Why I Make Music Nobody Asked For
The Human Frequency — Notes on making music by hand in the age of AI
By Roger Ricks
The Wood in My Guitar: Human Music and Guitar by Roger Ricks
I release a song every three or four months. The last guitar I built took a year and a half. In the time it takes me to finish one track, on one streaming service alone around ninety thousand fully generated songs are uploaded every single day. Measured on output, I'm the least efficient music operation in the country. I'm not competing with that and I'm not pretending to.
I've been writing here for a while about what machines can and can't do, and somewhere along the way I noticed I'd never answered the more obvious question underneath it. Not whether a machine could make my music. Why I make it at all.
So this one isn't about the industry. It starts a long way back.
When the feelings were too big
My father died when I was twelve. I was standing on the platform at the train station in Sargans when I was told. What followed was more feeling than a boy that age has anywhere to put. I didn't have the words for any of it, and I'm not sure I'd have used them if I had. What I had was music. It turned out to be somewhere those things could go, and it asked nothing of me while they went there.
I started on the organ, which is not the instrument anybody expects at the beginning of this story. Then I found rock, and rockguitars had something the organ didn't. Energy. I couldn't have explained then what I was getting from it, and I can now: it was the first thing that was as loud on the outside as it felt on the inside
The headphones on the wall in Lucerne
I was on holiday at my cousin's place in Lucerne, and I went into a record shop.
They had a wall of headphones, the way shops did back then, where you could listen to a CD before you bought it. I put a pair on and picked Get a Grip by Aerosmith, which had come out that year. It took the top of my head off. That energy. That voice. That power. I bought the disc and a completely new world opened up, and I don't think I've ever quite come back from it. I couldn't get enough. And then I started playing that music myself, on my guitar, loud, in my room, and I owe my neighbours from those years an apology I've never made.
Because here's what happened when I played. I could breathe. The worry went. For as long as my hands were doing that, I was free of everything, and nothing else in my life has ever done that for me, before or since.
That's the whole reason. Not a career, not an audience. A thing that let a kid breathe.
Which is why I'd like to pass it on
If music did that for me, the obvious thing to do with it is aim it at somebody else.
That's the actual project. Not to be heard, exactly. To get something out there that might land in somebody the way Get a Grip landed in me on a Tuesday afternoon in a shop in Lucerne, on a day I hadn't planned for anything at all.
And every stream I earn goes to children who need it. The amounts are small, because streaming pays what streaming pays and one independent artist releasing three or four songs a year isn't funding a hospital wing. I know the arithmetic. It's still the point. If a song of mine can make somebody's day bearable, that's one thing. If it also puts a little money towards a child who needs it, that's two, from the same three minutes.
The dream, if you want the honest ambitious version, is that enough people hear the music that the second thing gets big enough to matter. Not fame. Funding.
What I don't want
I should be plain about this, because it explains a lot of the odd choices further down.
I don't want to be famous. It has never been the engine, and if I'm honest it frightens me a bit. I don't play live, partly for time and mostly because I don't like being the centre of attention. I'm almost invisible on social media. I post next to nothing personal and show basically nothing of my private life, which I'm aware is a strange way to run a music project in 2026. It isn't a strategy. It's just what I'm like.
In the perfect version of this, I sit in my small room and create tracks and out there as many people as possible are listening. Not to me. To the songs. Not because they find me interesting, but because something in the music got to them.
I realise that's a slightly unusual thing to want in this business. It's what I want. Maybe later in life, I will try to play live...who knows, but for now it is how it is.
What that actually looks like on a Tuesday night
I have a day job and I like it. Then there's my family, and that time isn't up for negotiation either.
Which leaves the night. I record in a flat, so I can't put a microphone in front of a loud amplifier at eleven at night without becoming a problem for people who've done nothing wrong. Some things don't change. I've just moved from being the neighbour who plays too loud to being the one who's found workarounds. So there are three ways I get a guitar sound: a Kemper, a Line 6 Helix Stadium, or my actual Marshall and Fender amps run through the Universal Audio Ox Box that keeps the room silent.
