I thought love ended at goodbye. Then I taught AI to sing about it.
Some songs make you want to tap your feet. Others make you stare at the ceiling and wonder why somebody put those particular words in that particular order.
This project started with “You Are Not Alone,” a song about loneliness, distance, and the promise that somebody is still there. I wanted to take the emotional seed of that song and grow a new story from it.
Not a copy of the original. A different story. A man loses his wife, loses his sense of direction, and eventually discovers that love hasn’t disappeared. It has changed where it lives.
That sounded simple enough.
Naturally, I invited artificial intelligence into the workshop. What could possibly go wrong?
Quite a bit, as it turns out.
2/9
THE STORY CAME FIRST
I began by turning the lyric into a short story.
Daniel has been without his wife, Mara, for eleven months. The house is full of the little things that make grief so persistent: a photograph, a familiar sweater, an empty chair. He can’t figure out how to live in a world where the person who made it feel like home is no longer there.
Then he dreams of Mara in a garden.
She tells him he doesn’t have to be alone.
I could have ended the story there, with a comforting dream and a man waking up to a slightly less miserable morning. But I wanted something more tangible than a mysterious voice in the night.
So I gave Daniel a grieving teenager named Ben.
Daniel can’t fix Ben’s grief. He can’t even fix his own. But he can sit beside the boy, listen, and tell the truth about what it feels like to lose somebody.
That’s where the story turns.
Daniel discovers that he can carry Mara’s love forward by offering another person the kind of presence he himself desperately needs.
The roses will bloom again. Not because grief is over, but because life continues.
Lab note: I like stories in which the main character doesn’t suddenly become wise, brave, or emotionally well-adjusted. Sometimes the breakthrough is simply realizing that you can help somebody while you’re still a mess yourself.
3/9
THEN I HAD TO MAKE A SONG OUT OF IT
A story gives you room to wander. A song gives you a limited number of minutes before the listener starts checking the refrigerator.
The first lyric was much too long for a two-minute song. It had too many lines, too many explanations, and too much narrative detail. I was trying to squeeze an entire short story into a musical container.
The solution wasn’t to sing faster. It was to decide what the song was actually about.
The answer came down to one idea: love doesn’t disappear when someone dies. It finds another way to live.
I wanted a line that could carry the whole project. Something simple enough to remember after the music stopped.
That became:
Love doesn’t leave, it learns to live on.
There’s the hook.
It has a little internal tension, too. Love doesn’t leave, but it does change. The person is gone. The relationship isn’t erased. What remains can become part of how we treat other people, how we endure our own losses, and how we keep moving.
I built the chorus around that line, then cut the verses down to the essential images: the empty rooms, the grieving boy, and the roses waiting for spring.
The bridge needed only two lines:
I’ll plant the roses, watch them grow
There’s still so much that love can show
Sometimes the hardest part of writing is knowing when to stop explaining.
4/9
CHOOSING THE SOUND
The lyric needed a musical identity.
I didn’t hear a dance track, a cheerful pop song, or a delicate little piano piece that sounds as though it’s afraid of waking the neighbors.
I heard a mature male voice.
The singer needed to sound as though he had actually lived through something. Not theatrical misery, not a polished crooner pretending to have a broken heart, but a warm, expressive baritone with enough strength to carry the chorus.
I settled on cinematic adult contemporary pop, with a soulful ballad arrangement and a touch of gospel influence.
The plan was straightforward:
Piano-led verses. Warm orchestral strings. Restrained live drums. A gradual build toward a powerful chorus. A vocal that begins close and personal, then opens up as the emotional meaning becomes clear.
The final chorus should feel earned.
I didn’t want the singer to start at maximum intensity and stay there for the entire song. That’s the musical equivalent of typing everything in capital letters. After a while, even the important bits stop sounding important.
Lab note: Dynamics tell the story, too. A quiet verse gives the chorus somewhere to go. Restraint can make power more convincing.
5/9
DIRECTING THE AI ORCHESTRA
This is where the video alchemist puts on the hard hat.
AI music generation isn’t quite like hiring a studio musician, handing over a score, and getting precisely what you imagined. You describe the sound, cross your fingers, and see what the machine thinks you meant.
