Falling Down the Suno Rabbit Hole
Falling Down the Suno Rabbit Hole
Over the last year I've found myself using AI in ways I never expected.
As a piano technician, musician, teacher, recording enthusiast, and small business owner, I've been using ChatGPT for all sorts of things. Sometimes it's helping me think through studio workflow. Sometimes it's helping me organize website content, write FAQs, research piano technology, compare recording gear, troubleshoot technical problems, or brainstorm ideas for lessons and blog posts.
What surprised me wasn't that AI could answer questions.
What surprised me was how useful it became as a conversation partner.
Not because it tells me what to think.
Because it helps me think.
I can throw ideas at it, challenge assumptions, explore alternatives, and work through problems. In many ways it functions like a giant whiteboard that talks back.
Recently that process led me down a completely different rabbit hole.
Music generation.
Like many musicians, I was curious about Suno. I wanted to see what AI-generated music could actually do. I expected a novelty. Maybe a songwriting tool. Maybe a source of ideas.
What I didn't expect was spending several days obsessing over a song called Horizon Line.
The original concept was simple: write a song inspired by Amarillo, Texas.
That should have been easy.
Instead, it turned into an endless cycle of experimentation.
I'd generate a version in Suno.
Listen.
Take notes.
Then I'd open ChatGPT.
Not to ask it to write a song.
To ask questions.
Why does this section work?
Why doesn't this line land?
What if the chorus were simpler?
What if it leaned more Americana?
What if it leaned more punk?
What if the bridge was more personal?
What if I changed a single word?
Then I'd take those ideas back into Suno.
Generate again.
Listen again.
Take more notes.
Repeat.
Over and over.
At some point I realized I was using AI to drive AI.
Suno would generate possibilities.
ChatGPT would help me analyze them.
Then I'd feed those observations back into the next round.
Meanwhile I was sitting in the middle making decisions, rejecting ideas, chasing themes, and trying to figure out what the song actually wanted to become.
One thing I learned very quickly is that the first result is rarely the interesting result.
The interesting result usually appears after you've listened carefully, asked better questions, challenged your assumptions, and followed a few unexpected paths.
The song slowly accumulated details:
Route 66.
Cadillac Ranch.
The Big Texan.
Freight trains.
Feedlots.
Pantex.
Palo Duro Canyon.
Local stories.
Old memories.
People I knew.
Some of those references would only make sense to people from Amarillo.
That was fine with me.
I wasn't trying to explain Amarillo.
I was trying to capture pieces of it.
What surprised me most was how much time could be spent debating a single line.
Entire conversations revolved around a handful of words.
I would find myself saying things like:
"I like this part."
"Let's take that out."
"Try a different rhyme."
"That feels forced."
"This line is better."
"Why does this work?"
"What if the song is actually about this instead?"
Sometimes changing one word completely altered the emotional feel of a section.
Sometimes a line survived revision after revision and quietly refused to leave.
Those usually turned out to be important.
The process felt strangely familiar.
Generate.
Listen.
Evaluate.
Adjust.
Repeat.
That's not very different from mixing a recording.
Or building a website.
Or editing a blog post.
Or tuning a piano.
The tools are different.
The process isn't.
One thing I think creators often misunderstand about AI is that it doesn't eliminate the need for creativity.
If anything, it increases the need for judgment.
The tools can generate endless possibilities.
You still have to decide which ones matter.
Taste matters.
Storytelling matters.
Curiosity matters.
The ability to recognize when something resonates still matters.
In the end, Horizon Line became much more than a songwriting experiment.
It became an exploration of memory, place, storytelling, and creative process.
More importantly, it taught me something about how I use AI in general.
Whether I'm working on my studio, my website, my piano business, a recording project, or a song, the most useful interactions rarely come from asking for answers.
They come from having conversations.
That's the lesson I'm taking away from this particular rabbit hole.
Don't just prompt.
Explore.
Question.
Refine.
Experiment.
Stay curious.
Sometimes the most interesting destination isn't where you intended to go when you started.
Sometimes it's what you discover after following the horizon a little farther than you planned.
Below is one of my favorite versions of Horizon Line so far.
It's not necessarily the final version—and that's part of what makes this process interesting. Every new prompt, lyric revision, and arrangement idea opens up another possibility. This particular version captures a lot of the Americana, punk, ska, storytelling, and Texas imagery that I was chasing when the project began.
If you'd like to hear other versions, alternate arrangements, and some of the many experiments that came out of this process, you can find them on my Suno profile. One of the most fascinating parts of working this way is hearing how the same song can evolve into something completely different while still preserving its core identity.
Thanks for listening, and welcome to the rabbit hole.
