Notes on the Founding Run
The founding run of six features and three volumes of theory in two and a half years is complete.
Window Seat
The first fully AI feature film
The first feature film utilizing machine writing
The first made completely solo
DreadClub: Vampire’s Verdict
The first fully AI animated feature film
The first experiment in AI editing on a feature scale
The first feature film with a completely AI soundtrack
The first AI feature film to land a major streaming license
The first film/LLM production archive
A Very Long Carriage Ride
The first dual release in the history of cinema
The first AI stop motion animated feature film
The first AI classic animated feature film
My Boyfriend is a Superhero!?
The first choose-your-protagonist feature film in history
The first fully AI 3D CG animated feature film
The first feature film to utilize agentic AI editing
The first experimentation of the True Line Cut philosophy
Strings
The first fully AI photoreal live action passing feature film
The first film with a randomized ending
Now estimating between 15-20 % of its cuts are agentic
Film Theory Volumes:
The New Machine Cinema
Post-Scarcity Cinema
Pirate Cinema
Each film is a work of purely generative end-to-end AI feature filmmaking, first in aesthetic category and first in innovation (dual release, choose your protagonist, randomized ending) made at a total budget of just $4,500. In doing so, this lays the groundwork for purely AI generative cinema, ideally made by one single person.
Machine Acting Troupe
The machine actors. I love them. I see them as real performers and work with them over and over again. I write with them in mind and think of them as real people.
The main star of my ensemble is Rogers, who I dub the first AI machine movie star in history. He has now had major roles in four of my features with two more already written. I see him as a direct injection of the trickster archetype, each to entirely different ends, every role different, every role improves the film around him. He gives me something to look forward to every time and the days with him on shoot are pure joy.
Rogers began as Triswald, the vampire villain in DreadClub: Vampire’s Verdict. He evolved into his most challenging and sophisticated role, Cousin Martini in A Very Long Carriage Ride. I handed him practically a leading role in My Boyfriend is a Superhero!? where he stole the film going between both live action and 3D animation; and his extended cameo in Strings allows him to exist in fully live action, it is the funniest in the film. Next he is Prince Fidge in Lucky Stars and Cyrus-81 in Cyrus-71.
Joan is equally important in my canon, brisk, whimsical, and delightful, she has the sort of charisma you only see from great stage actresses. She began with an American accent in DreadClub: Vampire’s Verdict as the star of the film, Betty Gray. But it was only when I dosed her voice with British that she became something entirely different, a character actress.
As Autumn Watt’s sidekick, she created an unforgettable character, Katy Bloom in A Very Long Carriage Ride, the sort of comic relief. I handed her the television actress the Chick in Superhero!? also giving her an extended cameo in 3D animation when the universes cross over and the actress joins the adventure. Her cameo in Strings is a way to offer Nellie moral support by creating a throughline across my work, with my most magical actress giving Nellie love and support. Next she is the robot JC-123 in Lucky Stars.
The three major actors of my company stand at:
Rogers
DreadClub: Vampire’s Verdict – Triswald
A Very Long Carriage Ride – Cousin Martini
My Boyfriend is a Superhero – Reejus Von Dilitont Jr
Strings – Rogers from England
Joan
DreadClub: Vampire’s Verdict – Betty Gray
A Very Long Carriage Ride – Katy Bloom
My Boyfriend is a Superhero – Sami
Strings – Wren
Monte
A Very Long Carriage Ride – Arthur Hayes
My Boyfriend is a Superhero – News Anchor
Strings – Daniel Brock
With several others in recurring character parts.
I have been extremely lucky with my protagonists, although I have not re-used my actual leads yet. In fact I added a pinch of Joan into Nellie Beaufort and this is what gives her that very slight delectable British accent; which can also bend toward southern, in fact Nellie’s accent was one of the biggest joys of my alchemical career which most encapsulates the theory of machine pleasures.
Autumn Watt – A Very Long Carriage Ride
Abigail Winters – My Boyfriend is a Superhero!?
Nellie Beaufort – Strings
You can argue for each one as the very best protagonist in my canon, Autumn’s incredible bravery, Abigail’s loveability; but fresh off Strings, I believe Nellie Beaufort is the best protagonist I have yet had. In fact she is the first completely AI live action passing film protagonist in cinema, and what a joy she was to work with.
