The audience used to arrive at the end. You made the film, found distribution, launched the campaign, and waited to see if anyone cared. That sequence still shapes how much of the industry works, even as every part of it becomes less reliable.

This is bigger than a soft slate or a bad quarter. Hollywood is going through a shift it has not seen since silent films became talkies, and it is one of the biggest openings for creatives in a very long time.

AI lets us make more, but we already have more than anyone can watch.The distance between making the work and finding its audience is collapsing. People can follow an idea before it becomes a film. They can watch the visual language develop, understand what the creator is trying to do, react to early choices, and become invested before there is anything finished to promote.

We tend to discuss direct-to-audience as a distribution strategy. But it starts much earlier, as a creative strategy: build the relationship while you build the work.

The audience is not waiting outside anymore. It is already in the room.

The relationship is the asset

This is a total reset. For about thirty years, roughly 1980 to 2010,a small group of industry lynchpins controlled the gates. Dig a little and you found the same thing: those lynchpins kept a small, controlled circle around them and were careful about who they let in. That circle still likes to imagine it is protected. It is not, and it can no longer keep others out.

We are all still half-playing the game that they hold the keys, and we all half-know that they do not. Part of the reason is that for most of that period, entertainment depended on media buys and marketing campaigns, and the machinery behind those has changed. Television ad buys mean less every year, as fewer people watch linear cable and many pay extra to block ads. Billboards no longer command attention the way they once did. All of it drove the attention ecosystem for decades, and no real system has replaced it.

Online video, streaming, and social platforms had already pushed the old funnel wider. AI arrived while that system was under pressure. At this point it is broken, and AI finished the job.

A broken funnel does not make discovery easier. The old promise that good work will eventually find its audience becomes even less dependable.

You have to build your own hype. A platform can buy a project without putting much behind its promotion. A strong film can appear, disappear, and become one more thumbnail in a library before people know it exists. Creators cannot control every platform or guarantee reach. But they can build a direct relationship with the people most likely to care. That relationship travels. It can survive a weak launch, an algorithm change, or a move from one format to another.

We care less about creators owning distribution than about creators owning the relationship, because the relationship is what gives people a reason to follow.

The process can build attachment

The finished film is no longer the only point of contact. The process can carry part of the story. People want to feel like they are part of it, like they went along for the ride.

The chance to share your work, get immediate feedback during a livestream, build your audience, and bring them along with you has never been easier, or more fulfilling. Nobody stands between you and the audience now. No PR team tells you what you can and cannot say, no marketing department forces you into an outdated template built from focus groups, and no development executive who has maybe shepherded one or two titles out of development hell tells you how to change your characters. You live or die on what you create and on the audience and visibility you build around it.

That freedom still takes judgment. Creators do not need to turn every creative decision into content. The useful question is which moments help people care more deeply about the work. A visual test can reveal the world you are building, and a production problem can show what the project demands. Even a rejected direction has value: it clarifies the taste behind the final choice. Shared with context, these moments give an audience a way into the intention, not only the output.

Creative direction still has to protect quality and coherence. Someone has to decide what to open up and when to invite a response, and to recognize where participation would weaken the work instead of strengthening it.

Visibility should reveal a point of view

Going direct asks creators to be visible before release. For AI creators this can be tough, because you will likely find some AI haters in the comments who did not look at your work but want to voice their opinion on it anyway. Do not let that public sentiment keep you from putting your work out, or from creating right now.

Being visible does not mean feeding an algorithm every day. It means leaving enough evidence for people to understand who you are creatively. Share the work and the thinking that shaped it. Over time, those choices should let someone recognize your hand in them.

This matters more as production gets easier. As more people can generate a polished image, polish becomes less useful as an introduction. People need to see what you notice, what you reject, what you return to, and what you are willing to pursue past the first impressive result. That consistency builds a direct audience relationship through a recognizable way of seeing, rather than a rigid visual formula.

Recognition is not the same as attachment

The industry often treats familiar intellectual property as the safest and most predictable way to get attention. Recognition helps, but it does not create loyalty on its own. A familiar title can still feel empty. A new project can arrive with real momentum when the creator has spent years earning trust. People call those projects overnight successes. Only the release is new; the relationship behind it took years to build.

You can see it in what broke through this year. Outside of maybe The Odyssey, the films people cannot stop talking about are Kane Parsons’s Backrooms, Curry Barker’s Obsession, and Markiplier’s Iron Lung. None of those are legacy franchise IP. Each was built on years of a personal audience that kept coming back, felt invested in the creator, and above all was invested in the story that creator was telling.

Markiplier is the clearest case. Much of his audience had followed him for years, including watching him play through the Iron Lung game with all of its emotional ups and downs. They knew his love of it. In their minds he was trustworthy enough to bring them the final piece of a journey they had taken together, and casting himself as the lead only deepened that parasocial bond. It was brilliant marketing, and by choosing how he produced and released the film, he kept most of the profits. The film was not universally loved. For his audience it did not matter. They trusted him, he delivered, and whenever the next one comes, whether he steps further into a version of the studio system or not, he brings that audience with him.

That is exactly the audience studio marketing departments chase with every release, and lately it is like clawing at a cliff with sharp nails, watching the dirt crumble underneath with nothing left to hold. The reason is trust, and trust is running low. People are tired of being lied to in most parts of their lives, and that extends to entertainment. The curtain has been pulled back, and everyone can see the wizard is a smaller figure with far less power than the show implied, hoping to keep the illusion going a little longer.

You can put a whole lot of budget behind something, but if you are not engaging with what people actually want, rather than what you want to sell, you will find that gap. Staying engaged is not the same as handing creative control to an audience or chasing whatever performs this week. It means staying close enough to understand why people are there, which part of the world feels alive to them, and what promise they believe the creator will keep.

Release becomes the payoff

Going direct can work alongside studios, distributors, and platforms. It changes the terms of the relationship. A creator with an audience brings people who already understand the work and care what happens next. A distributor can extend that relationship instead of manufacturing interest from zero.

It also changes how a project develops. Audience-building cannot be left to the final weeks, when the creative team hands finished assets to marketing and hopes a campaign explains why the work matters. The story around the story has to begin while the project is still alive and people have room to enter.

Creators should not have to become full-time marketers, but communication belongs inside the creative system. The person sharing the work needs the same clarity as the person directing it: who is this for, and what are we inviting them to stay for?

As production keeps getting faster, this relationship only grows more valuable. Tools will improve and interfaces will converge; today’s specialist feature will become tomorrow’s default setting. Access to production will keep spreading, and attachment will stay as hard to earn as ever.

The creators with the strongest position will start asking people to care long before release. They will build a point of view in public, and they will let the right parts of the process carry meaning so the audience is part of the project’s life from the beginning.

Then the release is not an introduction. It is the moment a relationship becomes a film.

That was the itch behind Page One. Not another event, not another showcase. We take direct inspiration from Y Combinator, TechStars, Antler, and more from the tech ecosystem. This is a cocoon, a support network, and an incubator for great storytelling, ideas, and medium. At Machine Cinema, we believe strongly that stories can change the world.

But the thing is, we’ve never done this before. We’re in uncharted territory. When we first embarked on doing GenJams, no one had really figured out how to make films in a short period of time with AI. We were all figuring it out.

From the moment we did our first GenJam, we always thought how do we expand the scope? GenJams are usually 4-hour film sprints to make a music video, short film, or commercial. But we’ve done longer ones too. 12 hours, 48 hours…but Page One is basically an 8-Week GenJam on steroids. And we’ve taken the same approach we took to GenJams, let’s learn how this actually works as we go.

That’s why we flung ourselves head first into running an accelerator program. We thought, let’s build this plane as we jump off this cliff. Let’s work with creators, lawyers, business people, our community, and our mentors to figure out what does an accelerator program look like for the AI era. Let’s just put it out there and see what happens.

More than sixty people applied. We took twenty-three. Eight weeks later, nine finished projects walked out the other side.

A massive thank you to Digital Garage. Our kind hosts gave us space and support. We could not have started without them. They’ve supported multiple communities in the Bay Area who want to grow and we’re looking forward to doing more with them.

