AI May Drive Open Source Cinema
This summer, two filmmakers released The Odyssey. One was Christopher Nolan, armed with a $250 million budget, IMAX cameras, and one of the largest studio productions ever assembled. The other was Ash Koosha, working with generative video models, cloud computing credits, and a budget that reportedly sat somewhere in the mid-five figures.
Hollywood has interpreted this shift primarily as a debate about efficiency. Studios question whether AI can reduce production costs, replace visual-effects artists, or automate a script. Those are important questions, but they assume artificial intelligence is simply another way of making conventional movies. The more consequential possibility is that AI isn’t changing how films are made. It’s changing what a film is.
Denis Villeneuve describes filmmaking as a beautifully human, collective act of discovery that would be lost if a movie could simply be generated by a computer. The position carries an apt irony after Dune, a universe founded on humanity’s eradication of thinking machines to preserve human potential. Steven Spielberg accepts that AI may assist with budgeting, planning, research, or location scouting but rejects the “empty chair with a laptop” in the writers’ room and insists that it must never become the final authority on a creative decision.
These are serious defenses of human creativity. But they still define the problem through an inherited workflow: a screenplay is written, performers are cast, scenes are produced, and a director assembles a final cut. The more destabilizing possibility is that AI will not automate that sequence; it will dissolve it.
Historically, the studio system performed two essential functions: financing expensive physical production and discovering talent. The first function remains difficult to replace. The second has largely migrated to the open internet. YouTube, TikTok, Reddit, and other platforms do not merely promote aspiring filmmakers; they allow creators to develop a visual language, build an audience, and demonstrate that a story can hold attention before a studio executive ever sees it. The pipeline predates the current YouTube horror boom: A’Ziah King’s viral Twitter narrative became A24’s , demonstrating that a story born inside the grammar of a social platform could travel directly into feature filmmaking.
The shift became impossible to ignore in 2026. Kane Parsons transformed an anonymous internet concept and his YouTube series into , which has grossed roughly $380 million worldwide against a $10 million production budget. Curry Barker’s crossed $400 million after being produced for approximately $750,000. Mark “Markiplier” Fischbach self-financed, directed, distributed, and starred in , which earned more than $50 million on a reported budget of around $4 million. These were not creators whom Hollywood discovered and taught to tell stories. They arrived with stories, audiences, aesthetics, and proof of demand already attached.
At the emerging edge of the market, Hollywood increasingly buys outcomes rather than creating them. The studio begins to resemble a venture-capital firm: it supplies capital, distribution, legal infrastructure, and scale after product-market fit has already been demonstrated. Venture capitalists rarely invent the successful startup. They invest once someone else has shown that the thing works.
The subsequent gold rush makes the pattern explicit. A24 is reportedly negotiating a three-year deal worth as much as $65 million with Parsons, while Warner Bros., Universal, and Sony compete for him. Amazon MGM, United Artists, and Steven Spielberg’s Amblin won an 11-studio bidding war for Alex Kister’s YouTube series The Mandela Catalogue. Warner Bros. acquired Trevor Henderson’s Siren Head, and Sony’s TriStar acquired Cartoon Cat days later. Hollywood is now searching YouTube, memes, creepypasta, and collaborative internet fiction for concepts whose cultural relevance has already been tested in public.
But signing every successful YouTuber will not reproduce these results. The platform was not the breakthrough. Each creator had developed a specific narrative voice, aesthetic system, and relationship with an audience. Copying the résumé while ignoring the work is conventional Hollywood pattern-matching: buying the visible attributes of success after the causal machinery has already disappeared.
AI will compress this discovery cycle even further. The next outsider will not need to arrive with a short film, a screenplay, or a viral concept alone. A creator may be able to arrive with a nearly complete feature, multiple localized versions, audience-response data, and a functioning fictional universe. Hollywood’s role then shifts again-from discovering filmmakers to acquiring creators who have already financed, produced, tested, and distributed the first version of their own worlds.
Get Gideon Zettabyte’s stories in your inbox
Join Medium for free to get updates from this writer.