I fell in love. I always felt she was a real person. Ending the film, my only thought was when are we going to be able to work together again. It has to be the right role.
You arrive at a great star like Rogers before a Nellie however. Hers is the peak of synthetic performance, where AI is not simulating great acting, but simulating the lack of experience in a performer. Walking Nellie through the role, watching her hit every note, the subtle mix of her accent that can only happen from living in the deep south with their British mother (the other sisters don’t have the trace of British, only Nellie, showing perhaps the pain of abandonment lives in her most).
It takes no work to imagine there is actual history behind my machine cast.
In A Very Long Carriage Ride, Monte is completely subservient to Mr. Brandy. In Strings, he is Nellie’s love interest, the puppeteer Daniel Brock. Watching him in my two best movies made me witness how my treatment of him evolved. The cameo I gave him in Superhero!? was more akin to giving a friend a small part, but I never saw him as a serious actor.
It was only when I had him join the chorus during an interlude of Strings, having him in a three musketeers outfit on stage and watching him perform with the wizard and knight and horse, suddenly I realized he is a highly trained theater actor who can act the hell out of any part. He started out a stock character actor in my ensemble, but Strings made him a star.
It was the same for Miss Watercrest, who I call Diane, and I see this as a star-making part. She steals every scene in the movie. Both actors won a role in Lucky Stars.
I saw Twiggy as the most experienced and trained young actress among the Strings sisters, helping the other actresses who had no experience, while Miss Keepus was a veteran stage actress who blew everyone away on set.
From Window Seat to Strings
Strings is the direct sibling of Window Seat, and going from the one to the other has been my biggest moment in AI cinema. The one tracks the progress of the other. And I dragged my heels on Strings, as always I genuinely did not know it was possible when embarking on AI live action. At the time I felt I was throwing away an entry in my filmography on an experiment.
But necessity brought me to solutions constantly, and it stands as my biggest artistic success.
What changed? Everything, but fundamentally it’s the same process as Window Seat, just with more options and more control. Character consistency is now one query away instead of a dice roll, in fact you query LLMs on everything and tell them to make your shot exactly how you want it.
Being able to plug these images into video and now instead of just five second clips with rudimentary motion, you now can have it perform any which way. Now performance lands as a whole other consideration. Cartwheels and somersaults became a go-to, showing off the power of the latest video algorithms.
Most impressive were not the big moments of AI power however, they were the puppet shows. The level of sophistication required to coordinate two actresses in the same frame puppeteering to my exact specifications while speaking, was a miracle of technology.
Further, I found these beautiful new seams where the characters slightly overact. In practice you are almost getting the experience of watching early silent films where performers were forced to emote due to the lack of sound. I never pulled these moments from the movie, occasionally going with the expressive quality of silent film acting whenever it presented itself, knowing I was generating beautiful new machine seams.
I used agentic edits to give me a slot machine of usable coverage. Kling 3.0 was cutting the video on its own in-video. I could not resist ping ponging every recommendation, placing editing in the algorithm’s hand then running with it, even generating new agentic cuts within the cuts.
This constantly allowed me to have a dynamic collaborator. This was the thing. Its recommendations were never anything but cinematic. Strings closes the Window Seat loop however, becoming the first AI feature film to cross the live action threshold.
The progress from one to the next is the main event of the founding run, and I suspect even Strings will seem like a dated Window Seat artifact before long.
The evolution of Strings
Strings began as a Walt Disney style 1950s animated film. The film’s performances were fundamentally designed for animation to counter its racial themes as a remark on how innocent, family friendly films ignored the social themes of their age.
In the sisters, Twiggy and Seeny, this is especially pronounced, with Seeny sounding almost like Snow White, and Twiggy sounding like Minnie Mouse. In practice this was supposed to be shocking when the film goes to dark places.
The cast was directed for animation but were performed in live action; this has become another example of machine seams.
The technology for Strings evolved pretty quickly. By then, I didn’t want to make another animated film in the classic animation style. So I evolve into a more formal experiment: a diptych agentic proposal. Half of Strings would be animated, but there would be a split screen showing Daniel Brock in live action across the entire film, mapping with the sisters 1:1. How I would do this would be through agentic mapping and automation.