That math is not a failure. Some people combined into teams. Some got pulled under by their actual jobs and lives. A few realized the thing they came in to make wasn’t the thing they wanted to make. All of that is the program working. Finishing is supposed to be hard, and the nine projects below earned it.

What we were trying to do

Here is the part most launch posts would never admit: we didn’t fully know what we were building.

Startups have had Y Combinator for twenty years. Filmmakers have had film school and the studio ladder for a century. There is no equivalent for a creator building original IP with tools that reinvent themselves every few weeks, because nothing around them is settled. There is no accelerator for AI. Yes, there are AI studios, events, competitions, funds, and AI tool companies financing creators. There is no dedicated accelerator for AI IP. Nobody can tell you with confidence how to copyright AI-assisted work, what protection you have over a character you generated, or how an AI-native creator is supposed to make a living. We are building on ground that is still moving.

So we did two things. We brought the hard questions into the room: lawyers, investors, and agents, because when someone asks whether they can protect the character they just built, that deserves a lawyer, not a shrug. And we told the cohort from day one that we were figuring this out alongside them. They weren’t just there to finish their projects; they were co-authors of how the program ran. The structure of Cohort 02 is something the first cohort built with us, mostly by showing me where the first version was wrong.

The mechanics were unglamorous. Eight weeks in San Francisco, in-person sessions, Friday mentoring calls, a WhatsApp thread that never slept, and an irresponsible volume of pizza. The Friday Mentor calls are where the two worlds met: Fred Grinstein on the television and creator landscape, Khayyam Khan on pitching in the AI era, Will aka Blvcklight on world-building, Evan Baily on producing and IP, Martin Gaiss on brand storytelling, Jesse Z on the microdrama market, Bailey Kobe on screenwriting. Halfway through, Seedance dropped and half the room rebuilt their pipelines overnight. That is what this medium is like right now. The goalposts move while you are running at them, and the people who finish are the ones who keep running anyway.

What we learned

Clarity beats freedom. We kept the definition of “done” loose early on, thinking it gave people room. It just gave them anxiety, and anxious people polish instead of building. The moment we got concrete about what would be on screen at Demo Night, the work got better within a week.

Teams beat soloists. Some of the strongest work came from people who arrived alone and combined forces inside the program. Finishing is partly an emotional problem, and it’s easier to keep a promise made to a collaborator than one made to yourself.

Hard questions raise ambition. The weeks with a lawyer or investor in the room were the weeks the work got more serious, because the question shifted from “can I make this” to “can I own this, and can I sell it.”

And one on me: the work was finished in May and this post is going out in July. That delay is entirely mine, the cohort has been extraordinarily patient, and it won’t happen again.

The Cohort

Tracy Swedlow walked in with a musical about Philo Farnsworth and walked out with The Living Room, a psychological thriller anthology where every story happens in a different living room and the television has become an ominous, static-fizzling presence with power over the room. She pitched us by video link from her own television conference, which nobody planned and which was perfect. Setting down a project you’ve carried for years because the room showed you a sharper one is not a small thing.

Adam Mutchler took us somewhere else entirely with Sigil Protocol, a feature-length anthology of Jungian dream logic, cryptography, gnosticism, and hard science fiction. It sounds like the fringe, but the fringe is where a lot of the audience now lives, and the question underneath his project, how idiosyncratic work finds people without being handcuffed to platforms its creators don’t own, is the right one.

Vesica Theissen made The Glitch, an Art Deco surveillance thriller where the state’s magic is cold and geometric and the heroine’s is chaotic and alive. Vesica is an anthropologist and former dramaturge who lived on both sides of the Berlin Wall, and the piece is built out of that. A slow burn, deliberately far from the AI pyrotechnics that have dominated the last two years of demos.

Daniel Jacobs brought the energy back after the pizza with Galactic Fighting Championship, a vertical fight promotion where humans, robots, and beasts beat each other senseless. A crab creature. A goth witch. The moon. It should be disposable and it isn’t, because Daniel spent his time on the backstories, and it’s the only project in the cohort that is really a format rather than a story, with an open universe for other creators’ fighters.

Anand Parashar was the reality check in the room. While the rest of us chased the esoteric, he calmly explained that preschool video on YouTube is enormous and badly served, and that BamBam Rhymes puts the child inside the nursery rhyme instead of parking them in front of decade-old animation.

Blair Adams has more GenJam wins than anyone in San Francisco and could have coasted on a sizzle reel. Instead he made Adventures in ADHD, rendering the inside of an ADHD brain as a kaleidoscope of thoughts competing for a kid’s attention, because he has ADHD. It has heart, the thing AI work is constantly accused of lacking, because he put it there.

Paul Byrd built Katsumi the Electric Stealth Assassin, and the word that kept coming up was lore. He’ll tell you he wants kids grabbing Katsumi merch at Target, then show you the commercial. Most people are still arguing about whether AI can produce one good shot. Paul built an IP with a theme song and a retail campaign.

Mark Day came in solo and left as one third of Prompt Ponies, alongside comic creator and vertical drama specialist Eileen Kaur Alden and AI creative director Roshni Singh, fronted by The Mermaid Princess. They were the only team that refused to treat distribution as somebody else’s problem: the social strategy came baked into the pitch. Producing a team that didn’t exist eight weeks earlier is the second best thing an accelerator can do.

And then there was David Ronan, who drove up from LA most weeks, and who spent the early program playing with science fiction before turning toward a story he’d been circling for a long time. The Space Between is based on true events chronicled in the memoir My Son and the Afterlife: a young man struggling with depression, his death by suicide, and an aftermath where skepticism and the supernatural refuse to stay separate.

I’m not going to be clever about that one. Sometimes the act of making is what pulls a story to the surface, and I’ve never seen a clearer case. The room went quiet. It was the last thing anyone expected an AI accelerator to end on, and the only possible ending.

Step back and look at the shelf these nine sit on together. A Cold War surveillance thriller next to preschool nursery rhymes. An anime assassin with a Target commercial next to a gnostic anthology feature. A fight promotion where the moon has a backstory, next to a story about grief so real it silenced the room. No studio would have developed this slate. No genre, no format, no demographic connects it. What connects it is one community, and that thrills me more than any single project could, because it means the AI creator movement is not a niche with an aesthetic. It’s the whole spectrum of people who always had stories, finally holding the means to tell them.

What’s Next for Page One Accelerator

Becoming a filmmaker used to mean being born near an industry or moving to one, finding money before you had anything to show, and surviving a decade of gatekeeping. That filter mostly tested for endurance, proximity, and wealth. It’s now obsolete, and nobody has built the replacement. A seventeen year old in Lagos or Hanoi or São Paulo can produce a shot that would have cost a studio six figures a decade ago, and still, almost none of them will make anything anyone sees. Not for lack of imagination. Because there’s no path. No one finds them, teaches them, gives them a deadline, buys the work, or pays them.

So here is what we want to build, plainly.

Find the best AI-native creators in the world, especially in places that have never had a film industry. That’s what the community is for, and why it lives in over 60 countries across our ecosystem. Two hundred GenJams is not just a marketing number, it’s a scouting network.

Educate them, which is why the mentors matter. The gap between someone who can prompt and someone who can direct is the whole ballgame, and it closes when a person who has done the work for thirty years tells you the truth about your script.

Equip them with the structure and tools finishing requires: tokens and credits, a deadline, a room, a cohort, someone who notices when you aren’t there.

Distribute the work, so finishing means more than a post that falls off the feed. We worked with five distribution partners that would consider the best from the cohort to stream on their platforms.

And eventually, finance them. Not placing finished work with partners, which is where we are today, but backing creators themselves, on the strength of what they’ve made rather than who they know. If Machine Cinema works, an AI-native creator anywhere in the world can be discovered, developed, funded, and distributed without moving to Los Angeles and without asking permission. Do you want to help us finance the next generation of creators around the world? Our aim is to finish multiple cohort experiments in America, and eventually launch online and global Page One cohorts, focused on helping creators everywhere. This is why we work on global initiatives like the GenJam Director’s Cup, we worked on with OpenArt.

We know how that sounds. It’s a lot to hang on nine people in a room in San Francisco. But the room worked. People finished their projects and produced an impressive range of very thoughtful and profound work. And they were so kind in co-crafting this program, looking towards the future of AI creation.

