FOR MOST OF FILM HISTORY, making a terrible movie was surprisingly difficult.
You needed money.
Cameras.
Lights.
Actors.
Locations.
Transportation.
Insurance.
Sound.
Post-production.
A person whose entire professional life involved preventing cables from appearing in shots.
This arrangement was unfair.
Plenty of talented people never got access to it.
Plenty of untalented people somehow did.
But the expense created a strange form of friction.
Before somebody could make a two-hour movie, enough people usually had to become convinced that making the two-hour movie was worth doing.
This did not prevent bad movies.
Obviously.
Humanity found a way.
But there was a bottleneck.
AI is removing it.
Which sounds wonderfully democratic until you realize something.
The bottleneck did not disappear.
It moved.
From production to us.
CONGRATULATIONS, YOUR LAPTOP IS A STUDIO
Consider Odysseus: The Fall.
Filmmaker Ash Koosha used generative tools to create a 135-minute reinterpretation of The Odyssey. Koosha says he made the feature in roughly three months of part-time work, largely by himself, with costs in the mid-five figures. The production used tools including Kling, and the character of Odysseus was based on Koosha's own likeness.
A 135-minute mythological epic.
Months.
One primary creator.
A laptop.
This is genuinely extraordinary.
There is no reason to pretend otherwise.
The production machinery required for one person to make something that even resembles a large-scale fantasy movie would have sounded absurdly inaccessible not long ago.
That is the democratization argument in its strongest form.
A person with an idea no longer necessarily needs $100 million and a small temporary civilization to put something cinematic on a screen.
Great.
Now comes the annoying question.
Who is going to watch all of it?
THE CAMERA WAS THE BOTTLENECK. NOW IT IS YOUR TUESDAY NIGHT.
The Astana AI Film Festival recently announced that it received 8,067 submissions from 4,756 creators across 125 countries.
Together, the submitted films contained more than 100,000 minutes of material.
The festival hasn't even happened yet.
The movies already exist.
That number is the story.
Not whether the films are good.
Some almost certainly are.
Not whether AI belongs in filmmaking.
It already is.
The story is what happens when producing audiovisual material becomes so inexpensive that supply loses its old relationship with human attention.
For decades the cultural fantasy was:
Everyone should be able to make a movie.
Wonderful.
Now imagine everyone does.
Then imagine everybody makes three.
The liberation becomes a scheduling problem.
AI may democratize filmmaking faster than humanity can democratize having four extra hours every evening.
THE INTERNET ALREADY RAN THIS EXPERIMENT
We have seen this movie before.
Writing became cheap to publish.
Suddenly there were blogs.
Then websites.
Then content farms.
Photography became effectively free.
Suddenly there were billions of photographs.
Music production became cheaper.
Distribution became nearly free.
Suddenly every platform had more music than anyone could hear in several lifetimes.
None of these things were bad.
Quite the opposite.
Lower barriers created extraordinary culture.
They also created slop.
Because whenever production cost falls toward zero while attention retains value, an obvious business model appears.
Produce more.
Then more.
Then enough that probability starts doing the creative direction.
One thing catches.
The rest disappears.
Generative video is arriving inside an economy that already knows exactly how to exploit abundance.
A FEATURE FILM IS NOT A TECHNOLOGY DEMO
This is the category error haunting AI cinema.
A two-minute demonstration can be fascinating because the technology itself is the subject.
Look what the model can do.
Look at the lighting.
Look at the faces.
Look at the camera move.
That novelty can carry two minutes.
Maybe five.
A feature film asks for something fundamentally different.
It asks a stranger to voluntarily sit still for two hours.
The machine does not get partial credit because rendering the shot was technically impressive.
The audience does not owe gratitude to the pipeline.
Nobody walks out of a bad movie saying:
The compression algorithm was incredible.
A movie stops being a demonstration the moment it asks for your time.
That is when all the boring old human questions return.
Did I care?
Was I surprised?
Did the characters become real?
Did the scene change something?
Why was this shot here?
Why did this story need two hours?
Why did this thing need to exist?
Technology cannot waive those questions.
CHEAP IS A PRODUCTION ACHIEVEMENT, NOT AN AESTHETIC ONE
The marketing around AI filmmaking often emphasizes cost.
And fair enough.
Fountain 0 positioned Odysseus: The Fall partly as a demonstration of what becomes possible without a conventional blockbuster budget. Koosha's production cost was tiny compared with the roughly $250 million reportedly spent on Christopher Nolan's adaptation of the same ancient story.
That comparison is irresistible.
$250 million.
Versus five figures.
Look what technology did.
But the comparison also reveals a strange new cultural habit.
We have started treating how cheaply art was produced as part of the art's case for deserving attention.
Imagine doing this elsewhere.
This restaurant made your dinner for fourteen cents.
This violin cost less than a sandwich.
This novel was written in nine minutes.
This building required almost no architects.
Neat.
How is it?
Cheapness is incredibly important if you are financing the object.
The audience experiences something else.
The audience experiences the object.
Production cost determines whether something can exist. It does not determine whether anyone should care that it does.
THE ACTOR WHO NEVER NEEDS A TRAILER
Then there is Tilly Norwood.
