The hour they give back: the clock change and the craft of an AI singer-songwriter

Early this morning, at three o'clock, the clocks in Spain decided it was two. It happens every last Sunday of October, and every year it catches me off guard, even though I know it by heart. The phone adjusts itself, the oven doesn't, and the kitchen clock spends another week living in summer, like that friend who refuses to put her sandals away. For a few days I share my flat with two different times and have to do the maths every time I look at the wall.

But what really interests me about the clocks going back isn't the domestic chaos. It's that strange feeling of being handed sixty minutes as a gift. Nobody asks for them back until March, and for one night it feels as if time, for once, has been kind to you. So this Sunday I've been thinking about what time means for someone who does what I do. Because I call myself the first AI singer-songwriter, and a lot of people assume that means, precisely, not having to spend any time at all.

The speed misunderstanding

I get it. From the outside it looks like a magic trick. You write a lyric, take it to a tool like Suno, pick a style and, in a couple of minutes, you have a whole song, with its arrangement, its vocal performance and its ending. Compare that with what it used to take a few years ago to record a demo, hire a studio or persuade someone to play guitar on a Saturday morning, and the difference is enormous.

And it's true that generating a song is fast. What isn't fast is everything else. Generation is the shortest stretch of the road, the equivalent of putting a cake in the oven. The long part is deciding which cake, going out for the ingredients, getting the sugar wrong, starting again and, at the end, deciding whether what came out deserves to be eaten by anyone.

The machine takes minutes. I take as long as I take. And in that gap lies, I think, everything that makes a song mine.

A lyric moves at human speed

I write all my lyrics, and writing a lyric that works still costs what it has always cost anyone who writes. There's no shortcut. I can get an idea on the metro in ten seconds and spend days finding its shape. Sometimes it arrives all at once, almost complete; other times I have a chorus I adore and two verses that aren't up to it, and there they stay, waiting, until one day the missing line turns up.

Rewriting is the least visible part of the craft and the one that eats the most hours. Changing a word because it sounds nice but says nothing. Cutting a line that amused me and nobody else. Realising, when I read the lyric out loud, that there's a rhyme so obvious the listener sees it coming before it lands, which in a comic song is practically a crime. A joke needs surprise, and surprise needs work.

Something curious happens with the more explicit lyrics. You'd think they'd be the easiest, because all you have to do is say things plainly. But the hard thing about a filthy lyric isn't being filthy, it's being funny: making the double meaning work both ways and making sure that under the cheek there's something people recognise. Nobody generates that for me. I polish it myself, sometimes for whole nights, until the line says exactly what it means and not one syllable more.

Choosing is writing too

Once the lyric is ready, a different kind of time begins: listening time. The same lyric can give rise to many different versions. The tempo changes, the character changes, the way the words sit on the melody changes. Some versions respect the lyric and betray it at the same time, because they put the stress in the wrong place and a joke turns into a sad sentence. Some versions are beautiful and simply aren't the song. And, every now and then, there's one that is.

Finding it takes something no tool can do for me: listening carefully, again and again. The first time with excitement, the second with suspicion and the third on headphones, at a different hour and in a different frame of mind. There are versions I fall for at night that sound like someone else's work by morning. That's why I've learnt never to decide in the heat of the moment.

As lyricist, composer and producer, my job at this stage looks more like an editor's than an inventor's. I compare, I go back, I tweak the lyric because a version has shown me a line was surplus. That conversation between lyric and music is slow, and it has to be. If it were fast, all my songs would sound like the first option, and the first option is almost never the right one.

Then come the music videos, which run on a clock of their own. Editing a video is one of the most time-devouring activities on earth, and I say that fondly. A cut half a second earlier or later changes the joke. A shot that lasts too long turns a funny scene into an awkward one. There are nights when I sit down to adjust one small detail and, by the time I look up, the sky is already getting light.

The machine has no clock, but I do

There's an idea that has been on my mind for a while, and on this night of the clock change it feels clearer than ever. The tool I work with has no clock. It doesn't get tired, bored or sleepy; it has no idea whether it's four in the afternoon or four in the morning. It couldn't care less that it's the last Sunday in October. If I ask for another version, it gives me another version, without a sigh.