This is only the first inning.

Cohort 02 in Los Angeles: Coming Soon.

Cohort 02 is in Los Angeles, and Fred is running it. The focus is horror and thriller: the genres having a genuine moment right now, and the ones that have always rewarded craft over budget. One of our biggest learnings from Cohort 01 was focus and structure would help a lot with everyone. A great horror scene is built from restraint, timing, and what you choose not to show, none of which can be bought. If AI-native production is going to compete on quality rather than novelty, horror and thriller is where it has to prove it.

Cohort 02 also opens up what a Page One project can be. Teams will build verticals, microdramas, episodic series, and social web series, the formats native to how people actually watch now, alongside traditional forms. The mentor network comes with it, Hollywood and creator economy both.

Cohort 03 goes back to San Francisco, and it’s predominantly about video games. Generative tools are about to do to game development what they’ve done to film, and almost nobody is building the talent pipeline for it. After that we keep rotating: new genres, new mediums, new cities.

Applications for Cohort 02 are open now at machinecinema.ai/page-one-accelerator. If you’ve been sitting on something and you’re tired of it being unfinished, this is the one.

If you’re interested in helping with future cohorts, with mentorship, partnership, distribution, or anything, please reach out to us. Thank you.

Cohort 01 ran at Digital Garage in San Francisco. Thank you to Digital Garage, Google DeepMind, OdysseyML, ArtCraft, Rosebud.ai, FlikAI, and ImagineArt, to our mentors, and to our distribution partners Fairground, Escape.ai, and Alchemy.

But let’s hear it from their own words. Mark Day, one of our amazing creators from the Page One Cohort 01 recounted his experience in his own words:

The night Machine Cinema became pure cinema.

It’s the climax of the AI creator-community Machine Cinema’s first Page One Content Accelerator program.

Demo Night, on a Tuesday night, at Digital Garage, a coworking tech space on SF’s Market Street, where the walls are lined with physical media.

And not just any physical media, but shelf after shelf of Japanese manga. A genre where every page feels like a storyboard, packed with perspective, character and cinematic action.

According to the ubiquitous Luma invite, the Page One cohort had been “heads-down in San Francisco — writing, shooting, prompting, training, breaking things, fixing them, and building original IP with the full stack of AI tools. No theory. No vibes-only demos. Real IP, real pipelines, real proof.”

And it was that.

But, as the AI would tell you if it was writing this, it was also much more.

It was (with all due respect to my fellow cohort members) the classic “rag tag crew of misfits with individual dreams and a common purpose, coming together to do something special” story.

You know the trope. The kids who come together to stop the developers from demolishing the rec center to build a business park. The Goonies hunting treasure in the caves under Astoria. The Seven Samurai, but with laptops and prompts.

And whoever scripted the final act (the universe, perhaps?) stuck the landing.

A room filled with the raw energy of ideas being smooshed into an assortment of content shapes using tools that speedrun their own Darwinian evolution.

You can see why folks like noted AI skeptic Seth Rogen make shows about people making movies.

Because in a cohort like this, everyone shows up with a story.

Or a story about a story.

And how they want to tell it.

Why it’s personal. A passion project. Or why it’s just perfect for that whole exploding vertical drama market. Or some combination of the above.

There was something slightly old-school San Francisco about the whole thing.

Tech may have displaced a lot of the odd events discovered via alt-weeklies. My memories of weird cable-access TV are dim. And even I have to admit that the bulk of contemporary Burning Man art is a little less chaotic and DIY than it used to be.

But the spirit of “want to see a thing? Just make the thing”? That’s very SF.

Sometimes GenAI feels like “cable access CGI,” even when the projects are aimed at emerging distribution channels, vertical screens, well-established genre audiences and beyond.

So what happens when you put out a call for creators with “taking the next step” aspirations?

Things get nerdy pretty quickly.

Yes, there’s typing words into boxes to make pictures come out.

But there’s also node based workflows. Workarounds that might not scale. AI video platforms that have failed to generate coherent UX. Frustration at bottlenecks that eat up gen-credits and patience.

Then delight when things go right.

What if the much-anticipated Seedance model solves everything?

Spoiler, it doesn’t, entirely.

But it showed up midway through the program, and it does solve a lot of things. Up to a point.

The goalposts keep moving, even as the AI-averse sit in the stands, grousing.

This is the culmination of a couple of months of in-person mingling, Friday afternoon Zoom mentoring, WhatsApp message threads. Machine Cinema founder Minh Do’s bad jokes. And pizza.

Between Page One and assorted GenJams, my diet seems to consist of AI pizza.

And tonight was no exception.

There was pizza. There were people, including a smattering of industry folk. Creators, hungry for a new narrative. Maybe a redemption arc or two.

Tracy Swedlow started the program with a musical based on the life of TV pioneer Philo Farnsworth, before pivoting to the place many a TV calls home, The Living Room.

It’s a peek at a surreal, Black Mirror-esque anthology where the television is no longer a World’s Fair wonder but an ominous, static-fizzling box that becomes a dominant presence in the room.

To add to the prevalent screen-theme, she was telepresent on Zoom from the TV of Tomorrow Conference, which she executive produces. A scheduling issue, but it added to the vibe.

From the Twilight Zone/Outer Limits intrigue of the claustrophobic Living Room, we went deeper into the strange conceptual multiverse of experimental, out-there AI, with Adam Mutchler as tour guide.

We live in an era when the liminal spaces of the Backrooms goes from meme to hit movie, and AI artists like Gossip Goblin are getting the kind of mainstream coverage once reserved for the shock tactics of the Dor Brothers.

It’s a curious constellation of creativity that Adam Mulcher’s ambitious project hopes to connect the dots between.

Ready for movie-length anthologies of content rooted in Jungian fever dreams, cryptography, gnosticism, hard scifi and conspiracy theories?

That may sound fringe. But the fringe is a draw these days.

My own anime-loving teen digs liminal spaces, body-horror manga and esoteric story telling. My own AI fandom leans towards the colorful surrealism of niceaunties.

Do we need alternative ways of enjoying esoteric creativity? Ideally not handcuffed to platforms creators don’t own? Adam has thoughts about that.

Two presentations in, we’re not in Kansas anymore, Dorothy.

We’re getting off-the-wall thinking. The thinking behind the thinking. Video that might make you want to lock the TV in another room. And storytelling that, like good science fiction, uses futuristic scenarios as a prism to frame personal experience.

Next up, The Glitch is art-deco, surveillance state scifi infused with the first-hand experiences of life on both sides of the Berlin Wall, from former dramaturge (look it up) and anthropologist Vesica Thiessen.

With nods to cult sci-fi noir Dark City, A Handmaid’s Tale and international academy award winner The Lives of Others, The Glitch is more slow-burn than the showy AI pyrotechnics that have dominated AI demos, a trend we seem to be moving beyond.

Following a break for carb-loading AI pizza, Daniel Jacobs brought the energy back with his nascent Galactic Fighting League IP.

It’s an unhinged take on UFC-style combat paired with classic wrestling character tropes, smacked on the head with a folding chair made of AI graphics, and built for TikTok, YouTube and guilty-pleasure energy drinks.

Watching a crab creature or a goth witch go at it in the GFL arena is fun. But it’s the backstory moments that make it more than Mortal Kombat gone mad, as Daniel’s own ragtag collection of unlikely comrades, including a school teacher and the moon (yes, the moon) give it a Fast & Furious “it’s all about family” vibe.

To remind us that the esoteric Bay Area is also home to focused AI entrepreneurship, cucumber-cool BabBam CEO Anand Parshar showed us all what market-minded business building looks like.

With nursery rhymes. F’real.

Reality check! Pre-school kid content on YouTube is huge. And the BamBam platform gives that overlooked media reality a personalization spin, letting parents create videos where kids get to be in the videos, not watching years-old YouTube clips with dated animation.

Old McDonald had a … timeless, agentic, scalable business plan. Or at least, he does not.

It’s not quite the personalized storytelling seen elsewhere on the night, but Anand’s customers have no complaints, and the output’s personal to the kids watching the videos, over and over.