Who else is in there
I don't play the drums on my records. I don't like programmed drums, recording a real kit in a flat is out of the question, and I'm not a good enough drummer for the songs I write. So I work with Chris Brush in Nashville. I send him the stems, he plays over them, and it comes back to me to carry on with.
I don't mix my own songs either. Joe Carrell, also in Nashville, one of the best mixers I know, gives them the last thirty percent I can hear is missing but can't put there myself. Thom Wettstein masters them. Sometimes the lyrics come from Emme Pierre rather than from me. And sometimes there's another artist on the track entirely, singing something I could never have sung.
Who's involved changes from song to song. It's rarely fewer than three of us and it can be a good deal more. Being outclassed is part of why I work this way, and it's the less interesting part. The real reason is that each of them brings themselves into it. Chris puts a fill somewhere I'd never have thought to, and that fill is a piece of him. Joe hears the song differently because he isn't me. A lyric from someone else's head carries a way of seeing that isn't available to me from the inside. So the song ends up with more people in it than me, and layers I couldn't have put there because they were never mine to put.
A generator would have handed me a drum track in about four seconds, and it would have been competent, and it would have been nobody. Ask a machine for a drummer with feel and you get the average of every drummer it has heard. Ask Chris and you get Chris.
The small numbers
Since this is supposed to be honest, let me undercut myself properly. My reach is modest and my listener numbers are more or less flat. I pitch songs to radio stations by email, one at a time, like a man applying for jobs, and I get almost nothing back. A couple of stations have played me and one small regional chart was kind to a song of mine in August. I was genuinely happy about both, and both are the exception rather than the pattern. I
And ninety thousand songs a day is not a number you out-produce. If I thought of this as a fight against that, I'd have stopped long ago, because it isn't one and I'd lose it. But I've come to think the comparison is just the wrong instrument. Volume isn't the unit. One person is.
If a song of mine reaches one listener on the right day and something in them shifts, then the four months were worth it, and so was every email nobody answered. That's not bravery about small numbers. It's what the thing is for, and it was always going to be counted one person at a time. It happened to me once, in a shop, through a pair of headphones on a wall. The rest is patience. What I'm doing now is planting, and I might not be the one who finds out whether anything grows.
And the machines
Suno released a version of its studio this month that builds instruments. Synthesisers, effects, a chat bar that invents you a plugin on request. It's genuinely impressive, and between that and everything else it can now do very nearly all of this, faster.
What it cannot do is need to. Nothing is at stake for it. No twelve-year-old on a station platform. No record shop. No neighbours, no day job, no family, no children waiting at the other end of the money.
And it's only ever one thing. However many instruments it plays, there's nobody else in there. My songs have several biographies in them, a different set each time. AI can produce the shape of all of this perfectly. What it can't be is a group of people who each decided to be in it, making something for one person they'll never meet.
Where I land
None of this obliges you to listen to me. Nobody owes an artist attention because he worked hard, and a boring song that took four months is still a boring song.
What I think is that in a few years the question won't be whether you can tell a machine made something. You mostly can't already. The question will be whether you care where the things you love came from, and that's a question about you rather than about the technology.
I've decided I care. I could be wrong about how much anybody else does.
So here's what I'd actually like to know: is there a song that arrived on a day you needed it? Not your favourite song. The one that turned up at the right moment.
Tell me which one. I'll listen to it. And if you want to hear what a handful of stubborn humans in three countries do after hours, my music lives here.
Sources & further reading
Deezer newsroom, 21 July 2026 — around 90,000 fully AI-generated tracks uploaded per day, more than half of all new uploads, and only 1 to 3% of what anyone actually listens to.
Suno Studio 2.0 release notes, 13 August 2026 — MIDI, a built-in synthesiser, effects, and generating instruments and plugins from a chat bar.
Kommentare