Sometimes, the most inspiring things I hear come from private or casual conversations. They get stuck in my head in an obsessive loop as I work out their implications.
Here are three. Turns out they are all connected…
1. NO to AI… but be ready.
The first came from a VFX supervisor who spends his time inside the major shops—the vendors that handle the heavy lifting for studio tentpoles. He was relaying what those shops are hearing from their clients:
“They are being told by the major studios… right now, we do not want any AI work in the workflow. But you have to be prepared so that as soon as we say ‘go,’ you can turn on these systems.”
It is not a story about whether AI is good enough. It’s an order to stand by. This is all about errors-and-omissions insurance and chain of title.
E&O coverage is a condition of distribution at Netflix, Amazon, Apple, and Disney+. To write the policy, a carrier needs a chain-of-title review establishing that every element in the film was either created originally or properly licensed. A model trained on scraped material cannot produce that. So most 2026 E&O policies now carry explicit AI exclusions or require a representation that the production contains no AI-generated material that can’t be cleared. No E&O, no distribution. Full stop.
Technical problems resolve gradually and unevenly, which is what creates early-mover advantage. Legal problems just slow down what gets released in the world and where.
In some cases, the cost of readiness is being pushed down onto the vendors.
The VFX shops are being told to build, staff, and maintain a capability they cannot bill against a single shot, for an indefinite period, at their own expense.
That is a mandate they have to pay for themselves in a business on the point of extinction. Remember Rhythm & Hues? They won the visual effects Oscar for Life of Pi eleven days after filing for bankruptcy. That is the historical baseline for how much cushion this business carries.
At the same time, some—maybe all—of the studios are developing or buying their own tech. Netflix. Amazon. Is it a coincidence that they are both tech-first companies? Maybe not. They may well be trying to connect AI to their stock price.
Will the AI floodgates open all at once? I don’t think so. There will be fits and starts. Tech that seems bulletproof will turn into lawsuits. Other technology will quietly work its way into workflows without pomp or circumstance.
For filmmakers? Stay infinitely curious. Whether it becomes an advantage or mere survival, you have to keep up.
But who needs to wait?
There are podcasts made entirely with AI that are monetizing. Surely they don’t have clean copyrights. Ditto on social media platforms.
But here’s the example I’m most curious about—just one, but with much bigger implications. Friends of mine in an unnamed country in Latin America are using AI throughout their pipeline in ways that would not pass muster in the US. They finance independently, release their movies theatrically, and then sell them to the streamers. They have at least three movies on streamers now using AI in questionable ways.
But they are walking through the back door, through post-production delivery departments outside the US that don’t subject the content to the same scrutiny those platforms apply in other territories. I assume Europe is as strict as the US.
So while US filmmakers are learning the tools but can’t implement them, filmmakers in other parts of the world are doing reps and getting paid for it…
2. The Youtube Floodgates are Open but the Deluge has just Begun
The second came from a friend who, among other things, helps studios organize their content strategies. Over lunch, she casually offered:
“There will come a moment when every Hollywood bro, every tech bro, every studio, every production company with any sized library, every AI creator will be uploading their content on YouTube—and we are going to see another huge evolution in the business model.”
I guess we all know this, right? Somehow it hit me harder in that moment.
YouTube took 13.8% of all US television watch time in May—the largest share of any single distributor, ahead of Netflix, and a record. YouTube is not the future of distribution. It’s the present tense of it. The libraries follow the audience. They always have.
So picture the supply side: every studio shelf of unmonetized titles, every production company back catalog, plus a generation of AI creators producing at a marginal cost approaching zero. Producers launching shows direct to YouTube… An enormous amount of content arriving on one surface in a short window.
Here’s what I can’t reconcile, and what makes this interesting rather than obvious.
At the exact moment supply is about to go vertical, the biggest creators on the platform are moving in the other direction. MrBeast announced a 2026 reset built around what he called “ultra grind mode,” saying plainly that recent videos were “not as good as I wanted.”
Not more videos. Better ones.
Colin and Samir have been tracking the broader return to vlogging—the most sophisticated operators in the business going back to the least sophisticated format available.
Are these two connected in some way?
When content becomes infinite, content stops being the asset.
A library is inventory, and inventory in an abundant market prices toward marginal cost—which AI is driving to zero. What cannot be flooded is a specific person being specifically themselves on a Tuesday. The vlog is the least scalable thing on the internet. That is precisely why it’s appreciating.
To me, the bigger implication is handmade content. This is why I would bet on assisted AI to bring prices down before I’d bet on AI Gen. I just wrote about this last week.
But the impact is broader still and we watched this movie before—in music.
Streaming made every recording ever made instantly available. Catalog value went down. The money moved to live performance and the direct artist-fan relationship. The recording became the marketing; the show became the business.
YouTube is running that play on film and television catalog right now. The library becomes the marketing. The open question is what the show is.
If studios start strategically putting their content online as the market floods with content from all sides, they aren’t looking only to monetize what they have. They have to build the next business from it.
Which, to me, all comes back to one simple thing: Does your content have—or can it build—a real community?
All of a sudden, the relevant question in this discussion—whether the content is AI-generated, comes from studio libraries, or is Bro-Hollywood, whatever it is—is whether it can become part of a bigger flywheel.