Blair Adams is a stalwart presence in the SF AI scene, with a sizzle reel that shows why he’s the current holder of the Galactic Fighting League belt for most GenJam wins.

He’s got traditional film making chops, lives alongside Paul Byrd at the edge of what the technology can do, and the moxie to call his AI creative studio Imitation Films.

He’s also got, it turns out, a little ADHD.

And while his reel has a “could turn his hand to anything” vibe, his work often feels subtly personal.

So when he applies his technical prowess and evident good nature to something that feels more like IP than stand-alone showcase, with a first look at his edutainment series Adventures in ADHD, it’s got heart and soul.

You know, those things AI ain’t supposed to have.

Alongside plenty of “how’d he do that?” animation that captures ADHD information overload perfectly, with a kaleidoscope of thoughts vying for his young protagonist’s attention.

No two projects felt the same. Though everything carried personal fingerprints of some form or another.

Katsumi the Electric Stealth Assassin, Paul Byrd’s high voltage anime series in development may not flag up the creator as an undercover assassin.

But with the word “lore” getting thrown around a lot, Paul’s deep love of the genre feels as personal as anyone’s leap from thinking about it to making the thing.

There’s nothing stealthy about the world building on offer, and Paul won’t just tell you he’d love to see kids grabbing Katsumi merch at Target. He’ll show you.

Me? Having connected with comic-book creator and vertical drama maven Eileen Kaur Alden, and dedicated Roshni Singh, we studio’d-up as Prompt Ponies around Eileen’s slate of vertical drama and horizontal fantasy IPs.

With interest in vertical drama peaking, and more opportunities for AI content opening up in general, there’s already interest in The Mermaid Princess.

While this commercial AI studio slate may be more focused on “what the market’s looking for” than some of the Page One cohort’s solo outings, our preso came with social media strategy baked in, and the versatility to move beyond the standard AI short-film calling cards into, well, what Machine Cinema promised.

“No vibes-only demos. Real IP, real pipelines, real proof.”

And if none of this feels quite like the cinematic ending I talked about earlier, perhaps you had to be there, sitting at tables with folks mostly outside the traditional media infrastructure, trading prompt advice and hot takes on new AI models. Getting to know people, and still able to be surprised when they lay it all out there.

And for me, the real season finale/end of the movie moment arrived quietly and captivated the room, via David Rohan and his project The Space Between.

David had been driving up from LA most weeks. He and I messed around with some scifi serial production, genning robots, cryo chambers and astronauts on a space breeding mission (yes, really, see the Prompt Ponies for details).

Eventually, though, he started applying new AI skills to a story he’s been orbiting for some time. Based on true events, chronicled in the memoir My Son and the Afterlife, it’s the story of a young man struggling with depression, his suicide, and an aftermath where skepticism and the supernatural intertwine.

I mentioned Burning Man earlier.

Having documented the creation of the 2016 incarnation of the event’s sacred space, known colloquially as the Temple, I recognize that sometimes stories call people to create.

And sometimes the act of creation is what draws stories to the surface. Particularly when those stories are deeply human, vulnerable and perhaps hard to articulate.

Which is why I believe creativity is for everyone, and damn the gatekeepers.

You spend a couple of months making your thing. Alongside people making their thing.

Then one night, out of nowhere, you’re watching someone talking about life, death, the aftermath of loss, and experiencing it through sadness, and humor.

And wanting to tell that story visually, so it can travel, from screen to screen, touching other people in the process.

Whatever expectations we had coming into the Page One program, I’m not sure any of us imagined that sort of very human ending.

On a night like this, Machine Cinema is just cinema. Cinema by people, for people.

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

IN THIS EDITION…

This was the week the group chat got up, walked outside, and met itself in person. AI on the Lot turned Culver City into a two-day community reunion — and the channels have not stopped buzzing since.

But it wasn’t all swag scrambles and karaoke. Amazon’s big GenAI announcement sparked the sharpest debate we’ve had in months, a Cannes feature got a hard look at its real economics, and the community kept circling the same question: when does an AI creator actually make consumer money?

Let’s get into it.

Machine Cinema Channels

TOP TOPICS…

AI on the Lot Turned Into a Full Community Reunion

For two days, Machine Cinema basically relocated to Culver City. AI on the Lot became one long IRL hang — morning coffee at the Culver steps, a chaotic GenJam, late-night Ktown karaoke, and a swag scramble for the famously scarce MC totes and tees (”Caught a wild Minh and all I got was a tote bag” pretty much summed it up). One member called the whole thing “AI Disneyland” and credited it with launching her AI journey. The consensus from everyone who showed up: the digital hangs are great, but the IRL energy hits differently.

[LA]

The Avatars Were Male, the Tickets Got Pricier

Two smaller AI on the Lot subplots got the chat going. First, the event’s avatar generator — which masked attendees as animals — kept defaulting to male, suit-wearing creatures (one member’s arctic fox arrived with human breasts in a tank top; another had to re-roll repeatedly just to get a female). Second, the last-minute jump from $500 to $700 tickets stung, though several conceded that for arguably the biggest AI entertainment conference around, it might still be underpriced.

[LA]

Higgsfield’s Cannes Math Got a Reality Check

Francesco opened the week with a numbers puzzle: Higgsfield’s Cannes feature reportedly burned ~108,000 generations for a 90-minute film — roughly one usable shot per 200 attempts, a 99.5% discard rate. Did that scan? The room mostly landed on “high, but not crazy”: a controlled TV shoot can run 30–50:1, and directors often want variety across their “takes.” Minh added the kicker — at ~$500K (mostly platform credits), the film reads less like a business and more like a Higgsfield marketing line item.

[BASECAMP]

Amazon’s GenAI Fund Landed With a Thud

The week’s biggest debate: Amazon MGM’s new GenAI Creators’ Fund. Within roughly 24 hours of the announcement, Punky Duck creator Jorge Gutiérrez scrapped his AI Amazon project under a wave of backlash — and the chat split hard on who to blame. Several argued Amazon botched the rollout, leading with self-congratulation instead of showing the work. Others saw a familiar anti-AI pile-on. Danny offered the level-headed read: the narrative only shifts once “a wider audience is truly engaging with the content.” John capped it: “Success is a powerful deodorant.”

[BASECAMP]

When Does an AI Creator Make Consumer Millions?

Minh kicked off a meaty thread off the back of Obsession — a reported ~$1M production now tracking toward $100M at the box office. The real question underneath it: when does an AI-native creator pull consumer money like that, not just brand deals, commissions, and merch? Blvckl!ght, who’s logged 30M+ views, argued the bigger opportunity isn’t longer and longer films but generative experiences and games — “the riches are in the niches.” Predictions for the first AI creator to hit consumer millions ranged from six months to two years.

[BASECAMP]

🏆 COMMUNITY WINS…

A big week for the community on stage and on screen:

• Max Gallen — Took home the $10,000 top prize at the Campus Greenlight Pitch Fest at AI on the Lot, the student pitch event Fred Grinstein helped dream up.

• Jennifer Delalande Lim & Tom Pustel — Won a Hollywood realism contest and launched a new YouTube channel off the back of it.

• Ibraheem Diab — Picked up an award at WAIFF in Cannes.

Sutu — Launched Cyboracle Prime, a digital-twin experience built with afrofuturist icon Nona Hendryx (Labelle, “Lady Marmalade”), letting visitors collaborate on a song with her cloned voice.

