Slop or Not? Music-Making with Humans and AI

After a remote recording session with vocalists, sound engineer, and producer

This month, I’ve taken my song “Anger” on a journey from something I threw together with AI, to a fully produced song with human collaborators. It’s been amazing. It’s enriched my life and convinced me both that AI has a very valuable role in the creative process, and that skillful human collaborators can take a song to a level that AI simply cannot touch.

I want to trace the journey of that song as transparently as I can, which is going to be a little bit embarrassing because you’ll get to hear a very rough recording I sang into Suno without ever planning to share it with a human being.

Our story begins in July 2025, with a prompt from the Braver Music Network to share a song answering the question, “What does the world need to hear right now.” I wanted to write something about the toxic impact of anger on our civil dialogue, and the historical reality that it’s been deliberately used to divide us by players with a malicious agenda. But wow. That’s heavy. Who would listen to such a song?

Then the idea came to write it as a villain song for a musical. The intro kind of appeared in my head just as I was headed out with my family on a full-day road trip to a family reunion. The lyrics mostly wrote themselves as I was driving, and I jotted them down whenever we made a pit stop.

Usually, when I write a song, I have some sort of tune in my head. But I don’t remember having one this time. I just plugged the lyrics into Suno AI after we got to our destination, and I let it come up with a melody.

It took at least a couple tries before I got this.

Then I started getting feedback, because some things were working and others weren’t. My son Erik was especially helpful, suggesting that “anger” started to lose its impact after too many repetitions and I could add some punch with some good synonyms, like “outrage.”

I tightened up the lyrics. But I still didn’t like the melody. The notes need to rise on “anger” rather than fall. So I sang a new melody into Suno for the chorus, gave the program a style prompt for “Comical doo-wop style music, Vocals sound villainous“ and let the AI get creative. Here’s what I got. You’ll notice the lyrics for the bridge had problems. I got some more feedback, reworked the bridge, and had Suno do this remix. At last, the lyrics were right. Then it was just a matter of doing a bunch of remixes until I got my favourite version, which is what I then uploaded to YouTube and the Braver Music Network. It wasn’t perfect. I didn’t like how it slaughtered Ira’s name. But it was good enough for now.

The time it took for me to go from plugging my initial lyric into Suno to settling on my official AI version was all of about 42 hours. And I had something I could share on Youtube, that’s gathered all of 410 views in the last 11 months.

That was what spawned the idea of a musical. In September, when I found myself teaching a Braver Angels seminar on navigating media, I realized “Anger” belonged in the curriculum. Partway through that course, I started longing to write more songs for the same musical that could also be used in the course. I wound up starting that on March 3, with a reboot of the workshop.

In mid-April, I realized that this was the project that could help Summit Stages get some traction, and the music needed to be industry standard. I reached out to Daniel Blomberg, a Summit Stages subscriber and musical genius, and arranged to get started on the process.

I sent him the AI demo and described the musical, now set in 1950 with a time-traveling villain. We met by Zoom and he laid down this piano track that brought the song into the 50s. That took him 3 hours. A few days later, I was able to travel to the studio in Provo Utah where I had a blast watching Daniel and sound engineer Dave Zimmerman add the rest of the instruments. That took another four hours.

Lead vocalist Amy Whitcomb

Now we needed vocals. The right vocals. Amy Whitcomb lives about an hour outside of Nashville, which is a long way away. But one video was enough to convince me that she was the right one to sing Ira. Daniel booked with her and we connected by Google Meet last week. She used a tool, Audiomovers, to record remotely. It was so much fun watching her give life to the part! She didn’t just sing; she almost danced, and laughed about how much fun she was having with the character. We could have done just one take, and I’d have been happier than I am with Suno’s impressive but sterile vocals. Nonetheless, we did take after take after take. We also wanted to record video of the session but some technical glitches got in the way, so I can’t share that with you. But it was incredible watching her be Ira. We did about 6 takes of every section of the song, although the first take was better than Suno’s.

And then, last Friday, we finished it all up with villainous background vocals, sung by Aitana Alapa, Megan Murff and Neal Stucki. Once again, the connections felt real and joyous.

I’ve heard AI music described as “slop” before and that seemed, well, snobbish to me. Suno’s production has sounded really impressive to my newbie ears. But that was before I got to watch professionals paying attention to every. single. note. I’ve now watched them in post-production, finessing the timing and the pitch. I’ve been party to the discussions about why to use this particular instrument or that specific drumbeat. And I’ve also seen the passion in their eyes and heard it in their voices as they’ve talked with me about how deeply they relate with the message of the song. I’ve felt our hearts drawing together as we created something that was bigger than all of us. It’s been an experience I want to repeat, again and again.

That means the media workshop I’m launching next week needs to take off, because it’s costly. Especially when compared to AI. I calculate that the Suno version of Anger cost me in the ballpark of $2. And it was useful. “Anger” would not exist without it. By contrast, the human production of the song cost me just over $3,000. And it was an inspiring, connective, unforgettable human experience.

I think AI is a powerful tool that can enable us to make better and better music. And it can’t replace people. It can’t even come close.

---

Thanks for your support for Summit Stages! If you liked this post, please consider sending it on to a friend.

If you are not already subscribed to my weekly newsletter, I'd be delighted if you'd do so here.

And a deeply, heartfelt thanks to those who've contributed to my tip jar