For this project, the music prompt was:
Cinematic adult contemporary pop ballad, soulful male baritone, rich expressive vocals, intimate piano-led verses, warm orchestral strings, subtle electric guitar, restrained live drums, gradual dynamic build, powerful soaring chorus, memorable melodic hook, gospel-inspired harmonic lift, emotional vulnerability, deep resonant delivery, grief transforming into hope, dramatic crescendo, polished cinematic production, heartfelt and uplifting, no dance beat, no toe-tapping rhythm, no airy or fragile vocals.
That’s a fairly specific set of marching orders.
But a prompt isn’t a guarantee. The AI may understand “powerful” as loud, “emotional” as melodramatic, or “cinematic” as an invitation to summon every orchestra member who has ever owned a tuxedo.
The job is to listen critically, not merely accept the first output because a computer made it.
The technology supplies possibilities. I decide which possibilities serve the story.
6/9
THE VIDEO STILL HAS TO EARN ITS KEEP
A song can work perfectly well with a static image. But I wanted this project to have a visual life of its own.
The story offers several images that could carry that life: an empty room, a photograph, two people sitting together in silence, a garden waiting for spring, roses beginning to bloom.
The trick is to avoid illustrating every lyric literally.
If the singer mentions a photograph, I don’t necessarily need to show a photograph. If he sings about loneliness, I don’t need a sad man staring through a rain-streaked window for the four hundredth time in cinematic history.
The visuals should carry the emotional movement rather than act as a series of pictorial subtitles.
The beginning belongs to absence and stillness. As Daniel reconnects with the living, the world can gradually open up. The garden becomes more than a garden. It becomes a way of seeing what remains possible.
And the ending should resist the temptation to wrap everything in a neat little bow.
Mara is still gone.
Daniel still misses her.
The difference is that grief no longer gets to tell him the whole story.
Lab note: When the emotional idea is strong enough, the visuals don’t need to shout. Sometimes a single image that leaves room for the viewer’s own memories does more work than an elaborate sequence of effects.
7/9
THE HUMAN STILL HAS TO BE IN THE LOOP
One of the odd things about making art with AI is that the machine can produce an astonishing amount of material without necessarily understanding what matters about any of it.
It can generate lyrics. It can generate music. It can generate images and moving pictures.
But abundance isn’t the same thing as direction.
The creative work happens in the decisions: which lines survive, which sounds belong, which images carry the right emotional weight, and which apparently impressive results are actually taking the project in the wrong direction.
I had to shorten the lyric because the story was crowding out the song. I had to specify a male baritone because the emotional character mattered. I had to describe the arrangement in enough detail to establish the intended musical world.
Each decision narrowed the possibilities.
That’s not a failure of AI creativity. It’s the nature of collaboration with a tool that can produce more options than any one person could reasonably use.
My role isn’t to pretend I can control every note and pixel. It’s to recognize what works, discard what doesn’t, and keep the project pointed toward its purpose.
I’m the video alchemist. The AI orchestra can make plenty of noise. Somebody still has to decide what the music means.
8/9
THE REAL LESSON
The line I keep coming back to is the hook:
Love doesn’t leave, it learns to live on.
It applies to the story, of course. Daniel discovers that Mara’s influence hasn’t ended with her death. Her kindness continues through the people she helped, and Daniel becomes part of that continuation.
But the line also says something about grief itself.
We sometimes imagine that healing means reaching a point where the loss no longer hurts. That may happen to some degree, but it isn’t the only measure of healing. We can miss somebody terribly and still laugh. We can carry sorrow and still help another person. We can accept that something precious has ended without declaring that everything it gave us has ended, too.
The roses don’t bring Mara back.
They don’t have to.
They grow because Daniel is still here, and because he has decided to participate in the life that remains.
That’s the emotional destination I wanted the song to reach. Not the denial of loss, but the discovery that love can outlive its original circumstances.
And if an AI-assisted song can help somebody feel that truth for three minutes, then all the prompting, editing, second-guessing, and occasional creative wrestling has served a purpose.