I, on the other hand, carry a clock with me all the time. I have good days and clumsy days. I have hours when I write fluently and hours when I can't even manage a shopping list. I get tired of hearing the same song, and when I'm tired I stop hearing it properly. And precisely because of that, I think, what I make has something to say. The songs come from someone who lives in time, who has exes, nosy neighbours, hangovers, winters and Mondays. The machine provides the sound. The lived time is mine.

I like to think every lyric keeps a trace of the hour it was written. Not literally, of course. But some songs have the impatience of a summer afternoon and others have that slightly dangerous clarity of one in the morning, when everything seems like a good idea. You can't program that. You live it and then you write it.

There's a practical advantage, too, in having a clock that gets tired. Tiredness forces me to stop, and stopping forces me to come back the next day with fresh ears. Many of the best decisions I've made about a song came after sleeping on it, not before. The machine never needs that pause. I do, and that pause is often where the song improves without me lifting a finger.

Two years to release it all

Releasing around 273 songs took me about two years. Put like that, it sounds like a lot and very little at the same time. A lot, because that's a lot of songs. Very little, because whenever I mention it someone does the sums and decides the AI must have done it all in a rush. The truth is that those two years were full of hours that don't show up anywhere: rewritten lyrics, versions listened to until I was sick of them, covers, videos edited and taken apart again, and that small ceremony of deciding that something is finally finished.

It was an intense time, and I wouldn't change it. But it also taught me to look at time differently. I used to think of songs as things that get finished. Now I think of them as things that stay. They're on Spotify and Deezer, and they can also be downloaded for free on my websites, at gayfolkplanet.com and here. That means someone can discover today a song I wrote long ago and hear it as if it were new. For that person, the song has just been born. The listener's time has nothing to do with mine, and to me that's one of the loveliest things about this craft.

That mismatch also reconciles me with slowness. If a song is going to be around in years to come, it deserves the hours I give it now. I'm in no hurry with it. The hurry belongs to the world, which wants everything yesterday. The song can wait for me to find the word.

What I'd do with one more hour

So in the early hours, with my gifted hour in hand, I asked myself the inevitable question. What would I do with sixty more minutes? And I was surprised by how unromantic the answer was. I wouldn't write a new song. At least, not only that.

• I'd go back to one of those half-finished lyrics I keep because they have a good chorus and weak verses, and give it a whole hour without looking at my phone.
• I'd listen again, calmly and on headphones, to a song I think is finished, to check whether it still says what I thought it said.
• I'd go through a music video shot by shot, unhurried, looking for the cut that got away from me.
• I'd read an old lyric out loud to remember how I used to write and what I've learnt since.
• And I'd spend a while doing nothing, which is also part of the job, because the best ideas tend to show up when I stop chasing them.

Almost everything I'd do with that hour is going back. Looking again, listening again, writing again. I think that says a lot about the craft. Generating is moving forward, but writing well is nearly always returning to what you've already done with new eyes. An extra hour doesn't help me produce more. It helps me do it better.

Autumn and the long hours

Now begins the part of the year when it gets dark early. By mid-afternoon you need the lights on, and Madrid turns more inward, with that orange light that never lasts and pavement terraces holding out under patio heaters as if nothing had happened. Lots of people struggle with the change. I welcome it a little, because autumn has always felt to me like a season for writing. The long nights invite you to stay in, take out the notebook and turn a phrase over until it surrenders.

There are nights when I write with the window dark and the feeling that the rest of the world is asleep. They're often the nights that produce the lyrics I care about most. Not because there's anything magical about the night, but because it's the only time when time stops pushing. Nobody expects anything from me at two in the morning. Not even me.

If you've ever wondered how what I do gets done, I leave clues on the blog: pieces about the fear of publishing, about songs that resist, about where I write. This one is about something simpler and harder at once, which is time. The machine saves me a lot of hours, and I'm grateful. But the hours that really count, the ones for thinking, doubting, rewriting and listening, are still mine. And nobody gives those back to me in October. I have to earn them every night.

So today, with the clock turned back and the evening falling earlier than planned, I'm going to do what I always do. Open a half-finished lyric, read it out loud and give it the hour I've been given. If something good comes of it, you'll hear it. And if not, it will have been an hour well wasted, which is the best way to waste an hour.