Is There a Wrong Way to Make a Movie?
Carl von Cosel attempts to save Elena’s life in a scene from Von Cosel.
AI filmmaking, creative freedom, digital filmmaking, experimental cinema, filmmaking ethics, filmmaking technology, generative AI, horror cinema, Indie Film, low-budget filmmaking, Ronni Thomas, Von Cosel
For twelve years, I’ve tried to make a movie about a man who stole a corpse and slept with it for 7 years. A TRUE story. Already, I realize I have not made things easy for myself.
The film is Von Cosel, based on the true story of Carl Tanzler, the eccentric radiology technician who became obsessed with a young tuberculosis patient named Elena Milagro de Hoyos in 1930s Key West. After she died, Tanzler eventually removed her body from its tomb, brought it home and spent years attempting to preserve it (it’s weirder than that too…).
So: necrophilia. Corpse theft. Obsession. Death. Delusion. The usual ingredients financiers look for!!
For years I researched and <a href="https://www.vice.com/en/article/carl-von-cosel-ronnie-thomas-documentary/” rel=”nofollow noopener” target=”_blank”>wrote and failed at raising funds. I tried, on and off, for more than a decade to figure out how to make the thing. And the problem wasn’t writing it. I knew what the movie was, although at first I thought it was gonna be a documentary using puppets to do recreations (an idea I still love!). The problem was that the movie I could see in my head and the movie I could realistically afford to make lived in two completely different economic systems.
This is not exactly an exotic problem for independent filmmakers.
You compromise. You rewrite. You reduce locations. You ask what can happen in a hallway instead of a ballroom. You learn that your sweeping historical epic has, through the miracle of independent film financing, become three people standing next to a rented houseplant.
I know.
I’m sorry.
But suddenly, this strange little movie that had been sitting around my head for twelve years became possible. Not easy. Not automatic. Not “type a sentence and go have lunch while the computer makes your movie.” But possible. So I made it. Over many, many dark nights behind my laptop screen.
And somewhere along the way I discovered something I hadn’t anticipated: I had apparently chosen a filmmaking technology even more objectionable than the subject of necrophilia. That’s impressive!!
I should make something clear before this turns into one of *those* articles. I am not here to convince anyone that AI is good or bad or evil or omniscient. I’d be happy to be left out of that conversation entirely.
I am also not claiming that festivals are rejecting Von Cosel because it uses AI. Maybe they’re rejecting it because they don’t like it. That remains one of the oldest and most reliable reasons for rejecting a movie.
I’ve been doing this long enough to understand that “they didn’t understand my masterpiece” is generally useful for one’s ego and nothing more.
Maybe the film misses the mark.
Maybe it sucks.
Entirely possible.
What interests me is that, before anyone watches a frame, there is now another question hovering over the film: How was it made?
And increasingly, the answer to that question can determine whether the film is worth viewing at all.
Some festivals have explicitly announced restrictions or outright prohibitions on AI-generated work. Sitges, a festival that would otherwise seem almost genetically engineered for a bizarre film about Carl Tanzler, has been particularly emphatic about its position on AI (a position I suspect may change before this article is even published).
I submitted Von Cosel anyway, accompanied by a cover letter explaining the film’s twelve-year journey and why I’d ultimately chosen to make it this way.
They passed.
Now, I have absolutely no idea whether that decision had anything to do with AI. Maybe they read my letter. Maybe they didn’t. Maybe they watched the film and simply didn’t like it. Film festivals have been rejecting movies since long before a computer could generate a convincing human hand.
But it raises a strange question. When did independent filmmaking become precious about tools? Because my understanding of the job has always been almost embarrassingly simple: **You use whatever you have to make the thing.**
That’sit. That’s independent filmmaking.
I started my career at Troma in 1999. I was there. I watched films get made with no money, impossible schedules, questionable rerey Parker, Matt Stone, and Eli Roth, who simply refused to accept that something couldn’t be done. Nobody asked whether we had the right tools. We used whatever we could get our hands on
And we had Lloyd Kaufman yelling at us.
You borrow a camera. You steal a location. You convince your friends to stand in the rain. You use stock footage. You build something out of cardboard because wood costs too much. You discover that the prop you want costs $4,000, so my buddy Shane Morton spends three hours making one out of foam.