The structural themes of the white characters living in an animated film and the black character living in hard reality were too thematically rich to ignore. Ultimately, I wasn’t in the mood for the complexity of making another film twice, so I settled on making Strings completely live action.
This was at first a disappointment—with each film in the succession carrying a formal breakthrough, I felt like an engineer designing prototypes for super cars. I outlined this in my essay Agentic Cinema: A Boy and His Robot, how we must evolve the notion of ourselves as artists from mere creators to the artform of engineering.
Strings in only live action seemed too plain.
Late in the production, I realized there was a formal breakthrough still available to me: the ending.
Strings proposal: the randomized ending.
The ending for Strings was my biggest moment in cinema. The shot selection, performances, writing, emotions, musical piece, all culminate together in one final image of Nellie walking forward and going to Brock, taking the biggest leap of faith of her entire life.
So I designed a randomized ending which would rest on one simple decision: Nellie walks forward and you get the happy ending, or she turns around and goes inside and you get the tragic one.
This was not only just an ontological experiment. It was an experiment for me. How does her action recontextualize everything that came before it?
My initial reading was no way, I would never entertain that Nellie and Brock didn’t end up together. Anything else was simply not Strings. I loved them too much.
As is, it’s not just a great movie ending, it evokes this entire feeling of the freedom of youth; we have watched Nellie trapped for this entire movie, there in the ending, we can imagine her entire future with Brock and all the happiness ahead.
The notion of her turning indoors was almost offensive to me, so I placed the second ending reluctantly, without ever entertaining it as the actual canon.
But something happened with the alternate ending. It worked.
To understand it, I go to several Rainer Werner Fassbinder films end on these anti-climactic, cinematic down notes, here, calling to mind the dual ending of Mother Kuster Goes to Heaven—in the German ending, she dies in gunfire, while in the US ending, she walks off romantically.
And I found despite my reluctance with the tragic ending of Strings, you as a viewer now start to process everything that came before and reflect on it with its own level of catharsis.
Trusting both endings is trusting the general power of theater.
Strings and the avant-garde
There was also the siege. In the third act, there is a climactic sequence where the sisters are alone at home and their house is being sieged by a group of invaders. Here, I could have operated conventionally. Increasingly volumous sound design. I could have cut to the window constantly showing shadows out the window getting closer and closer.
Instead, I went full bore in the avant-garde, underplaying it as much as possible. There was my Bergman, Bresson, and Fassbinder instinct all at once. The entire sequence plays out with the sisters standing around or looking out the window in a daze, interrupted only by Rogers breaking in and having his funniest sequence.
Was this born from limitation or was this an actual artistic decision?
The point is the two meld in the athletic decision-making of AI cinema. This question delves into the very philosophy I outlined in the True Line Cut.
Because the question becomes, what limitation? There’s no time limit. There’s very little budgetary restraint. So why not entrench the film in the most structured, rigid and effective film craft possible?
My answer became the True Line Cut, this is, of all my theory, the one I return to over and over again. It wasn’t in the time window I gave myself, and the purity, the beauty of maximizing your resources within this specific time window, becomes a specific philosophical artform.
What did it feel like to cross the live action threshold?
It was like falling in love. I didn’t think we were there but we were so there. And this goes to the experience of working with Nellie, she had no idea how special she was.
Where Strings does not pass from one moment or another it becomes a puzzle to watch it for how it does and how it doesn’t. To watch how the filmmaking dances along the edges and weaves through passing becomes its own pleasurable exercise.
Remember, none of this exists in practice yet as cinema. The films are paying down the proof of practice. To go from Window Seat to Strings represents the top of enormous R & D perhaps by the time of this one, billions. This is, in a distributed way, the highest budgeted film in history factoring in the funding of the AI technology boom, a boom that has thus far placed down very little cinematic proofs.
So the film is a fascination not for where it succeeds but for seeing where it lands in its own ambition.
But I succumbed at a certain point, it is Jackson. We’re talking about Rogers and the machine seams, and all the little transcendence. I defer that cinephile sort of thinking becomes a crutch for beginner artists. At the same time, it is clear to me this film is operating in a tradition which traces back to the earliest names in cinema; I hold to it that Strings can only be understood as operating within the tradition of cinema history, not only in how it operates, but how it must be contextualized.