Share

🔗 This Week’s Link Drops

Industry News

• Obsession Is About to Hit Two Wild Box Office Milestones — Forbes

• AI Film “Dreams of Violets” Lands Tribeca Festival Premiere — Deadline

• Films More Likely to Star a Man Named Chris or a Talking Animal Than a Woman Over 60 — Variety

Tools & Tech

• ElevenLabs Music v2 — ElevenLabs (the week’s big AI-music talking point, alongside Suno, Sunauto, and Audiogen)

• Runway’s Latest Update — RunwayML

• Kling AI Rollout — Kling AI

• AI Creative Market Map: The Tools (2026 Edition) — Machine Cinema

Worth Watching

• Doug Liman + 30 Ninjas “In the Chair” — YouTube

• A 1-Hour AI Game of Thrones Video Winning Over Superfans — YouTube

• Star Wars Superfan AI Art — YouTube

• Magehold: A Vertical AI Dream Series — YouTube

Resources / Deep Reads

• Drawing the Line: Principles Launch (Virtual, June 3) — C4OSL

• Cyboracle Prime: The Nona Hendryx Digital Twin — EyeJack

Upcoming Events…

The Lot may be over, but LA isn’t slowing down. A few worth a look:

AI Policy June Meetup 📅 Wed, June 3 · 6:00–8:00 PM — IP law, synthetic likeness, and training-data compliance → RSVP

From Months to Days: How AI Is Rewriting Commercial Production 📅 Fri, June 5 · 8:15 AM → RSVP

World Models Panel @ Augmented World Expo 📅 Thu, June 18 (AM) — reach out in Basecamp if you’re building in the space

And keep your calendars open: there’s a big Machine Cinema competition coming in July.

Machine Cinema Events

Success Is a Powerful Deodorant…

This week the community stopped scrolling and started shaking hands — and somewhere between the tote-bag chaos, the karaoke, and a very public backlash, the same lesson kept surfacing: keep making the work, keep showing up, and let the haters hate. As one longtime member put it, success is a powerful deodorant. Go earn yours.

— Machine Cinema Newsletter.

(Note: Click here you want the full PDF version of this doc.)

Let me start by saying that making any market map in the AI era is impossible. The space is moving so fast that as soon as you publish, it’s aged. This market map will probably be useless tomorrow. It’s not just speed, it’s the breadth. The space has become so vast that having a perfect 360 view is futile. But it doesn’t mean we shouldn’t try to conceptualize what’s happening.

At Machine Cinema we think it’s vitally important to track this because our goal is to aggregate this trillion-dollar creative economy as it goes through transition. When we did the first ever GenJam, our AI signature event, in 2023, the only tools at our disposal were: ChatGPT, Runway, Dalle3, and Midjourney. When I did this market map last year in Q2 2025, it was about 100 tools.

This year’s market map is over 250 tools! In 2025, I called this the “AI Storytelling Market Map” because the bulk of the conversation in early 2025 revolving around Hollywood and production. But the reality, the revenue and audience has expanded far beyond Los Angeles (arguably from day one). Elevenlabs, the leading AI voice model many AI filmmakers use for character voices, gets a majority of its revenue from enterprise accounts focused on customer service and internal comms. Midjourney, the world’s leading image model, by users, with revenue rumored to be above 10 figures, classifies a large bulk of its users as “art therapy”.

Companies have also disappeared from last year. OpenAI’s Sora2 is gone. Pika has shifted from a video model to a video agent. In the last year, many of the AI tool partners we talk to are interested in agency and enterprise sales rather than consumer.

Despite the fact that Hollywood, film festivals, and celebrities dominate the ideological airwaves on the ethics and permissibility of AI usage, the revenue and gravity of the space is elsewhere.

Subscribe now

Things to keep in mind:

• There are multiple logos in multiple places because that’s simply the nature of the space. Runway, in 2023, mainly placed all of their own models in their creative suite, but by 2026, they’re hosting Seedance and other competitors under the hood. Everyone wants to be everything. For lack of moat, become the ocean?

• Although categories like aggregators and editing have grown, image and video models have remained the same. Although new entrants like Alibaba’s HappyHorse and Grok have entered the fray, building models is an expensive endeavor for hyperscalers, unicorn-sized venture-funded startups, and ad-supported entities.

• Models remain the power center of the space as we saw with Seedance. To some extent, we are all downstream of model providers. When Seedance arrived onto the scene, aggregators and inference providers rushed to supply it to their users, for fear of losing them to other providers who had Seedance. As models get better they enable all the tools downstream to offer more capability to their users. Not only are the aggregators downstream of the models, but so are the AI studios and creators, although our stories, projects, and perspectives can be unique, the quality of our productions is more dependent than ever on what the models enable.

• Despite models remaining stable and a proliferation of tools that offer everything under the sun, there’s been a steady rise in niches. Companies like Zeely focus solely on AI filmmaking tools for ads. Similar tools exist for AI microdramas. These tools offer the same features and models as the general tools, but targeted at their niche. The application layer remains undecided.

• Given the above, it’s hard to keep track of many categories. Note that Production Assistance, AI Ads, AI Design, AI Production Studios, AI Social, 3D, Video Gaming, and Interactive could deserve their own market maps. This market map serves as a window into these rapidly growing areas.

• We have deliberately omitted AI Production Studios, which although many have raised venture money on tool building, have not released these tools to the public. We have also noticed a slow down in funding for production studios despite a continued rise in venture funding for tool companies, in particular aggregators, 3D, world models, video gaming, etc..

• We have seen a rise in dominance of Chinese companies coming into the space. Kling, Seedance, HappyHorse, Minimax have stayed in the top 10 for over a year. Now we’re also starting to see Chinese aggregators enter the market as well. Most of the logos on this map are from America, will this stay true by next year?

What originally started out as a space for filmmakers and designers has expanded far beyond and is not abetting.

For us, Machine Cinema has to move where the space is going. In early 2025, no one was talking about programmatic video, world models, 3D, AI video gaming, and multi-shot agents. But today, they are driving a lot of the intrigue in the space. In response, we have hosted events, workshops, and our GenJams focused on AI in immersive, games, world models, design, and more, actively experimenting with what’s coming out. Learning and creating at the same time has helped us to categorize where these tools sit within a larger creative pipeline.

Ultimately, we are driven by one central question: What does an AI creator look like tomorrow?

Enjoy the market map. We hope it serves as a guide as you plumb the depths and make your storytelling dreams come true, just don’t spend too many tokens on the way!

If you’d like to add a company, give us feedback, or just say hello, please feel free to jump over to our website: https://www.machinecinema.ai/ai-creative-market-map

Appendix

A few notes on each category.

Development / Pre-Production

Everything that happens before a frame is shot or a pixel is rendered: writing the story, planning the shots, and organizing the production.

Writing has overwhelmingly been the domain of ChatGPT, Claude, and Gemini. But more niche products like Sudo Write have arrived onto the scene, building core features around what screenwriters need. Cursor is also an unlikely entrant as a writing tool, since it’s originally for coding, but the way it handles multiple documents at once makes it a surprisingly powerful tool for character and story consistency and development.

Storyboarding has evolved considerably and frankly many of the tools that were formerly storyboarding tools have since migrated to being full-fledged aggregators simply because it has become so easy to add image and video models to any given tool.

Production Assistance is a newer category that we felt needed to be featured because we’re starting to see real competition in the space, with startups building the necessary rails to help streamline the work of producers.

Production

The stage where the work actually gets made — image, video, audio, 3D, and the interfaces that orchestrate them. This is the largest and most contested band on the map, and the one where AI’s impact is most visceral. As mentioned, image and video models remain heavily contested with a battle between mainly Chinese and American well-funded corporations. The rest of the field is a total war. Aggregators are super crowded with Infinite Canvas and Agentic Chat as rising sub-categories. Elevenlabs and Suno have retained dominance in their respective categories, but if an incumbent like Spotify enters the fray, what does that mean?

My favorite row in the production tab is the bottom 5. Programmatic Video is so exciting because it allows coders and vibe coders to tinker with bleeding edge APIs and agents to make their own tools or workflows. AI in Ads & Social should frankly be two different sub-categories but they have heavy overlaps and are still emerging. 3D Production is where I think it’s possible the best bleeding edge work for professional AI craft is. With AI tools having the reputation for being so flimsy and unpredictable, merging 3D with GenAI feels like the best of both worlds: precision and imagination. Speaking of worlds, World Models have been the hottest thing to talk about for both LLM researchers obsessed with AGI and AI filmmakers excited about the game-like potential of 3D worlds. The irony, of course, is that world models have possibly the most relevance to robotics, not the realm of creative production at all. Thus, underlining the truly sprawling nature of artificial intelligence. Interactive is my favorite space hands down. I think it’s where we most feel the inklings of AI as an entirely new medium that isn’t just skeumorphic and referring to the past. I expect that what we see in 2026 and describe as “interactive” is hardly scratching the surface of what’s to come.

