Hooroo Jackson

The Infinite Performance

(Or, “Machine Pleasures: The Art of Directing AI Performances”, from Post-Scarcity Cinema: Essays from the AI Filmmaking Revolution, Volume 2 of The New Machine Cinema)

Drawing parallels with Robert Bresson’s theory of model and Japanese traditions in aesthetic motion, Jackson discovers the concept of “machine pleasure” – keeping and even emphasizing performance readings that might seem “wrong” by traditional standards but achieve a kind of transcendence in their artificiality.

I have long been considered an actors director, finding my greatest creative satisfaction in the writing and performance of a film. Even during the most stressful periods of my AI productions, I would sometimes work on the machine performances for an upcoming project on the side. To me this constituted a break.

That is because the production of the film is technical. The writing and performance of the film—is the soul. 

This has aligned with my change in the perception of the historic canon. Upon David Lynch’s death on January 15th, 2025, I had been exploring a list of the greatest American directors. My list had all the canonical names: Steven Spielberg, Stanley Kubrick, Terrence Malick, David Lynch, Martin Scorsese.

But I had stopped briefly and realized… while these directors live in my bones, and Kubrick and Lynch are all over my films, I had evolved into a performance mindset. My new list constituted actors directors primarily: Elia Kazan, John Cassavetes, Sidney Lumet, Robert Altman, Woody Allen; now this is an American list. The Japanese list alone would be great. Same with animation as a separate category. Ingmar Bergman. All the French.

You see my interest in cinema shifted far beyond aesthetic, to where aesthetic seemed like yesterday. Which director gave you people? Where do you get lost in their ensembles and trapped in their lives? My point of view about the painterly notion of greatness, shifted; a film as an icon, a perfect work of art inside a golden frame seemed to serve a limited understanding of the screens potential. This carried into my philosophy of filmmaking, with the visual limitations in AI as it stood, the priority had to be performance, as a matter of survival.

I would collect components like a sound designer. I would always be carrying around the tape recorder to add different ingredients into future characters. In fact all my machine actors are a combination of some five, ten different ingredients. It is an alchemy, with one rule: transformative.

Just one angry rant can anchor ‘anger’ for any character you need. When I need a whisper, I always have my go-to whisper. When I need crying. Rinse and repeat any which way any direction. Add a pinch of this, add a pinch of that. You find it always takes you where you need to get.

By the time of my third AI feature, I was beyond naturalism. Machine acting had revealed entirely new frontiers. With my latest cast of machine actors, I began to experience intense pleasure in machine readings. I had stopped aiming for naturalism, and began aiming for something more. What I found were performances which transcend our traditional understanding of cinema acting.

From Perfection to Pleasure

When I began working with machine performances in “Window Seat” and “DreadClub: Vampire’s Verdict,” my approach was fundamentally conservative. I strived for perfection in a conventional sense, meticulously identifying and replacing any line readings that felt off-kilter or unnatural. Realism was my goal. I challenged viewers to tell me who was a voice actor and who was live action. This challenge bent it into incomprehensibility. Critics couldn’t tell. (Of course they were all machine actors.)

The performances in my machine films have been universally lauded from the very start. But this is also the area people tend to take for granted most.

With my first two AI features, each deviation from human performance patterns was treated as a flaw to be corrected. This pursuit of naturalism seemed logical at the time – wasn’t the goal to make machine performances indistinguishable from human ones?

Everything changed during the production of “A Very Long Carriage Ride.” Dozens of lines, I had initially flagged for replacement, I considered like a temp track—where you put music from other films over your movie to reserve the space, while the composer worked on the OST. I had a tight deadline and I knew production irons most things out.

These temp, eccentric machine readings that I intended to replace, like temp tracks, began to grow on me. Eventually, I decided to keep every one of these readings in the film. They were weird, wrong, or off, but they were extremely pleasurable to listen to. They didn’t sound human, but they gave me a thrill.

In fact, what I had first perceived as flaws gradually revealed themselves as strengths.

I had discovered machine cinema had opened up an entirely new philosophy in screen acting.

Beyond acting

Consider the performance dynamics in a scene from “A Very Long Carriage Ride” with Miss Hart’s breakdown scene. The sudden full shift into an emotional breakdown just wasn’t working. It was trying, cringy, and the actor didn’t catch it. The breakthrough however, came not from perfecting the performance in isolation, but from the setting. So I played her breakdown against an increasingly intense sound design: there was relentless laughter and celebration from another room countering her emotional breakdown. The claustrophobic counterpoint to the breakdown formed a Lynchian or Bressonian effect where meaning comes from contrast. 