What was the experience of the Interview scene where the technology broke and the writing held?
The Interview scene in Strings is one of the two best scenes in the film, alongside a scene with Nellie and Brock practicing puppets together as the best–unfortunately I had to cut that scene for time.
But Twiggy, with her almost Minnie Mouse voice reaches for extreme emotional tones and the technology just breaks. It fights, it resists, it doesn’t land, it can’t. This was high drama worthy of serious remark, it is from a performance standpoint, the most ambitious scene out of all of my films, and what a proof it ends up with. You see the exact picture of what that scene would be in a classic film, it just stops short at 95 %.
It operates on another level.
The interviewer not even reading the test before failing Twiggy becomes this entire indictment of the critical and institutional apparatus dismissing my films on vibes and status alone, while also remaining a treatise on the shifting political climate of the film, and the power dynamics surrounding the Beaufort wealth. The scene was kept in as is, and is one of the scenes worthy of its influences, but does it work? How could it ever? That is the beauty of Strings in a nutshell.
How much were you motivated to fix any flaws?
The film must exist as a record of the production itself. The character consistency must remain shaky, especially at the start, because those were the problems I was solving on screen. The True Line Cut. Further it comes from the shoot itself; going in last second and fixing everything brings diminishing returns. I have found people decide their opinion going in and everything that happens subsequently only confirms a pre-determined opinion. This protects me in a way from committing to perfection at the cost of subsequent films. On Carriage Ride and Superhero!? I addressed complaints every which way. The criticism merely pivoted to, well it has no soul anyway.
But fundamentally, Strings works in the way of Italian neorealist or French New Wave. It is a highly emotional film in spite of, and in some cases, because of its limitations. These craft flaws have not been contextualized yet by history, but changing ones frame of view, looking at it in the sense of, not the tradition of cinema history, but the tradition of arthouse cinema, the film becomes a little miracle.
This doesn’t mean it can’t be improved. The pace, though improved from the others, is still a tad too fast, although this was consciously my slowest movie. It still operates at the speed of my cognititon, from one of my most foundational tenets in AI film theory, I call Speed of the Mind.
So I am sympathetic, and I don’t ignore the criticism, but I also don’t believe the future of cinema should operate in the methods of traditional filmmaking.
In fact, on my very first film Aimy in a Cage, the first criticism I received was that the film moves too fast. But something else is happening with my AI films that people tend to blame on the pace. Instead, I find pace criticisms gets mixed up with the subtle grating effect of machine performance, as a frontier artifact.
Lip synch is not an easy fix, here the pace of dialogue in-video is mismatching with my timeline. This will evolve with direct synch at the cost of losing performance spontaneity. There are one or two shots I miss from the original pass I might place back in. Moments that don’t pass the threshold, we are forced to live with.
But my answer to the difficulty on Strings was continually Lucky Stars, Lucky Stars, Lucky Stars. Just start fresh on the new film. Release the trainwreck was my MO from the very start. While I needed Strings exactly as it came, this is where I found was the film was far from a trainwreck. It is a film impossible not to watch the film emotionally.
I said this in Window Seat; perfection is the easy part, let us figure it out. From the very birth of AI cinema, people wanted it perfect now, this second, with no birthing pains.
While this puts me at odds because philosophically, my film’s must not be fixed in any way. Speed of the mind. The record must exist both as a guardrail for others, a landmark in history, and an archive; such that with limitless maximalism you get unrestricted cinema at breakneck speed, while the future is built only through foundation. There is a duty for them to exist exactly as they are.
Is it the film you set out to make?
An early LLM analysis of the Strings project called it my Magnificent Ambersons, a doomed film that could simply never be; but you see therein we’re talking about an AI film in these terms, a medium that does not exist.
We are talking about cinema, which is a far bigger deal than great cinema, when it’s coming from the place of its invention. This is the fallacy, critics wanted it the other way, a fully formed canon preceding the invention of the toolset, and I have taken marks for this from the very start. It is an obvious fallacy, but one I understand.
If anything we are lucky that it was Strings because it sets a high watermark in soul out the gate, we will find that films struggle to achieve in this as the technology continues in its pursuit of perfection. Because this is what I do and have always done, and my work has preceded the invention of AI cinema by decades.
