Finn

Finn is a heartfelt story about insecurity, self discovery, and fish porn.

The AI-generated short film was created during the Cinema Synthetica Norway workshop in late October 2024. This innovative project brought together twelve of Norway’s leading filmmakers and creative technologists, divided into four teams, each tasked with producing a short film entirely using AI within a 48-hour timeframe. All films were based on a script by Norwegian writer Erlend Loe and premiered at the Norwegian AI Conference in Oslo on November 20, 2024.

My team included celebrated director Rikke Gregersen and Steinar Klouman Hallert, one of Norway’s most beloved comedy actors and writers.

It’s hard to overstate how early this was. In late 2024, AI video was still a novelty: short, wobbly clips where faces melted, objects changed shape mid-shot and no two generations looked like they belonged in the same film. A consistent visual style across a whole story was more or less unheard of. Getting a coherent felted stop-motion look out of those tools, with a full film delivered in 48 hours, meant training our own model and building the pipeline around it from scratch.

In the creation of Finn, I collaborated closely with the team to develop the visual concept, aiming for a unique felted wool stop-motion aesthetic. I trained a custom LoRA (Low-Rank Adaptation) model to achieve this specific look and set up the technical pipeline to integrate the AI tools effectively. Additionally, I guided the team in utilizing these generative AI tools and played a key role in generating the final images and videos, ensuring the film’s visual coherence and quality.

Cinema Synthetica is an international concept conceived by Todd Terrazas, Max Einhorn, and Mike Gioia, which premiered at AI on the Lot 2024 in Los Angeles. The Norwegian edition was made possible by collaborations between the Norwegian Film Institute, The Norwegian Film School, Det Norske Filmakademiet, DAGSLYS, Reimagine AS, and Oslo Kreativ AI.