It’s worth noting there’s a lot of other sub-categories I wanted to put under Production, including XR, Immersive, Events, and especially Video Games. But we chose to omit these because of space and because they either deserve their own market map with their own pointed categories or because they don’t fit under our loose definition of the production pipeline. We’re working on it.

Inference

The infrastructure layer that everything else runs on: the GPUs, model-serving platforms, and APIs that turn a creative tool’s request into an actual generation.

You probably think I’m crazy not to put companies like Nvidia and AMD here, but we had to draw the line somewhere, right? These companies, especially Fal, serve specifically inference for creative production thus making them a vital part of our ecosystem. It’s worth noting that as inference has become more competitive, widespread, and easier to apply via vibe coding, it has completely changed the landscape in production tools. Because of tools like Fal, Baseten, GMI, etc. vibe coders and coders have been able to more easily build their own aggregators, canvases and tools for production.

Indeed, some of the best creators in the Machine Cinema network build their own tools for themselves and for their clients. If I’m a writer-director-editor I have completely different feature needs and workflows from an animator-writer-actor. Maybe the Cambrian explosion has barely begun.

Post-Production

Where raw output becomes finished work: upscaling, color, visual effects, editing, and the legal scaffolding that lets it ship.

Arguably, post-production are some of the most technical tools in the business and most useful for professionals. Upscaling and color grading have been relatively quiet categories, not seeing many new entrants. Gen VFX has broadened considerably as the model companies like Google’s Gemini Omni and Runway’s Aleph are competing to allow their users to edit objects, subjects, and landscapes in any given video but also tools like Metaphysic are so niche and powerful that they’re used in major films like Here to de-age Tom Hanks.

Lastly, as AI workflows and tools have gotten more powerful, it inevitably begs the ethical and legal questions. Can I protect my likeness? Can I protect my work? We will likely see further innovation in the legal/likeness category as the tools inevitably get more powerful.

Distribution

At the end of the day, the most important problems in the space are distribution and monetization. With people spending all these tokens, how are they making money off of it? AI is expensive!

YouTube stands to win the most as the largest video player in the world, but because of Google’s policy on monetization of AI content and given the still unproven and nascent nature of the space, of course there’s space for new players. AI-specific streamers are still very new and don’t command the kind of numbers that YouTube guarantees (at a certain size) but have an opportunity to nurture the new crop of AI filmmakers whose numbers will inevitably explode. There’s also some very unique strategies of putting AI videos on televisions or even airplanes.

Microdramas, which have also captured Hollywood and dubbed “verticals”, provide both distribution and monetization rolled up in one, but the financials might not make sense yet. Microdramas are notorious for their low budgets and smut-style content which fuels addiction and subscription dollars. Will this genre evolve into more sophisticated yet equally addictive storylines? We shall see.

Conclusion #2

We hope these few notes have added color to an otherwise intimidating wall of logos. Feel free to let us know if you’d like to add any commentary, missed anything, completely disagree, or just want to show us some love. We’re open to all feedback. Thank you again for reading. Happy generating!

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Last Week… Today. With Another One On The Way.

A week late but worth it. The community spent Saturday morning in a brand-new two-hour live text chat picking apart trolling, theft, and what art is actually for. Meanwhile, the “clean model” question finally got asked out loud, Jay Judah dropped an entry that broke the room, and Veo 4 (or Gemini Omni, depending who you ask) is starting to leak ahead of Google I/O.

The week the community stopped reacting and started arguing on purpose. Let’s get into it.

Machine Cinema Channels

🏆 COMMUNITY WINS

A big week for ships and signings.

• Charlie Fink — Dropped Linda’s Last Podcast, a 21-episode original microdrama, after three years teaching cinematic AI at Chapman and ASU.

• Chris White — BACKROOMS hits theaters everywhere May 29, with a sold-out Burbank pop-up walkthrough this week.

• Jagger — Signed for commercial representation with MC&V (Australia) on the back of a Claude spec ad.

Honorable mention to Jay Judah’s Impossible Commute entry, which got the rare “better than 95% of festival films” call-out from a notoriously hard grader.

Share

THIS WEEK IN ‘COMMUNITY’

GenChat Hour Is Live, And It Didn't Stay Polite

Minh launched a new weekly experiment: a one-hour live text chat in Basecamp, every Saturday at 11a PST. After a poll (Saturdays won 12 to 1), the first session ran for two hours instead of one, with members trading takes on the now-famous "real Monet posted as AI" trolling moment and the broader question of how to move the discourse. Danny Ratcliff argued for inclusive education; Carlos Cisneros pitched a moderated panel of pro-AI and anti-AI filmmakers; Brian Wankum quoted from the Pessimists Archive. Minh's read: "Most moderators play it safe. People don't like to disagree in public. It's always 'I agree and also.'"

[BASECAMP]

The "Clean Model" Question Nobody Wants to Answer

Jennifer Green asked the question that’s been lurking for months: are there any clean models left? She’d just clocked the Asteria/LTX partnership and noticed Firefly now pipes in Runway Gen-4, Veo3, Pika, Luma, and Kling 2.5 Turbo. Fred Grinstein noted that aggregator indemnification is more about due diligence than provenance, comparing it to E&O insurance. Niklas Fissel cited Adobe’s “trained only on data we own” claim; OverJK responded that “the data users uploaded while pinky swearing they own it” doesn’t quite count. Jennifer’s landing point: “Commercially safe is not the same thing as ethically sourced.”

[BASECAMP]

Jay Judah Drops, And The Room Loses It

Jay shared his Impossible Commute entry — fully generated, four reference pics from a single corporate headshot, character sheet from GPT Image 2. OverJK's reaction: "I went to a major AI film festival last month, and honestly, this is better than 95% of the films I watched there." Within an hour, OverJK had turned Jay into a meme series. Jay, gracious: "Plz don't steal my body." The thread doubled as a casual workflow breakdown — four images, same outfit, one session.

[BASECAMP]

SPONSORED

Try Flik Free

Veo 4 Or Gemini Omni? All Eyes On I/O

OverJK kicked it off mid-week — he’s been hearing the new Google model called “Gemini V” or “Gemini Video” more than Veo 4, with people pitching it as “Seedance 2.0 Pro with Kling 4K resolution.” Sunday morning, Eli dropped a more detailed leak: native multi-camera, synchronized dialogue, ambient sound, contextual scoring — a full multimodal system. LOGIK, who’s been deep in Flow, said the trajectory tracks. Minh stayed skeptical on Google’s execution pace, but conceded: “Their ad business is near infinite. And they have YouTube and TPUs.”

[BASECAMP / WELOVEROBOTS]

Is The Theater The Next Pop-Up?

Chris White announced a BACKROOMS walkthrough experience in Burbank — fully booked before most people saw the post. Michelle Gold ran with it: genre releases should be leasing dead suburban malls and building Stranger Things–style in-person experiences alongside the theatrical drop. Cameron took it further: with AI democratizing spectacle, in-person + digital hybrid entertainment might be what justifies a $20 ticket. The pop-up is the new poster.

[BASECAMP]

🔗 This Week’s Link Drops

Industry News

• Kuaishou plans to spin off Kling at $20B valuation — Reuters / TradingView

• Anthropic catches OpenAI in valuation — Techmeme

• Demi Moore on AI in Hollywood at Cannes — Variety

• “The AI Filmmaker Outrunning Hollywood” — profile of Charlie Curran — Pirate Wires

Tools & Tech

• Runway’s Big Pitch winners — Runway

• Krea + Recraft taking on Midjourney on style

AI Filmmaking

• Jon Stray — 3 music videos / 3 tools / 3 months case study

Machine Cinema Events

What Are You Actually Arguing For?

The week’s throughline wasn’t a tool or a release — it was the realization that “pro” and “anti” aren’t the only two options, and that nobody’s hosting the conversation that lives in the middle. If you’ve got a take you’ve been sitting on, bring it to GenChat next Saturday. The mics are open. The mod has no fear.

See you next Sunday (on time, this time).

— Machine Cinema Newsletter.

Avatar 3 opening credits: “No Generative AI was used in the making of this movie.”