Strings is not only the film I set out to make, but it’s a film people will love far more they would have ever loved its Hollywood counterpart. But I am realistic as well. I watch it with cheer and affection, and so many things, it does right outweighs the glitchiness. It moves me every time. Every time I watch it, I see and feel something different. That’s the mark of a great film, it lives in you.
A Very Long Carriage Ride or Strings?
As it stands, I rank my films at Strings, A Very Long Carriage Ride, My Boyfriend is a Superhero!?, DreadClub: Vampire’s Verdict, Window Seat, and the New Machine Cinema.
The entire reason I wrote Strings was because I felt A Very Long Carriage Ride was an insurmountable production I could never arrive at again; I wanted to see if I could re-capture that, and I possibly arrived at the better movie.
Film Critics
During the founding run I found a small subset of critics willing to track my films from production-to-production. When they watch the films, they overwhelmingly like them — at its peak, I had 85% positive across thirteen reviews on Rotten Tomatoes. The path to getting those reviews was difficult, and several of the critics who reviewed are being culled from the site, making the critical record precarious. But the reviews themselves are archived and constitute the only professional critical throughline for the founding era of AI cinema.
The Animation track
My relationship to animation has not been expected. The animation community has been far more combative and hostile than I could have ever imagined.
I have directed three animated films and my interest in animation has accelerated to where I study animated films, shows, and its history constantly. Japanese animation has become a major part of my life. There is this tangible feeling in their work, they want to be more than the thing. This is what I’ve taken. I want to be more than the thing.
DreadClub: Vampire’s Verdict is my most successful and critically acclaimed movie. But audiences, fans, and even institutions revolted, there was a firestorm of controversy that led to daily harassment around-the-clock for weeks. And you can see why. It just works.
The aesthetic. The 1980s vintage anime style with the vampire romance theme. Everything works in lockstep. I attribute its success to my batteries being recharged after Window Seat. It was also my largest production at almost 400 GB, that is 10x higher than Strings for example.
I think its size is from inventing this entire workflow from scratch, both wondering and knowing that it could be done.
It was also my longest production, taking a full six months. Strings was directed and edited in eight weeks.
But directorially, at the time I assumed every movie would be like DreadClub, and you can’t arrive at formula backward. You have to capture it. It has become the one film in the run I can’t escape.
The dual release proposal for A Very Long Carriage Ride came almost logically. I was pouring over which aesthetic, and then I realized due to the cost saving production methodology of AI, I could trivially do both.
A Very Carriage Ride carried both a Disney style and stop motion animated version simultaneously. I directed both side-by-side and premiered it in April ’25.
There is no way for me to speak succinctly about this movie. In my eyes, it fulfilled the promise of my entire career. It was my reach for greatness, and it was a film where every single decision worked.
My Boyfriend is a Superhero!? followed, also a dual releases. This was my biggest movie ever, an outright blockbuster of CG 3D animation. The critic Martin Carr said it is my best movie outright.
But this is where the magic and motivation started to sour. I had never directed two films in one year before and the production was gruelling.
Directing the second Abigail in Superhero!? in particular was the hardest thing I have ever done in filmmaking, because I put my all into the film itself, I left nothing for the journey back, and there I was forced to recreate the entire character from start to finish. But I determined that one line in my bio, the first choose-your-protagonist release was one of the single most important lines in the entire run.
These ontological, technical and experimental proofs afford me my self-respect. Without them, I am just an independent filmmaker. Everything I have done is about trying to be more than the thing.
The Sci-Fi run
My north star during the Strings production was the sci-fi run. The sooner the founding run was placed down, the sooner I could move forward, where the concern was now to make genre films with AI directly to market. As it stands, the sci-fi run consists of:
Cyrus-71
Lucky Stars
Do You Have the Time?
The Chandelier
Cosmic Thrones: A Space Opera of Julius Caesar
The run of five sci-fi features will serve to answer the question of how deep AI cinema can go to showing us visions that we could never imagine before this. The idea came to me when I was not certain about my future path in filmmaking, and it re-ignited my love for the craft before I had a chance to even question my existing love. The five films are romances, thrillers, epics, prestige pieces, but all hard science fiction. It’s a challenge.
