Warning: Extremely obnoxious video. It’s funny because it’s true.

Last night, I spent the evening doing a few things with suno.ai. You see, some of you may or may not know this, but long ago, I used to write and record music. You’ll never get to hear any of this because the things that I wrote and recorded were tasteless and rude and made by someone who was, shall we say, less enlightened than the fully-formed man you see before you.

Compounding the crude and obnoxious songs was the simple fact that I wasn’t a very good musician. Oh, I could strum chords on a guitar, but that was about it. However, dare I say, there was a bit of musical talent in me that, had it been nurtured or been provided with the right incubator, might have actually turned into something. But that is a Randall from another timeline; one that I can neither see nor visit.

So, sitting on my hard drive, even as we speak, is a good five years of my life where I made music intensely. I was recording all the time with a couple of friends of mine (one of them very talented…I see you, CC). But I have made it all disappear. The music will never again see the light of day. It’s gone. Kaput. Done for. Hopefully forever.

But I recently read an article on how the session musician industry in Nashville is dying out because of artificial intelligence. It’s a very interesting read. See for yourself. (It might be paywalled). The basics of it is that instead of paying session musicians to record demo tracks to sell songs to high-profile artists, songwriters are now simply using suno.ai to flesh out their songs to sell. It’s kind of amazing what this program does. It takes the sketch of a song and turns it into an actual song, fully-formed, so that artists can judge whether or not they want to record them or not. It’s the decline of yet another industry due to AI (that of the session musicians), while simultaneously doing something wondrous.

So last night, I thought, Why don’t I try to plug my old songs into Suno to see what would happen?

Would the AI be able to take my old songs and flesh them out into something worthwhile? I didn’t know, but there was only one way to find out. So, I plucked down $10 for a one-month subscription and got to work.

I chose as my first tune a stadium rocker that I had written way, way back in 1994. It was rude and crass, but I always had grand visions for what it might be had I only the musical talent to make it happen. It was a three-chord monster, with a big chorus and grand ambitions.

To say that my original recording didn’t achieve that would be an understatement. The first recording I did of this song with a full band was in the foyer to an old, 19th century rental house in my hometown, and I still remember that moment to this day. I had been writing stupid songs for a while and had convinced a couple of local, younger musicians to play bass and drums alongside my guitar riffs. We recorded it one evening (I would assume we were pretty sauced) and then went to another friend’s apartment to do the vocals onto an old four-track recorder that recorded onto cassette tapes.

I was enamored with the results. Listening to them now, it’s just immature silliness, but it was so much fun. Later, I re-recorded the whole thing solo for an album and it never quite worked. I had to use a drum machine. The song had no bass on it. I mixed everything without reverb or any kind of effects. It had a very dry quality and never got where my ambitions wanted to take it.

But what would AI do to it? I was about to find out.

I uploaded the song and added some kind of prompt that emphasized that I wanted the whole thing to be a late-seventies rocker that would get an entire stadium up on its feet. I added a few other descriptions that I wanted it to have, and then hit “create”.

It was shocking what it produced: my vision made real. It did everything that I wanted it to do. It took my original chords and turned them into an anthemic soundscape that was true to what I wanted my original song to sound like. The drums were there. I asked it to add harmonies to the vocals and it did. It translated my lyrics almost perfectly (though the timing was off). It was a full-fledged song.

I quickly fired it off to my good friend, Chris, in a single text that said, “Oh, my God!”

Though it wasn’t my voice on the track, it was still something amazing. I thought, I need to take ALL of my songs and redo them! And I had dreams of songwriting once again. I was going to turn my music into something that I had always wanted…full songs with full bands. I quickly uploaded several more songs…

…and the returns were diminishing from there.

Every song that I uploaded was cool. It was fun. But it was like a weird, generic cover band every time. And I don’t know if I was reading into the whole thing, but it felt…soulless.

Tracey hated it. I redid a song that I sang for her one night, long ago in 1997, and she couldn’t stand what it produced. Chris felt the same. Both of them said there was something that was janky about the original recordings that made them much better.

And though it was cool to hear those songs in a new light, I have to say that…they’re right?

My original recordings were terrible. The songs were terrible. But they were unique little pieces that exist in their own sphere. My flaws are what made them wonderful (and horrible). And it wasn’t just the songs…it was the way the songs were made. Recording that original track in that house in 1994 was what I remember most about the song. It wasn’t the finished product. It was the method by which it was made.

Sometimes, it’s not about the end-result of something. Sometimes, it’s about the process by which the end result was made that makes it special. There were memories about the recordings of those songs that made them unique. And whatever ethereal thing that created those memories was somehow embedded into the songs themselves. We recorded that song in the storm drain under a university parking lot one day. Tracey played saxophone on that song in my bedroom. My sister recorded the dialogue for one of those songs. Chris added an incredible harmony on that song.

All of those things are the patina of an existence that was real and that texture is what makes those things art, whether it is good art, offensive art, or bad art. The bottom line is that it is art.

And once again, I discovered the same old thing about artificial intelligence…there is always an initial burst of “Wow!” when I see it in action. But after that feeling wears off, it turns into something else. Something not good.

I can’t really describe it other than it feeling off-putting. Whatever it is that makes us human is what makes art meaningful. It’s the flaws in our artwork. It’s the crooked lines or the off-beat singing or the slightly janky camera angle. It’s the poorly-worded sentence or the wrong choice of guitar chords or the paint slapped onto the page in a haphazard way.

I always tell my film students, “Try to imitate your favorite filmmaker, and the ways in which you fail are what make you a unique artist.” The failure is a part of the whole thing; the inability to reach your lofty goals is what makes us special. And in using AI to make my music perfect, it removed the humanity of the whole thing. It’s a neat party trick, but one that leaves you feeling pretty hollow at the end of the day.

So yes, I continue to experiment with artificial intelligence, but no, I still do not think it has any real worth to society, aside from occasionally taking a garbled jumble of words and turning it into a coherent email. This may be the worst AI will ever be, but I can’t really see it being anything better than what it is right now either.

Anybody who feels otherwise is deluding themselves.