A24’s Heretic end credits: “No generative AI was used in the making of this film.”

Apple TV+ series Pluribus end credits: “This show was made by humans.”

But I don’t think it’s enough and it’s also disingenuous.

The futility of AI Labels in an AI-infused world

Let’s be honest, only a few people in the world are refusing to use ChatGPT altogether. There are people who, on ethical and environmental grounds, refuse to use ChatGPT, but even if you avoid it, when you Google anything, AI Overview is getting shoved down your throat. GenAI might be possible to dodge, but AI as a whole? Impossible. Any usage of a tech product has some degree of AI. It’s inescapable. In other words, people all over the world are using AI, including Hollywood.

Former Hollywood Reporter editor-in-chief Janice Min (now CEO of The Ankler) recently said: “The thing with AI right now in Hollywood: Everyone’s lying just a little bit. Studios are lying about how much they’re using it… I dare you to find a screenwriter who is staring at a blank page and not talking to Claude or ChatGPT at the same time.” Execs and assistants are using it appraise scripts. Producers are using it to plan. The list goes on and on.

The question is how much is everyone lying in Hollywood?

And for that matter, even if they were telling the truth, do we even have an agreed upon definition of what is AI usage? Or GenAI usage? Or algorithms? Or…? It’s bewildering.

This throws a big wrench into this idea that “No AI was used in the making of...” is a useful disclosure. AI is soon going to be like the air we breathe or water we drink. (insert cough cough joke here). This is why I think these AI labels are futile.

The Anti-AI folks will not find it to be enough (as more and more people use it and understand how it works) because everyone is lying and the pro-AI folks will want their work to be taken seriously. And knowing what degree of AI was used will become noise, easier to ignore, and therefore easier to appraise what we like about a work of art irrespective of how much AI was used.

A spectrum of Zero AI to 100% AI

AI usage should be judged on a spectrum. This will help artists and audiences clarify what is palatable and demystify the work. People will be able to judge the work with more transparency. The lack of clarity and transparency, which is where we are today with agencies, studios, and artists hiding to what degree they’re using AI, is causing suspicion and feeding into narratives about why use of AI is nefarious and ethically objectionable.

This goes both ways. What the Anti-AI crowd eventually has to grapple with is the reality that everyone is using AI to some degree. And the pro-AI crowd has to cop to how much they’re using it and to what degree it has infiltrated their “sacred” creative workflows.

What I think has to happen is an over-investment in behind the scenes footage, tracking, and recording and excruciatingly clear declaration of where AI was used in the process. It will be like the new DVD bonus feature. People will need to declare what percentage of AI was used in the making of a project from pre-production/development to production, and post-production. After all, if execs and their assistants are using AI to read scripts and judge creative work and streamers are using algorithms to determine the viability of projects, shouldn’t all the cards be on the table?

Imagine a world where every exec, cast, and crew member signs off on their AI usage before release, and to what degree and which tools. Disclosure as a credit, not a disclaimer.

What this will result in is a spectrum, even down to percentage of AI used. You might even be hearing one day someone saying “oh that 80% AI-made film was awesome!” Or “god. That 20% AI-made film was sh*t! Why even use AI at all?” Or: “that 0% AI film was awesome. Love pure human magic.”

However it manifests, that clarity will breed a new way to appreciate art.

A new world of transparency aka nutrition labels

From the 1900s to the 1990s, nutritional labels went from preventative measure to a requirement. By the 2010’s amidst the wellness and longevity craze, they’re background noise behind health marketing. Protein! Amino acids! Antioxidants! Vitamins!

I would not be surprised if the level of granularity that goes into nutritional labels today is where AI labeling ends up. Teams could even clarify, in detail, what tools they used to get the job done. Eventually, it will be so obvious that AI was used that it will be a boring detail. It’s like great, you got the basic disclosures right now what did you do with it? Is it another cliche-filled Hollywood flick? Or did AI help you be original? And am I getting an interactive immersive ecosystem along with this? Or merch? Or whatever buzzword or marketing gimmick happens to be hitting that season.

The thing is nutritional labels have the FDA, to regulate and enforce. The MPAA rates films for sex and violence. The FTC polices advertising claims. But we are currently dealing with an avalanche of AI usage across multiple fields including Hollywood, gaming, XR, immersive, music, and more. There’s no singular regulatory body. There’s unions, studios, agencies, corporations, etc. How will all these organizations agree on a standard? I think it’s unlikely. It’ll likely just have to be emergent for lack of coordination.

In AI, this isn’t a new idea. Researchers at Stanford and groups working on AI provenance (Starling, the C2PA standard, the AI for Documentary coalition) have been pushing for content authenticity frameworks for years. And the Academy of Motion Pictures Arts and Sciences specified new rules on acting and writing, but not a complete ban on using AI in things like VFX, etc.:

“”In the Writing categories, the rules codify that screenplays must be human-authored to be eligible.“

“Additionally, in the Acting category, only roles credited in the film’s legal billing and demonstrably performed by humans with their consent will be considered eligible.“

For the Anti-AI folks, who think the Academy is making a blanket ban on AI, reading the fine print is a rude awakening. This is precisely the level of granularity that is inevitable and needed.

AI isn’t going away and we still don’t have the frameworks to appreciate or appraise it. We’re in a honeymoon/demonization era of it but this won’t last. Everyday we’ll get another famous person or organization embracing it and another day another famous person or organization will ban or decry it. But eventually, it’s all going to be noise and we need to get back to work, no matter much AI you use. Just tell me how much.

Work in Progress is Machine Cinema’s weekly op-ed series exploring ideas at the intersection of AI, creativity, and culture. If you’re interested in contributing, please reach out. Selected submissions will be featured and we will promote the author’s blog or work alongside the piece.

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Sunday Edition

Happy Mother’s Day to all you Mothers! 💕

Sora’s gone. The Hollywood Reporter dropped its annual “Most Powerful in AI” list and the community had thoughts — sharp ones. The Impossible Commute deadline ambushed half of us, so we slopped accordingly. And in the middle of all of it, Minh declared the Slopacolypse has arrived: humans are now performing imperfection just to prove they’re human.

It was a week of eulogies, eye-rolls, and engineering takes. Let’s get into it.

Machine Cinema Channels

🏆 COMMUNITY WINS

A few well-deserved shoutouts from the week:

• Chris White — The Heretiks will show advanced footage at Cannes (per Deadline), and Chris was also featured in The Ankler’s piece on Backrooms, Kane Parsons, and the Curry Barker young-horror wave.

• Alex Naghavi & Ezra Li — Released their new short Le Drip, with Ezra coming in as a Gen:48 winner.

• Sultan — His MIT Reality Hack team was selected for an exclusive international tech accelerator through IBC, with a chance to show their XR + AI + Entertainment project at scale this summer.

Share

THIS WEEK IN ‘COMMUNITY’

The Hollywood Reporter's "Most Powerful in AI" List Got Roasted

THR's 2026 list dropped and the community wasn't having it. Monica suggested a retitle: "the most popular people the Hollywood Reporter is in bed with list." Multiple members questioned specific inclusions — particularly Justine Bateman. One member dropped a real piece of intel: Disney is actively onboarding Runway, with weekly calls for employees on how to use it better. Martin landed the cleanest read: the list isn't journalism, it's a funnel. "These lists serve to cement the likes of THR as 'a place where the minds meet' — an easy onramp for the next event and speaker series."

[BASECAMP]

Leave a comment

Sora's Gone. Long Live the API Sweater.

OpenAI quietly retired Sora, and the chat went into mourning. Wuz Good asked, “Wait is Sora gone?” Brogan confirmed: “Yeah… it got old yellered.” Mike delivered the eulogy of the year: “Sometimes I still catch her scent on her API sweater. For a second, I forget she’s gone. I look toward the door like she might walk in… It’s stupid, I know… but it feels like part of her is still here.” Later he conceded: “Sora 2 is still the best slop generator. RIP my love.”