Norwood is an AI-generated "actor" developed by Xicoia, part of Particle 6 Group. She is being positioned as the protagonist of an upcoming feature called Misaligned. The company has even made the digital character available for press interviews, creating the strange spectacle of journalists interviewing a promotional system about a movie involving its own synthetic persona.
This is where AI cinema becomes more interesting than the usual argument about replacing actors.
Because the real dream is larger.
The infinitely available performer.
No scheduling.
No trailer.
No aging unless desired.
No conflicting movie shoot.
No difficult interview.
No contract negotiation in the human sense.
No scandal unless someone writes one.
The actor becomes intellectual property with cheekbones.
From a production standpoint, this is magnificent.
From an artistic standpoint, it raises an awkward question.
Why did actors become interesting in the first place?
Was it merely because somebody needed a face in the rectangle?
Or was part of acting the strange fact that another consciousness was interpreting the material?
A machine can simulate hesitation.
An actor hesitates.
Those may eventually look identical.
They are not automatically the same cultural event.
AND YET HUMANS KEEP SNEAKING BACK INTO THE MACHINE
This is the funniest part.
Even AI films frequently turn out to contain humans everywhere.
Neill Blomkamp's Nightborne, a 13-minute AI filmmaking test created with Seedance 2.0, used human concept artists and the likenesses and recorded voices of 32 real people under agreements, according to Blomkamp's own description of the project.
Odysseus: The Fall uses human likenesses.
Its script is human-directed.
The editing decisions are human.
The prompts are human.
The selection is human.
Tilly Norwood has creators.
Every synthetic world seems to require people standing just outside the frame making choices about what the machine should produce.
Which makes the language of "fully AI-generated" strangely slippery.
AI does not remove authorship.
It redistributes authorship while making some of the people involved less visible.
That may eventually become one of the biggest labor fights in creative industries.
Not whether humans exist in the pipeline.
Whether the pipeline still considers them worth mentioning.
THE FESTIVAL OF EVERYTHING
Eight thousand films.
One festival.
This is where the future starts looking less like Hollywood and more like YouTube.
Film used to have an unusual cultural property.
Movies were discrete objects.
Someone made one.
It arrived.
People discussed it.
The production scale made movies comparatively rare.
Generative systems threaten to turn film into something closer to feed content.
Not:
Here is the movie.
But:
Here is another movie.
And another.
And another.
Customized for you.
Regenerated.
Alternate ending.
Different actor.
Shorter version.
Personalized version.
Anime version.
Version where you are Odysseus.
Version where Odysseus is a Labrador.
The technological endpoint is not simply cheaper Hollywood.
It may be the collapse of the movie into an infinitely mutable content object.
At that point the word film begins doing some very heavy lifting.
SCARCITY WAS NOT ALWAYS THE ENEMY
This sounds dangerous because scarcity has excluded people.
It has.
Film history is filled with people who deserved cameras and never got them.
The answer cannot be restoring a system where a small group of studios decides which human beings are allowed to make culture.
But abundance has its own pathology.
When everything can be made, making becomes less meaningful as a signal.
The important questions move elsewhere.
Who chose this?
Why this version?
Why now?
What survived editing?
What was rejected?
What took time?
What mattered enough that somebody kept working when the novelty disappeared?
Art has always involved possibility.
It has also involved refusal.
You could have written another scene.
You didn't.
You could have used another take.
You chose this one.
You could have painted forever.
Eventually you stopped.
Generative systems are extraordinarily good at possibility.
Infinite possibility.
Perhaps the new artistic skill becomes refusing most of it.
AI FILM WILL EVENTUALLY BE GOOD
Some of it probably already is.
That is another reason "look how ugly this is" will age badly.
The visual defects will improve.
The coherence will improve.
Long scenes will stabilize.
Characters will remain consistent.
Voices will gain texture.
Synthetic actors will become extremely convincing.
Someone will eventually make an extraordinary movie using generative AI.
This should not be difficult to imagine.
Humans have made good art with every ridiculous tool we have ever touched.
The real slop question was never:
Was AI involved?
The question is what the technology encourages once wrapped in an economic system.
More care?
More possibility?
More strange voices getting access to cinema?
Wonderful.
Or does it encourage the same thing every cheap content technology eventually encounters?
Volume.
Speed.
Imitation.
Trend-chasing.
Infinite supply.
The movie equivalent of the SEO article.
That is where AI cinema either becomes a new artistic medium or the largest direct-to-video rack in human history.
THE AUDIENCE IS THE LAST SCARCE RESOURCE
This is the part nobody can automate away.
Your attention.
Your evening.
Your boredom.
Your capacity to care.
The machine can generate ten films while you sleep.
You still wake up capable of watching roughly the same number of movies you could watch yesterday.
This creates the central contradiction of infinite media.
Supply becomes unlimited.
Demand remains attached to a human nervous system.
A nervous system that wants dinner.
Sleep.
Conversation.
A walk.
Maybe a book.
Maybe absolutely nothing.
The computer does not experience this problem.
The computer can produce while nobody watches.
And it probably will.
That may be the purest image of future slop.
Millions of films generated.
Thousands of synthetic actors performing.
Infinite stories unfolding.
No audience necessary.
Servers humming in the dark.
The great entertainment machine finally achieving total productivity.
And somewhere outside the data center, a person closes the laptop because they have work tomorrow.