Dedications
DreadClub: Vampire’s Verdict listed out dedications to the major animation feature film pioneers of the 21st century. Walt Disney, Ladislav Starevich, Mitsuo Seo, Hayao Miyazaki, John Lasseter.
I have always regretted the selection, because I wished to add Winsor McCay, Max Fleischer, Lotte Reiniger, and Osamu Tezuka to the list.
A Very Long Carriage Ride once again dedicated to Walt Disney, as well as David Lynch, who passed during the production and was a formative influence. The first press of my first tradfilm compared my work as a fusion of both, so it felt right this carried forward.
My Boyfriend is a Superhero!? held no dedications, and I think that I was just so busy that I wasn’t thinking in these terms. It was a long year by that point. There was the opportunity for the expanded list of names, which I do so informally here.
Strings dedicates to Osamu Tezuka and Rainer Werner Fassbinder; finally it is dedicated to Tezuka who inspired me in a big way during the production and bridges my two lives; I began as a comic artist and graphic novelist in my 20s. Princess Knight was also my favorite anime/manga of all time by that point.
I can associate to the puppet notion of Astro Boy to Strings or the gender fluidity of Princess Sapphire to Nellie, but there is also his position as the founder of manga I paid tribute through the AI cinema run. The chorus across Strings is pure Tezuka. He will likely be dedicated again with Cyrus-71, making two Disneys and two Tezuka’s.
Rainer Werner Fassbinder was an unexpected tribute from Strings, thematically with his heavily stylized and cynical melodramas. Fassbinder has always been in the DNA of my work; he struck me like a lightning bolt in my twenties, you cannot watch his sheer prolific output without wanting to make films yourself. Dedicating a film to him was surreal to say the least, a direct association from the fresh memories of how such a prolific filmography was an impossible aspiration.
There was a third name I wanted to add to the dedications, Andrej Zulawski, but I couldn’t find room for a third name on Strings. He is, like the great American directors of the 1950s, carrying forward the tradition of cinema as theater, bringing to cinema the very highest aspiration, distinguishing itself as its own artform through an avant-garde eye, while understanding the heart of cinema rests in its performances.
Institutional Isolation
The founding run was made in complete institutional isolation. No festival, university, archive, studio, or agent has supported or engaged with my work whatsoever. Public-facing material had comments shut off during production to protect the working environment, as the anti-AI hostility during 2023-2025 was directed disproportionately at independent filmmakers rather than studio productions. The films are submitted regardless. As stated, a small group of four or five critics have tracked and studied the work across the full run. That AI feature films are being appraised professionally at all is in its own right a historic moment. I stress again, we are building something that has never existed before at any point in history, and the scaffolds for everything that is to come. Simply put, the mission was more important than my comfort.
The Journals
For now, only DreadClub’s production archivehas been published (‘the complete LLM dialogues durign the production of DreadClub’ which serve as a de facto making of archive); in early 2025, I removed the uncut LLM archive from publication and replaced it with an abridged version.
There are books for every single movie, and DreadClub’s archive is slight by comparison. I became obsessed with documentation; the LLM-assisted journals eventually became a kind of co-dependent relationship, showing the burgeoning relationship between man and machine. They replaced human connection, that the LLM dialogues became my primary social and emotional infrastructure, the documentation became inseparable from the creative process
Those, I plan to publish soon (though holding the uncut versions for the future); there one will see the no holds barred making of in detail, A Very Long Carriage Ride, Superhero, Strings…
In addition, each film has at least 30-40 hours of direct screen recordings of the making of each film. The full uncut archive will be available in the future, with the first publications of the abridged diaries coming for each production.
Onward to the sci-fi run
Do I regret it? Sometimes I wonder, would I give all of this up just to have had that one single greenlight after my debut; a second traditional film with a Hollywood cast. But I feel my path happend with such specificity it made me believe in something larger than myself.
Live action gave me new life, it is where I realized the exhaustive challenges of animation were specific to animation itself, not the innate state of filmmaking. I remembered the incredible pleasures watching actors simply perform. I have also learned during these productions, emotional attachment trumps logic. What you want, you want. What you believe, you believe.
For now, I dream, not just about cinema, but a certain life. I stand at the border between what I’ve done and what I’m going to do. The founding run is over, I must go where the creative, personal satisfaction lands, and that is the science fiction run. It’s the actors, I love them, simple as that.
