13 Ways to Annoy People on The Commute

Inspiration

The idea grew out of years of commuting in London. Quiet tension. Polite frustration. Unspoken eye rolls. I wanted to turn those shared moments into something playful and relatable. I also drew inspiration from creators who elevate ordinary life into sharp, funny storytelling. That mix of humour, everyday truth, and visual craft became the spark.

What it does

The film presents a rapid series of commuting behaviours, each pushed just far enough to feel ridiculous while staying recognisable. The aim is to make people laugh at their own habits, to highlight how similar we all are, and to capture the strange social rules of public transport.

How I built it

I approached the project with my usual filmmaking structure. Shot lists, references, pacing choices, sound design, and clear performance intentions. I combined that with an AI-first pipeline, using Kling, Krea, Freepik, and Luma Labs to create the visuals. From there I shaped the timing, rhythm, and emotional beats in the edit, blending traditional directing instincts with the speed of generative tools.

Challenges I ran into

Consistency was the biggest obstacle. I made this before tools like VEO3 and Clean 2.5 existed, and models were updating constantly. The look of a character could shift overnight. Motion would improve one day and regress the next. Matching faces, lighting, and visual tone across platforms meant re-rendering, re-prompting, and grading shot by shot. I also had to juggle different tools with different capabilities, which made building a stable visual language harder than expected.

Accomplishments that I'm proud of

This is my first fully completed personal short in years, and finishing it matters. I am proud that the film feels cohesive despite the tool shifts. I am proud that it reflects my directing style in its timing, framing, and humour. Most of all, I am proud that it resonates with people and has already become a talking point about what AI filmmaking can achieve.

What I learned

I learned that iteration is everything. I learned that comedy becomes sharper when the behaviour is simple and honest. I learned that AI works best when guided by clear direction instead of treated as a shortcut. I also learned how to stabilise an inconsistent toolset and still deliver a film with a unified look.

What's next for 13 Ways to Annoy People on The Commute

I plan to expand the concept into a small series exploring other everyday micro-frustrations. I also want to share a behind-the-scenes breakdown of the workflow for filmmakers and clients who want to understand what AI can unlock. The hope is that this short becomes a stepping stone to larger projects built with the same speed, creativity, and craft.

Built With

• elevenlabs
• freepik
• higgsfield
• imagen3
• kling-2.0
• luma
• reve
• runway
• veo