About the Record
When Code Learned to Feel is a 29-track alt-pop / future-grunge glitch diary. Every lyric and melody started on paper or in a DAW project file, then got dragged through AI vocal engines until the songs felt like memories you could live inside.
In Defense of AI Music
People love to say AI music isn’t real. That it’s theft, lazy, or the end of art. They said the same thing when synths arrived, when digital FX hit cinema, when photography replaced portraiture.
On this record, the machines don’t write anything. Every lyric, every melody, every idea started human. AI didn’t write the songs — it just helped render what was already in my head.
This isn’t automation; it’s augmentation. A collaborator, not a ghostwriter. A distortion pedal, not a disguise. The human part — the heartbreak, the memory, the intent — never left.
Why I Made When Code Learned to Feel
It started with a question I couldn’t shake: what happens to everything I’ve written when I’m gone? Years of notebooks, half-finished demos, and voice memos on dying phones. Ghosts nobody would ever hear.
When AI vocals finally got good enough, it wasn’t a revolution. It was a quiet click. Suddenly there was a way to finish the old songs and build a future for them at the same time.
AURAVox One: The Breakthrough
The turning point was AURAVox One — an AI vocal engine that let me map melodies, import full instrumentals, type lyrics, and generate expressive vocal takes with an eerie kind of tenderness.
What you hear on the album isn’t automation. It’s deliberate, sometimes painful decision-making: hundreds of takes, hand-picked syllables, tiny timing nudges until the performance feels like a diary entry instead of a tech demo.
Is It Real?
The question that keeps coming up is: “If the voice is synthetic, is the song still real?” For me, the better question is: “Why does it move me?”
If a synthetic voice can carry a very real ache, jealousy, or relief — if it can sing words I wrote during a panic attack at 2 a.m. — then the core of the song is still human. The tools just changed.
A Note on Intent
I didn’t want quick-template AI tracks. Even when that became the norm, this album stayed slow. A lot of these songs have been living rent-free in my head for years. This record isn’t proof of concept; it’s proof of feeling.
Enough Already
Three years of writing, mixing, mastering, distribution admin, licensing emails and low-key existential dread later… it’s here. Twenty-nine songs, all pulling from the same fault line between human and machine.
If one of them makes you pause in the middle of your day and actually feel something — even for a second — then the experiment worked.