There is a strange place where creativity lives.
It is somewhere between a feeling you cannot explain and a picture that does not exist yet.
This project started there.
A simple idea:
What if someone wrote a message they were too afraid to send?
Not a dramatic Hollywood confession. Not a grand romantic gesture.
Just a quiet human moment.
The kind where you think about someone who changed your life a little, even if they never knew it.
The first video was called “The message I might never send.”
It was a spoken-word poem from the perspective of someone carrying a feeling they could not release.
The challenge was not making images.
AI can make images.
The challenge was making the images feel like they had a heartbeat.
And, as usual, I discovered that getting a machine to understand a human emotion is a little like teaching a cat to file taxes.
Possible?
Maybe.
Reliable?
Absolutely not.
2/9
The idea behind the companion piece
After finishing the first video, I had a strange thought:
What if the message actually arrived?
What if the person on the other side had been carrying something too?
So I created a second poem:
“The reply I almost never wrote.”
This time, the narration is from the other person’s perspective.
But visually, I wanted something different.
I did not want to simply create another romantic video.
I wanted distance.
I wanted the feeling of two people living separate lives under the same sky.
So I imagined the person receiving the message somewhere far away.
Scotland.
Not postcard Scotland.
Not castles and tourists.
The Scotland of empty beaches, stone cottages, wind, mist, and huge skies.
A place where a person could be alone without being lonely.
Lab note:
This was one of those happy accidents where the location became more than a location.
Scotland became the emotional metaphor.
The landscape was not just where he lived.
It was what he carried.
3/9
Step 1: Writing the emotional architecture
Before opening any AI tools, I needed the emotional structure.
The first video asked:
“Do you feel this too?”
The second video answered:
“You were never alone.”
That sounds simple.
It was not.
The danger was making the response too easy.
A Hollywood ending would have been:
They finally meet.
They embrace.
The music swells.
Everyone cries.
Cue the orchestra.
Roll credits.
But real emotions are usually much quieter.
Sometimes the most powerful thing someone can say is:
“I saw you.”
That became the entire emotional center of the second film.
The final image was not about romance.
It was about recognition.
4/9
Step 2: Building the visual world with Midjourney
The biggest challenge with AI image generation is consistency.
Humans understand that the same person exists from scene to scene.
AI sees every prompt as a fresh conversation with a stranger who has forgotten everything you said five seconds ago.
Which is actually pretty realistic.
I have met people like that.
For the Scotland sequence, I created a consistent character:
A solitary man in his early 60s.
Silver-gray beard.
Thoughtful eyes.
Weathered but kind face.
Rugged Highland clothing.
The goal was not to create a celebrity-looking character.
The goal was to create someone believable.
Someone who looked like he had lived.
Someone who had stories.
Lab note:
AI is surprisingly good at creating beauty.
It is much harder to create history.
A perfect face is easy.
A face that looks like it has loved, lost, worked, wondered, and waited is much harder.
5/9
Step 3: The images became a film
The images were created as cinematic stills.
But a still image is only the beginning.
The magic happens when you make the viewer spend time inside it.
I used PhotoFilmStrip to create the Ken Burns movement.
Slow zooms.
Gentle pans.
Small movements across landscapes.
Nothing flashy.
The camera movement needed to feel like memory.
Not a music video.
Not a slideshow.
A memory.
The opening images slowly discover his world:
A remote cottage.
A letter on a wooden table.
A man looking out toward the Highlands.
The audience is not introduced to him.
They find him.
That difference matters.
6/9
Step 4: The final human moment
The final shot was the one I worried about the most.
After 90 seconds of solitude, I wanted one moment of connection.
I created a six-second AI video clip of him looking toward the camera.
Then I slowed it down to twelve seconds.
That was the right decision.
A quick smile would have felt artificial.
A slow smile feels like recognition.
The emotion is not:
“I got what I wanted.”
The emotion is:
“Someone finally understood.”
Lab note:
This is where AI surprised me.
The tools did not create the emotion.
They did something more interesting.
They created enough space for me to discover the emotion that was already there.
The AI was not the artist.
It was the strange mirror that helped reveal the painting.
7/9
Step 5: The sound of an unanswered question becoming an answer
The music needed to match the emotional journey.
The first video was unresolved.
Suspended.
Like a question hanging in the air.
The second needed to feel like that question had finally reached someone.
Not with a giant orchestral explosion.
Just a subtle release.
A warmer piano.
Softer strings.
A little more light.
The music was created with AI tools as an instrumental cinematic score designed to sit behind narration.
The female voice came from ElevenLabs.
The goal was never to replace human emotion.
The goal was to create a container where the words could live.
8/9
My creative laboratory
The stack for this project:
• Midjourney — cinematic still images and environments
• ElevenLabs — emotional narration voices
• Google Flow Music AI — soundtrack exploration
• Grok imagine – one image-to-video clip
• PhotoFilmStrip — Ken Burns camera movement
• Kdenlive — final editing, timing, and assembly
The actual cost? A few dollars.
Much lower than traditional filmmaking.
But the real cost was attention.
AI makes creating things cheaper.
It does not make creating things easier.
You still have to make hundreds of tiny decisions.
You still have to throw away ideas.
You still have to stare at a screen and ask:
“Is this actually saying something?”
Sometimes the answer is no.
Sometimes the answer surprises you.
9/9
The real lesson
The biggest discovery from this project was not about AI.
It was about human connection.
Technology can generate a face.
It can generate a landscape.
It can generate a voice.
But it cannot decide what matters.
That part is still ours.
The best AI creativity is not pressing a button and waiting for magic.
It is collaboration.
It is improvisation.
It is being a video alchemist directing a very strange orchestra of machines, algorithms, accidents, and unexpected discoveries.
Sometimes the machine makes exactly what you ask for.
Sometimes it makes something better.
And sometimes it makes something that finally helps you understand what you were trying to say all along.
The moral of the story:
The message does not always need to be sent.
Sometimes creating it is how we finally hear it ourselves.
Steve Teare
video alchemist
TerminallyBored.Monster
Palouse, Washington USA