The Next Sopranos Will Be Made by One Person Using AI
TV has been A/B tested into aesthetic oblivion. The next renaissance won’t come from HBO or Netflix. It’ll come from outsiders wielding AI like a camera.
Television is dead. Not in the sense that no one watches it. In the sense that it has no pulse. But something new is emerging at the margins: the AI auteur.
TV has been A/B tested into a state of affective anesthesia: smooth, frictionless, behaviorally optimized content designed not to provoke, but to pacify. The goal is not to enthrall. It’s to retain. To keep you in your chair. To make your thumb twitch the “Next Episode” button before your prefrontal cortex has a chance to reassert itself.
This is not some cynical conspiracy. It’s just the logical result of optimizing for the wrong objective function. Instead of creating unforgettable moments of aesthetic or psychological rupture, studios now optimize for “completion rates,” “time-on-platform,” and “churn reduction.” The result is content that sedates.
You could not greenlight The Wire today. You could not pitch The Sopranos. You might not even get Breaking Bad past the first development meeting. Too slow. Too dark. Too morally ambiguous. Too willing to let the audience sit in silence and tension. These were shows that trusted the viewer. That time is over.
But, for the first time since the invention of the television, it is becoming technologically possible for a tiny team to make an entire serialized show.
Not just a script. A show. Full stop.
Here’s what AI is eroding:
Scriptwriting: LLMs can now draft compelling scenes, generate dialogue in-character, and iterate faster than a writer’s room.
Casting and Voice: ElevenLabs-style models can generate humanlike speech with emotional range.
Storyboarding and Animation: Diffusion models + animation interpolators + soon text-to-video make it plausible to animate full scenes.
Editing and Direction: Generative agents and AI-powered editing tools will soon support complete cinematic workflows.
Scoring and Audio: AI can already compose custom soundtracks and even simulate diegetic soundscapes.
It won’t look perfect at first. But perfection is not the point. Atmosphere, voice, and vision are what matter. And vision does not require a 10-person VFX team anymore.
We are staring at the possibility of a $1K pilot. A one-person operation with an idea, a laptop, and access to the right tools. This is the inverse of what it took to make Game of Thrones. And it’s far more exciting.
Now here’s the missed opportunity: Disney, Netflix, and HBO could dominate this space tomorrow, but they won’t.
The only thing more powerful than a solo creator with AI is a solo creator plugged into a beloved universe. Imagine if Lucasfilm let anyone make fan canon. Imagine if Disney allowed open-source Marvel spinoffs, with performance-based curation. You’d get thousands of experiments for free. You’d discover talent. You’d deepen the IP’s grip on the cultural imagination.
But they won’t do this. They’re too lawyered up. Too scared of brand dilution. Too controlled by middle managers whose job is not to find vision, but to mitigate risk. They will protect their IP with the zeal of a dying bureaucracy, even as the ground underneath them liquefies.
Instead of saying “yes, use our characters, make something beautiful,” they’ll say: “Cease and desist.” And so the next wave won’t happen inside their walls. It will happen outside them.
Prediction:
The next Sopranos will come from someone you’ve never heard of. Pseudonymous. Online-native. Working alone. Maybe on Hugging Face. Maybe on Farcaster. Maybe on something that doesn’t exist yet.
They will build a small but fanatical following.
The pilot will look weird.
The animation will feel uncanny.
But the ideas will be razor-sharp.
And it will grow. And it will spread. And the studios won’t know what hit them until it’s already over.
TV, as we knew it, was a cathedral. Beautiful, expensive, built by teams of artisans.
What’s coming next is a rave in the ruins.
Yes to making a pilot with AI visuals (and voices) as a showcase. I get that. But not the writing. If the writing is done by AI, is will be the same content that sedates. Because AI can’t create anything new, it’s all derivative.