What will the Netflix of the future look like? — Tomorrow Tuesday, Friday edition
Today I’m bringing you something different from the weekly newsletter. Last Friday I was at Campus Party talking about one of the things I spend the most time thinking about and researching: how do we consume content today, and how is that going to change when content becomes infinite?
The talk ended with a demo of a collaborative TV channel. Participants scanned a QR code, proposed what should happen next on the channel, and an AI system turned those ideas into what we were watching on screen, in real time.
You can download the slides here. Below is the explanation and part of what came out of that experiment.
Download charla-campus-party.pdf (9.79 MB, PDF)
The first disruption: distribution
To think about what’s coming, it helps to take a step back. The content industry just went through a huge transformation: getting a story to someone became much, much easier and cheaper.
During the 20th century, if you were a musician, say, putting out an album meant pressing records, putting them in boxes, shipping them to stores and repeating the whole process in every country. Every step was really expensive. Digital turned the entire infrastructure behind distribution from something physical into something virtual: you no longer ship boxes, just bits to a server. Netflix, YouTube, Spotify and social networks built their businesses on that shift.
The distribution disruption brought 6 main effects with it:
- Stagnation. Time spent consuming content isn’t growing anymore.
- Disintermediation. Creators skip the middlemen.
- Fragmentation. Attention gets split across infinite things.
- Deflation. Each hour of content is worth less.
- Globalization. Small markets (Korea, for example) compete with the main one (Hollywood).
- Concentration. A few platforms and a few hits keep almost everything.
Every problem the media industry is living through today is a result of this first disruption.
Second disruption: how we create
What happens when creating content is as fast as consuming it?
The internet made moving a bit from one place to another free. Generative AI is going to make making those bits free.
A good part of the framework for this talk comes from Doug Shapiro, who has spent years thinking about these changes and puts them into words much better than I do. He raises a central question: when something becomes abundant, what stays scarce?
The content industry is in the business of monetizing attention (how much time someone spends consuming a piece of content) and engagement (how much they interact with that content). Attention has plateaued, but engagement doesn’t have to. There’s no limit to how big a Harry Potter fan you can be. That’s the real opportunity behind AI for entertainment.
Source: Activate, Technology & Media Outlook 2025, p. 10.
There’s a trap we fall into every time a new medium shows up: copying the previous one. When the film camera appeared, the first thing we did was point it at a stage play. Now you could watch theater without going to the theater. If you look at movies from the early 1900s, they really were filmed plays. But cinema isn’t filmed theater. It’s a different medium with a language of its own, thanks to the unique qualities of the film camera.
AI today is copying cinema: cheaper movies, faster ads. That’s a small way of looking at it. The real opportunity is creating a completely new medium. Because AI has qualities that neither the pencil nor the camera have:
- It adapts to context. The same story changes format, length and tone depending on where it’s being watched and by whom.
- It can be real time.
- It’s spatial. The gap between generating a video and generating a space you can walk through is closing.
- It’s cheap and fast. More experiments, more voices, niches that didn’t make economic sense before.
- It isn’t limited by physics.
- It’s agentic. Content that can take the initiative.
Why would we want to create a new medium? Because an experience that personally involves each person in the audience can make them a much bigger fan of what we do.
Nobody knows what the format of the future will look like, but we can start to imagine it.
The demo: a live TV channel
Imagine a medium that happens in real time, with nothing pre-recorded. Where what the audience does changes what happens on screen, and it never ends: every time you turn it on, something different is going on. It’s a bit Black Mirror.
This was the demo we did live at Campus Party. Participants wrote what happened next in the story. In the background, an AI director (H3 Max Director, on fal) took the ideas, moderated them and generated video from them, faster than we could watch it. There’s no showrunner, no script and no crew, and what we saw that afternoon will never be seen by anyone else.
The result is weird and far from something anyone would sit down and watch like a series. But it lets us start looking at new ways to create and consume content. There’s a bit of watching TV, a bit of playing a game and a bit of improvising with strangers.
The medium of the future
At SOUTS we believe in creating a new medium, similar to what Matthieu Lorrain calls liquid content: dynamic content that adapts to the audience. This dynamic content has five axes:
- Format. Watching something on a cinema screen isn’t the same as watching it on your phone. The content doesn’t have to be the same either.
- Length. If I have 2 hours I watch the whole movie. If I have 15 minutes on the bus, the content can adapt.
- Staging. Elements of the scene can change for each viewer (say, a character drinking Coke for some viewers and Fanta for others).
- Interactivity. Sometimes it’s passive, sometimes it needs the audience to take the initiative.
- Narrative. The same plot can be adapted for horror or for romance, for example.
This is something every lab training video models is exploring. World models are one of the foundations for it. It isn’t cinema, it isn’t video games. It’s something else, and the ground is fertile for building.
So, what will the Netflix of the future look like?
Nobody has a clue. But my predictions are that it will only be possible thanks to AI, it will have something of real time to it, and it will chase engagement far more than attention. It will be built around communities. But above all, it will be defined by human creators.
Because cinema wasn’t made by Edison or the Lumières. They built the machine. Cinema was made by Chaplin, Lucrecia Martel, Spielberg: the ones who grabbed that machine for things nobody had imagined and invented a language.
Today the machine is being built. What’s missing is the language.
From SOUTS we ask you to think bigger: instead of looking at AI and asking how to cut costs, ask what we can do now that was impossible before.
This was a special edition of Tomorrow Tuesday, the weekly newsletter. If you want to subscribe, subscribe!
At SOUTSwe try to always stay on top of what’s happening, and every time there’s something relevant we’ll bring it to you in this newsletter.
See you next Tomorrow Tuesday.
Yves