9/9
A FINAL THOUGHT
I started with a song about being alone and ended up with a story about what happens when two grieving people sit together long enough to discover that neither has to solve the other’s pain.
There is something quietly hopeful about that.
We don’t always need the perfect words. We don’t always need an answer. Sometimes we need someone willing to stay beside us while we figure out how to live with what cannot be changed.
And sometimes, when we’re the ones who have been hurting, we discover that we can still offer that same gift to somebody else.
The AI helped me explore the story, compress it into a lyric, and define a musical direction. But the reason for making it was human from the beginning.
Love doesn’t leave. It learns to live on.
That’s the song I wanted to make.
Now I just have to make the video live up to it. No pressure, then.
Steve Teare
video alchemist
TerminallyBored.Monster
Palouse, Washington USA
PSYCHOLOGY
“Love Learns to Live On” is a psychologically precise portrait of adaptive grief rather than a sentimental denial of it.
The film (and the story behind it) follows Daniel eleven months after his wife Mara’s death. The house is frozen: empty rooms, her photograph, the faded scent of her perfume, curtains left unchanged. These are classic sensory and spatial anchors of the lost attachment figure. Time has not moved for him; the internal world still expects her presence. The opening establishes what contemporary grief research (continuing-bonds theory, Klass et al.) treats as normal: the relationship does not simply terminate. It persists as an active internal representation.
The dream and the internalized voice
The dream in which Mara holds his hand and whispers that love will find the light functions as a classic compensatory or transitional experience. Early grief often produces precisely these quasi-sensory or dream encounters. Psychologically they are not pure pathology; they are the psyche’s attempt to maintain the secure base that has been removed. Here the dream supplies the core insight the rest of the film enacts: the bond can change form without disappearing.
The later insertion of material drawn from “You Are Not Alone” (“Another day has gone… You are not alone… I am here with you”) externalizes that same internal companion. What begins as raw isolation (“I’m still all alone… Did you have to go and leave my world so cold?”) is answered by a reassuring voice that is simultaneously the deceased, the self, and a more general promise of presence-in-absence. The shift is not magical recovery; it is the moment the continuing bond becomes usable rather than only painful.
Sitting with Ben
The decisive turn is the appearance of the grieving boy. Daniel cannot fix Ben’s pain any more than he can fix his own. What he can do is sit beside him and share the tears both have been hiding. This is the film’s most psychologically mature move. It abandons the rescue fantasy (the wish to “mend” the other or oneself) and substitutes mutual regulation and witness.
In attachment and object-relations terms, Daniel is externalizing the continuing bond. By offering Ben the kind of non-fixing presence he himself desperately needs, he keeps Mara’s love alive in a new relational form. Love is no longer only a private, frozen memory; it becomes generative. The blog post that accompanies the film states this explicitly: “Daniel discovers that he can carry Mara’s love forward by offering another person the kind of presence he himself desperately needs.” That is continuing bonds in action—transformed, not relinquished.
The roses and the refusal of neat closure
“Plant the roses. Watch them grow” is the restoration-oriented counterweight (Dual Process Model of Stroebe & Schut). Grief oscillates between loss-focused immersion and engagement with the life that continues. The roses do not bring Mara back. They mark Daniel’s decision to participate in what remains. The film’s deliberate refusal of a tidy ending—Mara is still gone, Daniel still misses her, grief no longer gets to tell the whole story—aligns with the modern clinical view that healing is not the disappearance of pain but the integration of the loss into an ongoing life.
Creative process as meaning-making
The making of the piece itself is part of the psychological work. Starting from a song about loneliness, growing a short story of transformed love, then compressing it into lyric and image is a classic act of narrative reconstruction (Neimeyer). The constraints of song form forced the abandonment of explanatory excess in favor of the single usable line: “Love doesn’t leave, it learns to live on.” That line is both aesthetic hook and psychological summary: the relationship survives, changed in form, and finds new places to live.
In short, the video is not about “getting over” loss. It is about discovering that the love can be relocated—into memory, into presence with another grieving person, into the decision to plant something that will outlast the immediate season of pain. That is a sophisticated, non-pathologizing picture of how continuing bonds actually function.