When people talk about AI filmmaking, the conversation often begins with technology. Which model produces the most realistic results? Which tool is improving the fastest? Which workflow saves the most time? These are useful questions, and the answers will continue changing as AI develops.
The more I worked with AI, however, the more I realised that I kept returning to a different set of questions. Does this shot belong in the story? Is this scene already good enough to move forward? Should this sequence be regenerated, or can it be improved another way? How do we know when a film is actually finished?
Those questions felt much closer to filmmaking than to software. They eventually became the reason this website exists.
The technology keeps changing
One of the most exciting things about AI is how quickly it evolves. New image models appear, video generation improves and workflows that seemed necessary a year ago can suddenly become unnecessary. At the same time, entirely new ways of creating films continue to emerge.
I enjoy watching that progress. If better technology removes unnecessary work, filmmakers benefit. Simpler tools can allow us to spend more time thinking about story, performance and emotion instead of technical process.
What interested me, however, was not keeping up with every new tool. It was understanding which filmmaking ideas remained valuable even as the tools changed.
Feature-film VFX changed the way I see a shot
Before working with AI, I spent many years in feature-film VFX.
A significant part of that work involved solving problems that audiences were never supposed to notice. Computer-generated elements needed to match the photographed world so convincingly that viewers never questioned what they were seeing. Countless small adjustments were made until the work disappeared into the film.
The goal was rarely to draw attention to the VFX themselves. It was to help the audience forget they were watching VFX at all.
Without realising it, that way of thinking became part of how I evaluate every shot.
AI did not change that instinct
When I began working with AI-generated images and video, I found myself approaching them in much the same way.
My first instinct was rarely to ask which model had produced the result. Instead, I found myself looking for moments that interrupted the experience of watching the film. A character might change appearance unexpectedly. A shadow might behave inconsistently. An object might disappear between shots. Sometimes the issue was subtle, but still noticeable enough to remind the audience that they were watching something artificial rather than becoming absorbed in the story.
I did not see those moments as simply how AI is. I saw them as filmmaking problems.
Some could be solved through another generation. Others could be addressed during editing or through the kind of finishing pass that prepares AI-generated material for an audience. Some were small enough that they no longer mattered within the context of the finished sequence.
The important question was never whether the image had been generated by AI. The important question was whether anything in the shot distracted from the story.
Invisible craft has always been part of filmmaking
This way of thinking is not unique to AI.
Editors remove frames that interrupt the rhythm of a scene. Sound designers eliminate unwanted noise that draws attention away from dialogue. Colour grading helps create visual consistency and support the emotional direction of a sequence. VFX artists can spend enormous amounts of time making extraordinary work appear completely natural.
Much of filmmaking is invisible. When the audience notices only the story, the craft has often done its job.
That philosophy felt just as relevant when working with AI as it did in traditional production.
I wanted a different conversation
As I spent more time creating with AI, I realised that I was becoming less interested in individual tools and more interested in the production decisions that followed the first generation.
How do we evaluate what AI has produced? How do we decide whether something is already production-ready? When should we regenerate a shot, and when should we improve what already exists? How do we move from an interesting AI generation to a convincing film?
Those questions seemed to remain relevant regardless of which model happened to be popular at the time. They also happened to be the conversations I enjoyed having most.
This website became a place where I could explore those ideas.
Filmmaking first, AI second
Everything published here is built around this simple belief.
AI is changing how films are made, and I hope it continues making filmmaking more accessible and capable. New models will appear, familiar workflows will evolve and some of today's techniques will eventually become obsolete.
What I hope remains constant is the filmmaking.
Stories still need to engage an audience. Scenes still need to work together. Creative decisions still determine whether a film succeeds. Technology changes how those decisions are carried out, but it does not remove the need to make them.
That is why this website is not centred on a particular model or workflow. It is centred on the filmmaking questions that remain valuable regardless of how the technology evolves.
Closing thought
AI will continue becoming more capable, and I look forward to seeing where that progress leads.
Along the way, I hope this website continues exploring the part of AI filmmaking that interests me most. Not simply how to generate images or video, but how to shape those generations into films that audiences can become immersed in.
Long after today's tools have been replaced, I believe that will still be the conversation worth having.