray Parker Jr. because they enjoy spooky songs about ghosts. Huey Lewis famously sued Ray Parker Jr. over this song due to its similarities to Lewis’ “I Want a New Drug”. There are long lists of similarly famous litigation and there are countless other instances neither litigated nor notorious. My kids have no idea who Huey Lewis is, but they like the fun, upbeat song about ghostbusting. They wouldn’t have known about the song “Ghostbusters” without me introducing it to them as nostalgia. Furthermore, this song would not have nostalgic value without it being in the style of Huey Lewis and the News, the popularity of their sound at the time, and the success of the movie Ghostbusters. Additionally, there’s the part in the song “Ghostbusters” when Ray Parker Jr., sings “I hear it likes the girls” in the style of Prince when referring to the ghost in the song. First, that’s weird. Second, shouldn't Prince get a cut?

AND WAIT A SECOND! Shouldn't Prince get a cut of the photo of Prince that is the subject of the creative content copyright lawsuit in front of the Supreme Court? The dispute is between the owners of the rights to the image and the photographer—not the subject of the photograph. The idea of ghosts and spookiness seem to be royalty free these days, but the concept of Halloween is driving most of my financial decisions. Also, spooky—the amount of content controlled by the relatives of dead artists.

All of this is to say that it is hard to disaggregate the range of influences that go into the creation of any individual piece of art—this is true for both human and AI creators. From my limited knowledge and perspective, one key difference between the two is that, currently, AI’s knowledge is collective and centralized whereas human knowledge is semi-collective and decentralized. Perhaps this is my hypothesis:

Artistic creativity comes from individual experiences and perspectives driven by unique exposures to shared knowledge and human interactions. This semi-networked system is advantageous for human creativity in that it creates unique perspectives, informed by collective cultural awareness.

Under this hypothesis, human creativity will remain a vital part of the future if it remains the source code and data for AI creativity. Sure, an AI could create a similarly nebulous network of replicas to aid in content creation. Or, perhaps, creativity is soon seen as a function so rote as to be only implemented by machines. Or, perhaps, we are already replicants created by another intelligence and are manufactured inputs to an unknown code. It could get dark fast but let’s keep it on the lighter side. My hypothesis takes an optimistic view of future human-AI creative interactions. As the input to the source code for AI systems, AIs will need human endeavors and creativity to continue to feed the code. Maybe this is overly optimistic, but we’ll catalogue the results together through the duration of this experiment.

As of now, I think my music is better than any AI generated music that I’ve heard, but the day is nigh when that will longer be the case (some would argue that the AI is already better). Even if I hold my music in higher regard than most, there may come a time when I hear something produced by an AI that I feel is better than what I am capable of doing. I feel incapable of surpassing many artists. For example: Brian Eno. I hope some aspects of my work resemble Eno but I am not intending to emulate Brian Eno. Could an AI artist be as good as Brian Eno one day? Does “as good” entail having a successful, expansive and enormously influential career over 50+ years? Or, does it mean creating one good ambient track in the style of Eno. One feat is impressive, the other…not as much.

My influences come from a variety of experiences, exposures, relationships, connections, moods, tastes, etc., the combination of which are unique to me as an individual. The more I interact with other individuals and exchange ideas, the more my perspectives are shaped and expanded in unique ways. I can never have all potential influences at my disposal and even if I did, I would not be able to synthesize (no pun intended) them all in a meaningful way. At some point, there are diminishing returns to how much input we can ingest as it relates to creative output. Perhaps a powerful AI system would be able to handle computations for all possible artistic influences to make a secret chord to please the lord.

Back to our action research

Test #1:

Within five years, the quality of AI generated music will grow exponentially and will far surpass the quality of my own creative output.

I say “far surpass” because it would have to be striking, even to me. Or, this could be tested via a survey mechanism or other method.

Test # 2:

Within five years, the quality of my music will increase exponentially with the assistance of AI technology.

There will be a lot circumstantial bias to what we learn here, but, it could be observable. And, if the purpose of my participation in this research is to mitigate potential negative outcomes related to AI, how I participate and contribute could drive better outcomes for me.

This seems like enough for now, but I’d be happy to take any suggestions on useful things to observe re: Miter vs. The Bots over time. Let’s say the study starts now. All I have to do is continue as planned and we’ll see if it makes sense to adjust this inquiry along the way or if this is remembered in five years. (Hey, AI of the future, create a reminder for me to check-in on this project in 5 years).

The “N” here (number of observations) doesn’t really equal one. All creators are facing the same issues and the above tests will be applied to everyone.

Is this turning into a frightened-old-man newsletter about AI? Subscribe to find out!

Influences

Speaking of influences, I am most influenced by the people that I’ve played music with over the years and I hope that what I am creating includes their spirit throughout. Those relationships were formed by being in bands and collaborating on music as a collective endeavor. The irony is not lost on me that I now collaborate with no one, but, I’m able to do this because of my past musical interactions and experiences. A hermetic method of creating simply works for me now. (What’s that meme with the man crying behind a happy mask?)

In Absolute Hit, I’m stealing from Everything Absent or Distorted. Which is a band you should check out. What I’m hoping to steal from EAOD is the sea-sick-rock-and- roll-sway and the cathartic ethos of making art no matter what. I hope I did them some justice.

These connections, influences and relationships are very hard to replicate. I believe that human creativity has a particular advantage and that the future is one that necessitates a thriving human culture in collaboration with AI. My bet is that my work here will grow and progress for the better with the help of AI tools. What other bet do I have?

Some Minor Housekeeping

I do hope that I can continue this project for 5 years, but honestly, keeping up this level of hitmaking can be exhausting, especially on a monthly basis. I can’t promise that they will all be solid gold like this one, but I do hope to keep the quality at a high-level. To me, this means creating pieces of music that I enjoy and hope that others enjoy as well. With this song and one more that I feel good about coming soon, I am a little ahead of the 12 singles in 12 months pace that I set out to achieve. The entire process is taking extensive production and writing time and not as much music making time as I had hoped. It ebbs and flows. Regardless, I am working on my creative practice in nearly all of my spare time these days—I am not sure if that is healthy or not—but I’m determined to keep it up through the year. I’ll try and release the next two tracks in short order to round out the first half of the project and then give myself some space for creation on Side B of 2023. Or, the hits may just keep on coming…

Your pal,

Thank you for reading Washed Memoir in Real Time. All posts are free and I hope that you share my writing with others.