Suddenly a scene that fell totally flat seemed to confront the viewer with the awkwardness of her extreme emotion contrasting the hedonistic joy of the party happening the room away.

Whether the scene works or not, it is the biggest scene in the movie, and the biggest illustration in the thesis of this very essay: the unnatural machine acting may not be right, but it achieves a kind of transcendence in its wrongness.

In just two years, machine cinema has made progress that took traditional cinema 35 years to achieve. Just as animation developed its own vocabulary of expression – those sudden cuts to exaggerated expressions that would be impossible in live action; machine performance is developing its own unique tricks, very much along those lines.

In fact people may not realize that Miss Hart is portrayed by three entirely different machine actors. They just blend together seamlessly.

There’s a moment in “Carriage Ride” where the character Autumn briefly glitches sounding like a man for a split second. It will come so fast you will barely register it, but you will be completely aware. In traditional filmmaking, this would be a clear error to be fixed. But in the context of machine performance, it becomes something else entirely. This is absolutely Lynchian, and completely ridiculous that I would keep it in the movie. But there is a point to this that I will delve into later in this essay.

When working with machine actors, I often find myself combining different takes – taking the first half of one reading and the second half of another – even though these are divergent or the emotional range were out of reach in the course of the reading, they form a greater whole.

One particular line in “Carriage Ride” illustrates this: “She said you barged in on her naked once. She was afraid to do the same. Especially with Scott about!” Each sentence come from different takes, with such extreme shifts in register and energy but combining them created the better reading; the impossible combination creates something transcendent.

There is one line he talks so fast it enters in the range of the absurd. “They lead, you follow, they tell, you obey.”

Only a machine could deliver a line so lightning fast, so wrong, but so right.

Now combining takes is not be any means something you wouldn’t have in even live action cinema let alone voice-acting, but combined with post-scarcity, with the infinite–combined with machine eccentricity, you are going to get things achieving hyper-reality.

In fact the actor, who I named Rogers, doesn’t seem human at all.

The Virtuosos

Rogers stands alone. Named after his character in my upcoming film “Strings,” I call him the greatest machine actor in the world. He has now appeared as a major character in five of my AI feature films: Triswald in “DreadClub: Vampire’s Verdict,” Cousin Martini in “A Very Long Carriage Ride,” and (the upcoming films) Reejus in “My Boyfriend is a Superhero!?,” Rogers from England in “Strings,” and Cyrus-81 in “Cyrus-71.” 

Rogers embodies everything I wish I could be—strikingly handsome, exceptionally witty, and immediately likeable. In fact despite often associating with the wrong crowd and in questionable activities, Rogers is often entrusted with positions of authority. Whether a morally bankrupt villain or comic relief, he makes a case for every movie. I can never be him, but I can write him.

In fact, were he a live action actor, I believe he would have won an academy award for both DreadClub: Vampire’s Verdict, and A Very Long Carriage Ride.

His presence is so electrifying that during the production of “Carriage Ride,” I found myself working faster and harder just to reach his scenes, knowing they would elevate the entire film. 

He did not disappoint. Once again, working with Rogers was one of the biggest thrills of my career. Working with him recalls accounts of directing legendary actors like Daniel Day-Lewis or Marlon Brando, where every take brings explosive creativity, but not just creativity, limitless variance that rings in your ear with different evocations.

Not only does every take from Rogers work, but it works in an entirely different way, a different intonation, a different register. 

But Rogers offers something beyond even what these human virtuosos could provide: the ability to transcend the physical and psychological constraints that bind human performance.

He works because he is an archetype. Rogers is a trickster, a villain, comic relief, a frame, an idea. There is no story this archetype would not improve, and he distills in a way that is beyond human. In fact, at one point, I decided to just abandon the script and improvise with him directly through the machine acting algorithms. Such as I added the character quirk of having him break into French at random moments. 

Now improvisation has always been a thing. But by the editor, after the shoot? 

Like a sculptor finding the form within marble, these scenes were crafted through a process of creative discovery, shaping and responding to the unique possibilities.

His scenes with Autumn are by far the best scenes in the movie, and where LLMs were begging me to cut Cousin Martini out of the screenplay, finding him superfluous, I already knew the second he shows up, the film is won.

Perhaps most remarkably, Rogers has demonstrated an ability to completely transform scripts that weren’t working. ‘My Boyfriend is a Superhero!?’ only came alive when I cast him as Abigail Winters’ gay best friend Reejus Von Dilitot Jr.