I don’t know if any of this is surprising. But I can’t stop thinking about it…
3. Could Fede Álvarez Still Break In Today?
The third came from a major producer and former studio executive. She was talking about how good the AI tools have become for building concept trailers, how fast you can now get a look and a tone in front of a buyer.
“Ive never been so inspired by the tools we have today. All of a sudden we can make trailers for movies that no one ever imagined in places no one ever imagined and demonstrate their potential so easily…”
But we both pondered their value in the marketplace.
She talked about the olden days—20 years ago?—when certain directors used rip-o-matics to sell their movies. This was once a moat. They weren’t easy to make. You had to pull scenes and clips from other movies and string them together into a coherent trailer. That, plus a script, sold a movie. But you needed amazing talent behind the scenes to make them compelling.
She worked for one such director, who was hired to make the biggest commercials in the world. He had editors in-house. They used rip-o-matics for the commercial world. He would put them to work on his movie concepts and in their moment, they stood out because they were so hard to pull off well…
Then there was the Short as Calling Card phase. The best reference: In 2009, Álvarez made a four-and-a-half-minute short called Ataque de Pánico!—giant robots destroying Montevideo—for a reported $300 and put it on YouTube.
Within days, he had a widely reported $30 million deal with Sam Raimi’s Ghost House Pictures. He went on to direct Evil Dead, Don’t Breathe, and Alien: Romulus.
The thing that made that short was the to make something like that.
In 2009, a filmmaker in Uruguay producing a VFX driven short taking place in a part of the world where we couldn’t imagine it ever happening for $300 was attention grabbing.
The goal wasn’t just showing you an idea. It was proving a capability. The demo was the credential.
Here’s a piece of the puzzle that calls back to the point I was making about filmmakers in Latin America working with AI tools now: VFX shops in Latin America could prosper only if they bought some or all of their VFX tools on the black market. The licenses were just too expensive. This is kind of a version of the innovator’s dilemma… Those who play by the rules lose? But I digress…
As far as trailers go, anyone can generate a competent concept trailer this afternoon. It is no longer evidence of anything. When the cost of the demo goes to zero, the demo stops functioning as proof.
So the barrier relocated. The question is where.
Not entirely to craft—the machines are absorbing some of that. The bar gets much higher, for sure. A great concept trailer today needs more and more human intervention. The pure Gen AI stuff feels like pure Gen AI. It’s about specificity. It’s about pushing the tools to make something that feels unique—in every frame.
Most importantly, it needs a POV. It needs a real director making real, specific decisions, not leaving those decisions to chance.
In fact, most concept trailers look like trailers from the early aughts. But the best trailers today don’t even try to tell the story.
So part of the puzzle is more human intervention. Where a rip-o-matic was literally melding 30 different directors’ content into a single story that a totally different director would one day make, a concept trailer needs the most originality and precision possible to have a chance.
But that’s just the first piece.
Because mostly, the barrier has moved to validation.
A concept trailer with forty million views and a real argument in the comments is still proof—not proof that you can execute, but proof that somebody cares. The audience becomes the credential.
That’s not really new. First, Fede Álvarez created an amazing short. Then it went viral. Then Sam Raimi “discovered” him.
The old barrier was capability, and capability can be beaten by one talented person working alone. The new barrier is attention, and attention can be beaten only by an audience built over years. The tools that were supposed to democratize the entry point may have quietly raised it.
Where the Three Meet
We’re all talking about the same thing every day, and we don’t even know it. It’s about value. It’s about scarcity. How do we create value in a world of infinite content?
The first story says that technical capability is no longer the only thing determining what gets made. Legal permission and institutional readiness matter. While the studios wait, filmmakers outside the system are doing reps, learning the tools, and getting paid for it.
The second says that when every library, producer, and AI creator pours content onto the same platform, content itself stops being scarce. The value moves somewhere else—to the person, the relationship, the community around the work.
And the third says the same thing at the level of the individual filmmaker. When anyone can make a competent concept trailer, the trailer is no longer the credential. The proof moves from the image to the audience. Can you make people care?
AI doesn’t eliminate scarcity. It relocates it.
For a long time, the ability to create polished images was scarce. Soon it won’t be. But taste is still scarce. Specificity is scarce. A point of view is scarce. An audience that trusts you is scarce. So is a production pipeline that can actually be insured, distributed, and monetized.
That creates a strange split. Production is becoming more democratic at the same time that breaking in may be getting harder. More people can make something. Far fewer can make something that can be cleared, discovered, remembered, and turned into a lasting relationship with an audience.
The cost of creation is falling. The cost of significance may be rising.
Which means “be ready” has to mean more than learning the tools. It means being ready legally, creatively, commercially, and—maybe most importantly—with a community already forming around the work.
The tools were supposed to remove the barrier. Maybe they removed one barrier and revealed the next:
Can you make it?
Can you clear it?
Can anyone find it?
Will anyone care?
And will they come back?
That’s where the three meet.
Very insightful article. I loved how you tied these three themes. They are indeed related. I believe the most challenging questions for all content creators (of any kind) will be "will anyone care? and "will they come back?"
Interesting!