I’ve shared a recording of Patterns Beneath the Noise: VFX in the Age of AI, a talk I gave at Mundos Digitales this year. This is a slightly shorter English version, recorded during an internal session at Orca Studios.
There’s a lot happening around generative AI, and keeping up with every new tool could easily become a full-time job. Putting this talk together was a chance to step back a little and think about the patterns behind those changes, based on my experience in film production. Some were already there before this wave of AI, which seems to be accelerating them.
How do we explain what we want? How do we judge whether a result works? And what’s worth spending time learning when the tools themselves keep changing?
The recording is about 30 minutes long, with chapters if you’d like to jump to a particular part. Hope you find it interesting!
The starting point is a question: when execution becomes less of a bottleneck, where does the value of our work move? I look at this through tool-building, with SplatGenerator as an example, and through compositing, where so many technical and creative decisions come together in the final shot.
From there, I’ve grouped the observations into four patterns:
The last part brings these ideas back to production: testing workflows early, allowing for uncertainty, keeping a fallback, and being clear with clients about what’s being used.
It also looks at where our experience helps, and what might be worth learning. My feeling is that integration, judgment and the ability to translate an idea into a finished shot become more valuable. There’s still room for deep specialisation, with a lot to gain from learning a little more about the disciplines around it. Fundamentals like light, colour, camera and composition carry across tools.
One idea I come back to towards the end is the distinction between blue work and red work: deciding what to do, and doing it. I wrote about this in my earlier post on productivity, and I find it just as useful when thinking about these tools.
AI can help us execute more, but that doesn’t automatically improve our judgment. I’m interested in what happens when some of that saved effort goes into thinking more clearly about the image we’re trying to make.
These are observations from my own experience, and I’m curious to see how they hold up as the tools develop. If any of it resonates with what you’re seeing in your own work, I’d love to hear about it in the comments on YouTube.
You can subscribe to my newsletter to get notified about my new tools, blog posts, new courses and other relevant info I’ll post about.
EN 
ES