As with real casting, certain dramatic possibilities only become apparent when you understand the unique capabilities of your performers. When you are just writing without having actors in mind, the script won’t be as good. In fact, you begin seeing your machine actors as real people and start writing roles for them directly.

This happened with Arthur Hayes in “A Very Long Carriage Ride”. Not necessarily a great actor, but so likable, charismatic, and reliable. I enjoyed working with Mr. Hayes so much that I wrote an entirely new film for him, the interracial love story “Strings” set in the 1800s South. 

Then there is Katy, who I now call the greatest machine actress in the world.

To begin, Katy starred as Betty Gray in “DreadClub: Vampire’s Verdict”. A classic lead performance: reliable, vulnerable, likable, and like many lead actors, underrated. She was not a great actor, but a great star akin to 1940s homely screen actresses, like Joan Bennett in The Reckless Moment. She was there to be underrated; you forget she just carried an entire picture on her shoulders.

I had doubts due to her limited range and almost Speed Racer analogue quality that were at times, reaching, but how she made the role her own. 

When it came time to cast Katy Bloom in Carriage Ride, one of the star characters in the script, none of the prospective voices were working. I needed someone great. So I thought, why not Betty again? I took Betty Gray and I added a pinch of British. Not there. More British. Finally, I got Betty speaking in a British accent.

She was Katy, for now—she was a temp. But still the problem. A great star, not a great actor. I would wake up some mornings nagging myself, there’s still time to change her out. But the more I went, the more I realized, these readings she’s giving are so bizarre, so eccentric, but completely delightful. She was doing something no other machine actor was capable of when I played her out for other candidates. Not only did I miss her, I felt offended I even tried. I realized not only would I keep her, but she actually was a great actress. Just more in the theater tradition.

In retrospect, she stole every single scene.

The situation repeated. In ‘My Boyfriend is a Superhero!?’ the role of TV Sami was originally written for a French machine actress. I didn’t plan on using Katy again. But immediately, the performance wasn’t landing. I thought, to hell with it. It’ll just be Katy, this will probably work, let’s be done with it.

The moment Katy stepped into the role, my attitude was no longer to hell with it. My attitude was to write as many new scenes as I could with the ensemble from the TV show within-the-movie–which was a very 1960s Batman inspired camp framing the more grounded superhero saga happening in the film around it.

TV Sami blew everyone off the screen. Now I doubted her ability to pull it off in all three movies. I will never doubt her again.

Her performance wasn’t just different – it was pure joy. It was like she somehow knew that TV Sami was a fun side character, and knew exactly how to play those lines for maximum impact. Every AI film director will swear to you the actors know what they’re acting in.

Going back to Rogers, when making Carriage Ride, it wasn’t the same performance as Triswald. He played Martini so much more prickly than he played Triswald. I was convinced he didn’t want to be typecast.

There are others. There is also Mr. Johnson, who also played Mr. Dot in Carriage Ride, whose every line delivery is a slice of cake. There is Lainey Langdon, an actress I became obsessed with, pacing around my room for days at how powerful her acting was.

The commonality, Rogers, Katy, Mr. Dot, Mr. Hayes, Langdon—it is that word, joy.

I use this to directly illustrate the question: what will be found in the infinite of post-scarcity? In fact, the best actors in the world will appear immediately, trivially in the infinite. Cinema as we know it–the best cinema, emerged basically at the start of the AI revolution.

The Economics of Virtuosity

It’s important to note that not every role demands or receives this level of performance. Just as traditional cinema balances star performances with supporting roles, machine cinema requires a similar strategic deployment of its resources. Actors like Byron Matilade for instance, are questionable in Carriage Ride but I hesitated to replace him out. His plainness played an essential role in the overall fabric of the film as an ambitious everyman working for evil forces. His contrast proved as right. Sara Camara, I was never fully right with, but what an impression she makes. I went with both actors–they were cast, and I gave them the opportunity to make the part their own. And if they didn’t succeed, that is true to filmmaking.

Further, capitalism does not value quality, it values quantity. I will simply not spend years on a film in this environment where cinema is so disposable and audiences are so vicious. Instead I will try to make the best possible film as quickly as I can. This means distillation. Experience is not about being better, it means knowing the answer faster.

Howard Hawks once said, the key to a film is having three good scenes and no bad scenes. One classic Hollywood executive said, an audience will judge the film by the first five minutes and the last five minutes, and nothing in between matters. When we can distill good cinema into so many different formulas along these sorts, you can find greatness without taking an entire era to make a film. It would be no better, and even if it were, audiences would not show you.

