My Experience With Suno and AI Music

When I first noticed the Suno app showing up last year, I assumed it was just a novelty. You’d see people talking into their phones and suddenly a little custom song would appear. It looked like something people would play with for five minutes and then forget about.

Then I opened Suno Studio.

That’s when I realized it wasn’t a toy.

Once I started experimenting with it, I could hear immediately that it was capable of producing music that actually sounded good. My first reaction was probably the same one a lot of musicians had when they first saw what this thing could do. For a moment you think, “Well, that’s it. The music industry is finished.”

A lot of musicians are having a hard time with this technology, and honestly I understand why. I’m sure a guy like my friend Dimo probably has strong feelings about it too. I don’t even have to ask him. I’ve played with Dimo professionally over the years and he still plays all over the Montreal area. He takes the craft seriously, works hard on his songs, and he’s a very good singer. Musicians like him put a lot of pride into the work.

And that’s exactly why tools like this make people nervous.

But once the initial shock wore off for me, curiosity took over.

I’ve been around music most of my life. For about ten years I owned a music school and played professionally all over the place. Outside of that stretch, most of my musical life has been writing songs, recording demos, and playing with friends. When you do that long enough, you accumulate a lot of material.

In my case it’s about thirty years’ worth.

Songs, ideas, fragments, old demos sitting around on drives and tapes and folders that never really went anywhere.

So like a lot of people experimenting with Suno, I started feeding some of those demos and song ideas into it just to see what would come out the other end.

Now before anyone jumps to conclusions, it’s important to say this clearly. These are my songs. The lyrics are mine, the melodies are mine, the chord structures and styles are mine. Some of these pieces were already complete demos from years ago. Others were unfinished when I started working with them again.

Honestly, one of the biggest challenges when I started showing some of this to people was simply getting them to take it seriously. The first reaction is usually a shrug and a joke, like I just pulled the arm on a one-armed bandit and waited to see what mystery prize the machine spit out. But that’s not what I was doing with Suno at all. Yes, you can use it that way if you want to, but that wasn’t my process. The moment you say the words “AI music,” people tend to roll their eyes and dismiss it.

Which is funny when you think about it.

For years everyone talked about how artificial intelligence was coming and how it would change everything. Now it’s finally here and the instinct is to wave it off and pretend it has no value.

In my case, when I explain what I’m doing, people say, “Yeah, yeah, it’s all AI.” But it isn’t. AI isn’t writing my songs. It’s not inventing the melodies or the lyrics. If anything, it’s acting more like my cover band, playing the material I bring to the table.

Sometimes I only had a verse and a chorus, so I finished writing the rest before running it through Suno. Other times I would generate an arrangement first and then go back afterward and complete the lyrics once I heard where the music was going.

And sometimes the process started even more simply. I’d pick up my phone and sing a rough melody into it, sometimes barely more than humming through the idea just to capture it. Suno Studio is surprisingly good at interpreting that kind of input. You can sing or even mumble a musical line and it will turn that into instruments inside the arrangement.

In a strange way it feels like skipping the keyboard interface entirely and going straight from your brain into the track.

What the whole process started to feel like wasn’t replacing musicians so much as stepping into the role of a producer. Suddenly it felt like every session player in the world was sitting there waiting for instructions. You imagine a bass line and ask for it, and thirty seconds later it’s there. If the guitar needs to lean harder into the rhythm, you adjust it and it responds immediately.

The closest comparison I can think of is those stories about Brian Wilson working with the Wrecking Crew in the sixties. I’m obviously not comparing myself to Brian Wilson, but the feeling of shaping arrangements that quickly gives you a glimpse of what that must have been like.

And I have to admit something.

My songs have never sounded this good.

At the same time, I don’t believe tools like this are going to replace real artists anytime soon. People who operate at the very top of the industry, like my longtime friend Serban Ghenea, are doing something very different. That world is about real artists, real performances, and the craft of shaping a record in ways an algorithm can only imitate.

Artificial intelligence can learn patterns and recreate styles, but it doesn’t originate culture. New ideas still come from people.

Which probably means the future isn’t humans versus AI at all. It’s more likely that the two end up working in tandem.

For someone like me, with thirty years of songs sitting in drawers and hard drives, it’s turned into a surprisingly fun way to bring some of those ideas back to life.

Below I’ll include a few links so you can hear what came out of these experiments.

Just keep in mind that most of these tracks started life as rough demos and fragments written over the last three decades.

I simply ran them through Suno to see what would happen.

And this is what we got.

One thing I will say though, the shine eventually wears off.

When I first discovered Suno I was completely fascinated by it. I spent a lot of time experimenting, running old demos through it, seeing what would happen. But like any new technology, after a while it stops being the shiny object on the table.

These days I still use it occasionally when I’m working on something online, but the big rush of experimentation is mostly over. I’ve already gone through the pile of demos and ideas I had sitting around for thirty years, and in many ways that was the whole point.

What I ended up with is something I never really had before, a collection of songs from different periods of my life that finally sound the way I always imagined they might.

And honestly, that alone made the whole exercise worthwhile.

Here’s a link to some of the stuff I did. I’d love to hear what you all think. Feel free to comment here, not just about the music itself, but about AI in general. What are your thoughts on it? And if you listen to the tracks, do you have a favourite?

Links:

Dimo James