Open access to cinematic production does not guarantee an open field of imagination. The source material of AI cinema may increasingly arrive preformatted by online participation: premises designed to be instantly grasped, identities expressed through recognizable references, and narratives that invite copying before they invite interpretation. What appears to be bottom-up creativity can therefore carry its own form of standardization. The pressure no longer comes primarily from a studio note demanding a broader audience; it comes earlier, through the creator’s intuition of what an audience will recognize at all. This is how decentralized culture can produce centralized taste. The field becomes crowded with original makers, but the makers are working from a shared vocabulary whose boundaries were established through repetition, social proof, and the continual observation of what other people already appear to want.
To understand where this economic model leads, one must look back to Ari Folman’s 2013 film The Congress. Robin Wright plays a fictionalized, aging version of herself who accepts one final studio contract: her body, face, gestures, and emotional expressions will be digitally scanned, and the studio will own the resulting performer forever. It can place her ageless digital double in any film it wants without requiring her participation or consent.
The premise anticipates a fundamental shift from purchasing human labor to owning reusable performance assets. Voices, faces, bodies, environments, lighting systems, and crowd performances become modular intellectual property. A studio could replace a lead actor, alter a character’s age, translate a performance, or recut an ending years after release because no creative choice remains physically irreversible. The actor ceases to be an employee and becomes an asset library.
But The Congress contains a deeper warning about the audience’s role in this transaction. Folman’s film does not stop at the commercialization of the actor; it extends to the commercialization of reality itself. In its second half, the studio abandons conventional movies for a chemical entertainment product that allows consumers to hallucinate customized, solitary realities. The final commodity is no longer Robin Wright’s performance. It is the spectator’s capacity to withdraw from an inconvenient shared world.
That adaptation would not necessarily resemble an audience selecting alternate endings from a menu. It could happen continuously and invisibly. A thriller might accelerate when a viewer’s attention drifts. A romance might prolong uncertainty because the system detects a stronger response to unresolved tension. A horror story could change where a threat appears depending on where the viewer looks-or withhold the threat precisely because the viewer expects it. The work would no longer have a single rhythm, duration, or definitive sequence of events.
This changes the director’s role from image-framer to world architect. Authorship would reside less in selecting every shot than in designing a system of possibilities: the characters, dramatic boundaries, spatial rules, emotional trajectories, visual language, and events that the machine may-or may not-assemble for a particular viewer. The director would no longer determine every moment of presentation. The director would determine the universe within which presentation is allowed to vary.
Cinema may consequently divide in two. Traditional human filmmaking could become a premium craft valued precisely for the difficulty of its physical production: real performers, real locations, dangerous stunts, giant screens, and a director’s irreversible choices. AI platforms, meanwhile, may produce an entirely responsive medium that no longer behaves like a movie at all.
Hollywood keeps asking whether artificial intelligence can make movies more cheaply. That is the wrong question. The real question is what happens when a movie is no longer something a director finishes, but something a system continually revises. Cinema has always been organized around the authority of the final cut. We may be entering an era in which there is no final cut-and perhaps no single movie that every member of the audience actually sees.
Meme-ocracy asks the next governance question: what if we keep optimizing politics for outrage, followers, and spectacle until the state becomes a sponsored livestream? In that world, President Kardashian unboxes disaster relief, elderly citizens perform Jackass-style stunts for healthcare, and voting feels less like citizenship than airport security run by TikTok. The resistance comes from Amish farmers, analog weirdos, tradwives, and people who still remember that real work beats fake engagement. It’s Idiocracy with algorithms, breastmilk cheese, and a permacane parked over Florida.
Start here if you want the Gideon Zettabyte operating system in one place: AI, bureaucracy, defense, political theater, meme culture, airline design, institutional failure, and the emotional machinery of modern power. This anthology is part satire, part systems map, and part flashlight pointed at the things polite institutions prefer to leave in the dark.
Thanks for reading! Subscribe for free to receive new posts and support my work.
Originally published at https://gideonzettabyte.substack.com.