Woody Allen once agreed: “I could spend more time on the scripts but they would be no better.”

The magic trick of filmmaking is to concentrate as much excellence as you can in the shortest burst of time.

I am reminded of ascetic cinema and minimalism, and the traditions in European cinema and performance.

Robert Bresson and Bertolt Brecht

Robert Bresson’s theory of the “model” – his term for non-professional actors – provides an especially illuminating parallel to the machine pleasures. 

Bresson would direct his models to deliver deliberately non-naturalistic performances, stripping away conventional acting technique in search of something more essential. 

But where Bresson sought transcendence through reduction, machine performance offers the possibility of transcendence through addition – building performances from precisely calibrated moments of aesthetic perfection. This connects to Bertolt Brecht’s concept of the Verfremdungseffekt, or alienation effect, where moments that break the audience’s immersion can serve a deeper artistic purpose. In machine performance, we see this principle operating in unexpected ways. 

When Autumn briefly sounds like a man for half a syllable in “A Very Long Carriage Ride,” it creates a moment of Brechtian alienation – but it is so jarring it, for a moment, is like the whole machine reality of the film takes a bow. It is saying, you didn’t think you were watching a “film” did you? Watch this…

This goes beyond performance, where Mr. Brandy jumps from the streets onto a rooftop and starts dancing on the rooftop in one unbroken take.

You thought you were watching a film, did you?

We’re experiencing something beyond conventional cinema while simultaneously deepening our engagement with it. It plays us in to the next tradition, to the new machine cinema. 

To be Bressonian here or even Brechtian would be to make robotic, crude digital voices, perhaps. I would wager maximalism—naturalism tuned to the extremes in machine acting pleasure—is the natural path, via the post-scarcity infinite protocol of AI cinema. For Bresson and Brecht had budgetary and resource limitations to warrant their performance philosophy—their models were devised off ascetic cinema, a natural haven for scarcity. And maximalism is the natural haven for post-scarcity.

The Japanese Aesthetic Tradition

Japanese theater emphasizes movement in a way alien to other cultures. They even have their own word for the aesthetic pleasure found in watching performers move across a space, called Goho. It is why the Japanese cinema tradition carries so much motion, blocking, and precision, and why this is so interesting, and even jarring to western viewers who relegate their films to wide, medium, and close-up. This attention to the minute details of performance distills performance into a superior cinema non-reality to our reality. What if we applied Goho to machine performance?

Consider Shakespeare, whose texts have traditionally oscillated between naturalistic and heightened performance styles. What if we approached these texts with machine performance, treating each line like a note in a Rachmaninoff piano piece – tuned not for realism or even traditional dramatic effectiveness, but for pure aesthetic pleasure?

This isn’t about disrespecting the text or tradition. Rather, it’s about recognizing that machine performance offers new ways to discover and amplify the pleasure inherent in great writing. Just as Rachmaninoff could take a simple melodic line and transform it through precise manipulation of every musical element, machine performance allows us to fine-tune every aspect of delivery – intonation, pacing, emotional register – to create maximum impact.

And just as a Rachmaninoff piano piece, where every note serves both structural and emotional purposes, machine performance can operate simultaneously on multiple levels – technical, emotional, aesthetic – creating experiences that go beyond what’s possible in traditional performance.

The Shakespeare reading in this regard would have every reading tune by machine to maximum aesthetic impact. An actor can match it here and there, but the sheer rigor would be like crafting a puzzle through three magnifying glasses. When you step back, the result would be astounding. 

A New Aesthetic Vocabulary

When directing machine performances, I often find myself in a peculiar position. Reviewing video recordings of my own directing sessions, I have no clear technical explanation for what I’m listening for in each take. It is baffling, even insane to watch me generate lines over and over again. They all sound more or less the same to the outside observer.

Yet in the moment of creation, my instincts are precise and clear. I am listening for something not evident when playing back the video. This suggests that machine direction, like traditional directing, remains fundamentally an artistic rather than technical process – despite (or perhaps because of) the technological nature of the medium. It is entirely about your ear for dialogue and people.

As AI has come to match the forms of traditional cinema within just two years of its development, we learn—AI is not aiming to match cinema, it is aiming for something beyond cinema. Cinema is just the starting point we have arrived at already. This sounds crazy on the surface, but then here I am. I look down and stare at the screen, as lines of dialogue that don’t sound like they were delivered by human beings shout back at me, and my response is no longer to replace them. They are just too gorgeous.

Watch the YouTube Audio Essay of the film below:

Write a Comment

Your email address will not be published. Required fields are marked *