Inspiration
Many of us grow up surrounded by rules, expectations, and invisible walls—physical or emotional. Those walls slowly push children away from their own softness and imagination.
But deep inside childhood, there is always a “dream-making pig”— not a toy, but the silent companion we talk to when we don’t yet know how to talk to the world. It reminds us that dreams are not an escape, but another way of understanding life.
This film looks for that forgotten corner— the place where courage, tenderness, and a child’s first light still quietly glow.
What it does
The project uses AI’s visual imagination to rebuild a child’s inner world— the walls, the dinosaur man, the bending buildings, the maze-like city. What seems surreal is actually how a child truly perceives the world.
AI helps translate these emotional perceptions into concrete images, connecting the coldness of reality with the softness of dreams, opening up new ways to tell a story about growing up.
How we built it
We first built the visual tone of the film using Midjourney, DALL·E, and other concept-generation tools. During image-to-video production, tools like Runway, Pika, and Keling were combined to complement each other’s strengths. The dream sequences required multi-step AI iterations to stabilize the surreal, storybook-like atmosphere.
For audio, ElevenLabs provided voice nuance, while Suno generated the musical foundation. All footage was edited, composited, and color-graded in Premiere, After Effects, and Resolve.
Rather than a linear “AI generates, humans fix” pipeline, the workflow became a loop of collaboration: we fed ideas to the AI, and the AI kept expanding the visual possibilities of our story.
Challenges we ran into
The biggest challenge was that a child’s world is governed not by logic, but by emotion. AI struggled with subtle emotional beats—for example:
Walls must feel monstrous without becoming literal monsters.
The dinosaur man must be absurd yet gentle.
Buildings must feel cold but still bend compassionately for the elephant.
These emotional nuances often destabilized AI outputs, requiring repeated iteration and recomposition. Similarly, shaping the pig’s tenderness and presence took substantial refinement.
In the end, we learned to navigate AI’s unpredictability and transform it into emotional precision.
Accomplishments that we're proud of
People resonated with the film’s gentle, childlike perspective. Many shared memories of their own childhood companions—the toys that made them feel less alone. And that made us feel that the story truly reached someone.
What we learned
Every childhood has a quiet corner where a dream-making pig once lived. AI helped us revisit that corner, reminding us that technology cannot replace emotion—but it can illuminate it.
What's next for Walled Dreams
To share the film with more audiences, and continue exploring stories that blend realism with gentle surrealism— stories about walls, dreams, and the small forgotten selves living inside all of us.
Built With
- chatgpt
- dreamina
- elevenlabs
- kling
- midjoury
- photoshop
- runway
- suno
Log in or sign up for Devpost to join the conversation.