We have traditionally called this ingenuity.
Now, suddenly, there is a category of technological ingenuity that can make the work itself suspect.
And I understand why.
There are enormous and unresolved questions surrounding generative AI: authorship, training data, labor, copyright, consent and what happens to an industry when the cost of producing images approaches zero. Those aren’t imaginary concerns, and filmmakers have every reason to take them seriously.
But artists have also spent the last century enthusiastically adopting technologies that allowed them to make things they previously couldn’t afford to make.
Hip-hop would be almost unimaginable without sampling.
Electronic music would lose enormous pieces of its history without synthesizers, drum machines and sequencers.
A composer who cannot afford the London Symphony Orchestra can nevertheless write orchestral music because somebody sampled one.
Digital nonlinear editing allowed people without access to million-dollar post facilities to cut movies in their bedrooms.
Digital cameras eliminated film stock and processing costs. Visual effects replaced things we once had to physically build.
Every one of these technologies changed labor. Every one changed aesthetics. Every one created arguments about what counted as “real.”
And artists used them anyway. Because artists are scavengers.
Give us a broken camera, a Xerox machine, a sampler, Photoshop, a MiniDV tape, an iPhone, a synthesizer, a piece of software nobody understands yet—we will eventually try to make something with it.
That instinct seems much more fundamental to art than allegiance to any particular production method.
And yet, after spending countless nights making an entire feature film largely on my own, using generative AI to create the world, the characters, and the images that had lived in my head for twelve years, I don’t particularly want to make another movie this way.
That may be the strangest part of this whole experience.
Making Von Cosel was extraordinarily lonely.
Film production is usually communal almost to the point of absurdity. Even when it’s miserable, you’re miserable together. There are arguments and jokes and lunches and disasters. Someone can’t find a cable. Someone else has been holding the cable for twenty minutes. Someone literally disappears (true story). Everyone hates you. Then somebody solves the problem and suddenly everyone loves each other again.
A film crew becomes a temporary little civilization.
This wasn’t that.
There were days making Von Cosel when my closest collaborator was a prompt window.
I occasionally found myself gossiping with AI just to approximate the sensation of having collaborators.
That’s a sentence I wish I had invented for this article. I did not.
There is something profoundly satisfying about being able to reach directly into your imagination and pull out an image that economics had previously made impossible. There is also something profoundly depressing about turning around to share that victory and discovering there’s nobody standing there.
So this isn’t evangelism.
If anything, Von Cosel left me with a greater appreciation for human crews, human actors, human accidents and the strange chemistry that happens when a bunch of people bring their own ideas to something you thought belonged entirely to you.
AI gave me control. It also showed me how overrated total control can be. But it gave me something else, too: It gave me the movie.
After twelve years, Von Cosel exists. That’s the part I keep coming back to.
Independent filmmakers spend enormous portions of their lives being told what they cannot do. Not because the idea is necessarily bad, but because it is too expensive, too weird, too specific, too difficult to market, too period, too niche, too whatever-the-fuck.
Eventually you either stop trying to make those things or you start looking for different tools. I chose the tools.
Now the film, nearly ready to screen leaves the edit with two strikes against it: it’s about a man having an extremely unconventional relationship with a corpse, and it was made using technology that portions of the film world currently regard with approximately the same warmth.
Again: I don’t know why I do this to myself. (My next project is a hybrid theatre/film production about a suicidal cult that invites the audience to set themselves on fire at the climax)
Maybe Von Cosel will find an audience. Maybe it won’t. Maybe five years from now AI- assisted filmmaking will be utterly mundane. Maybe the backlash will intensify. Maybe we’ll discover that most of what we’re currently doing with this technology looks embarrassingly primitive.
Probably all of the above.
But I have difficulty accepting the idea that the artist’s responsibility is to wait for cultural permission to use a tool.
Our responsibility is to interrogate it.
To understand its consequences.
To argue about it.
To establish rules around it where rules are necessary.
And then, sometimes, to pick the damn thing up and see what it can do.
Because twelve years ago I had a movie I couldn’t make. Today I have one I can argue about.
I’ll take the argument.
If you want to learn more about the film or request a screener, email raygun76@gmail.com or visit https://storylab.voncosel.com/