[BASECAMP]

The Vibe Coding Reckoning

A study claiming "real engineers don't vibe code" sparked the deepest WeLoveRobots thread of the week. Mike introduced "Dead Codes Society" — the graveyard of zombie code that vibe-coded projects accumulate, the kind that breaks everything when you try to delete it. Rahul countered with a working framework: PRD → spec doc → architecturally-prioritized task list → execute. "Just because the AI can code for you doesn't mean you shouldn't have some sort of process." Mike's line was the keeper: "AI didn't kill programming. It killed junior levels and just exposed who actually understands it."

[WELOVEROBOTS]

The Slopacolypse Came Early

Minh shared a thread on humans deliberately performing imperfection to prove they’re not AI — typos, weird angles, verbal fillers — and named it: “AI helps create slop. So now humans have to be sloppy to be seen as humans. Slopacolypse came early.” Mike added: “Human SLOP!” John just nodded: “Wabi Sabi MF’ers!” The bit cuts both ways. We spent two years trying to make AI look like it was made by a person. Now we’re studying how to make people look like they weren’t made by AI.

[BASECAMP]

SPONSORED

Try Flik Free

🔗 This Week’s Link Drops

Industry News

• New Oscars rules for 2027 on AI actors and screenplays — Hollywood Reporter

• Trump’s moves on AI models — NYT

• Lionsgate, AI, and Kathleen Grace — Hollywood Reporter

• Elon Musk: xAI to be dissolved and folded into new SpaceXAI structure — FinanceFeeds

Tools & Tech

• Krea’s new image model (reply-to-tweet for access) — X

• Google’s Flow music partnership — Google

• Mike’s deep frame-rate / FFmpeg pipeline writeup — LinkedIn

AI Filmmaking

• How AI is transforming China’s entertainment industry (microdramas) — NYT

• India’s AI filmmaking landscape — Hollywood Reporter

• Backrooms, Kane Parsons, Curry Barker, and the young-horror wave — The Ankler

• AI music is flooding streaming services — but who wants it? — The Verge

• Chasing AI’s Creative Ghost — Kai (Substack)

Worth Watching/Resources

• Jack Clark: “End-to-end automated AI R&D by 2028” — Substack

• China’s “annotation army” — Sixth Tone

• China powers AI boom with undersea data centers — Scientific American

Machine Cinema Events

LEAVE A LITTLE WABI SABI…

This week we mourned a model, mocked a list, and made peace with the idea that the future of authenticity might involve performing imperfection on purpose. The community shipped fast, debated harder, and somehow turned a missed deadline into a meme. Whatever you're working on next — make it weird, make it yours, and don't be afraid to leave a little wabi sabi in the file.

See you next week!

— Machine Cinema Newsletter

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Weekly Edition

Last week Anthropic’s valuation chatter hit twelve digits while its best model quietly got throttled. The week an upscaler turned 19th-century African Americans white and the community refused to let it slide. The week Cannes prep dragged us all into a debate about whether actors are still necessary — and someone wrote a satirical reply so good it’s getting screenshotted across DMs.

Plus: a benchmark fight, a Women’s Gen Jam love-fest, and Theo’s prompt-to-scene moment.

Let’s get into it.

Machine Cinema Channels

🏆 COMMUNITY WINS

A strong week for festival energy.

Brogan Wassell, Andrew “Squid” Montanez, JKR, Jon Stray — all dropped AIFF Runway submissions. Andrew’s The Crunch Job even comes with a Santa Monica workshop on how he made it in two days.

Mike OverJK — has a film screening at AIFA Cannes on May 21 at the Majestic.

Larry Layfield — released Echoes of Havana, made with Seedance 2.0.

Christine Carr, Laura Tornga, Irina Logra — organized the inaugural Women’s Gen Jam, by all accounts a hit.

Share

THIS WEEK IN ‘COMMUNITY’

Will Actors Always Be Needed?

Guillaume opened a Cannes-prep thread asking for arguments on both sides for a French TV segment. The replies were the most thoughtful exchange of the week. Miriam delivered a long, candid list — stunt safety, retake economics, and a frank note on Hollywood's tolerance of difficult talent. Kelsey, a lawyer peer-reviewed on AI actors since 2019, offered to run a webinar on the legal stakes. Aaron Bhugobaun reframed it: "the technology is not going away, and this is bigger than just actors...it's every society and industry." Then Jennifer wrote a satirical reply imagining a future chat room of agentic directors debating whether humans are still needed — before turning sincere: nothing replaces the messy work of human collaboration.

[BASECAMP]

Leave a comment

The $900B Compute Crunch

Minh laid out a thesis that ate the WeLoveRobots channel for two days: Dario's blog and the Pentagon back-and-forth pulled enough enterprise demand toward Claude that Anthropic ran into compute constraints, throttled Opus 4.7, downgraded users to 4.6, and tightened Pro-plan token limits — meanwhile OpenAI's growth slowed as users defected. Aaron pushed back on the Zerohedge sourcing ("I generally agree with your premise"), Minh conceded the point, and then Fred dropped the CNBC story about Anthropic weighing a $900B raise. Aaron later flagged a follow-up: Pentagon says Anthropic is still blacklisted, but Mythos is a separate issue. The subsidization era, as Minh put it, is hitting its first discount constraint.

[WELOVEROBOTS]

The Upscaler That Whitewashed History

Sutu tried to upscale a 19th-century image of African Americans for a non-profit working on abolitionist-movement education. OpenArt's tool turned the subjects white. Sutu posted the side-by-side, referenced Coded Bias, and the channel responded fast — Topaz's Monji jumped in offering DMs to surface model issues, Fred called it "terrible output" and credited Sutu for making noise, and Sutu eventually confirmed Topaz delivered a better result. A reminder that bias audits are not optional infrastructure.

[BASECAMP]

SPONSORED

Try Flik Free

Benchmark Beef: FilmArena vs. The Methodology Police

Mehul launched FilmArena.ai with a 10K-participant, 1K-video benchmark putting Alibaba’s HappyHorse just below Seedance and PixVerse C1. Mike OverJK called BS — one iteration per user, English-only prompts, no JSON or Chinese, no production-context testing. The exchange got heated but landed in a good place: Mike asked for the conversation to move to DMs to design a better benchmark, and Mehul welcomed it. “It’s not 2025 vibes anymore. We’re replacing traditional pipelines. Let’s do it properly.”

[BASECAMP]

The Women’s Gen Jam Glow-Up

Caroline, Stina, Michelle, and Rachel all came back to the channel to praise Sunday’s first Women’s Gen Jam, organized by Christine, Laura, and Irina with Women in Tech sponsoring. Stina liked it enough to make a TikTok about it. Laura confirmed more women-led events are coming, open to the full community.

[BASECAMP]

EVENTS, SCREENINGS, FESTIVALS, & MEETUPS

AIFA Cannes — Mike OverJK has a film screening, and members are gathering at Marché from May 12 onward. May 21, the Majestic.

Cannes Lions Machine Cinema gathering — June, via Christopher Carew →luma.com/yoxf5s0o

AI on the Lot — San Francisco — Use code GENJAMMER for 25% off →aionthelot.com

Machine Cinema Events

🔗 This Week’s Link Drops

Industry News

• OpenAI misses key revenue and user targets ahead of IPO sprint — WSJ

• Anthropic weighs raising at $900B valuation — CNBC

• Pentagon: Anthropic still blacklisted, but Mythos is a separate issue — via Aaron

• Lionsgate’s Kathleen Grace on AI, John Wick, and Twilight — Hollywood Reporter

• Inside the World-Conquering Rise of the Micro-Drama — The New Yorker

Tools & Tech

• Topaz Astra 2 launch — prompts on upscale, credits available via Monji — Topaz Labs

• ElevenLabs Music platform — ElevenLabs

• Runway $500-credits API hackathon — Runway

• Where the Goblins Came From — OpenAI

AI Filmmaking

• Brogan Wassell — AIFF submission — YouTube

• Andrew Montanez — The Crunch Job — YouTube

• Larry Layfield — Echoes of Havana — YouTube

• Jon Stray — Resistance series — YouTube

Worth Watching/Resources

• Coded Bias trailer (2020 doc, freshly relevant) — YouTube

MAKE NOISE

Cannes is around the corner, valuations are eating headlines, and benchmarks are about to get a lot more rigorous because of conversations happening right here. The work this week wasn’t only the films — it was also the calling-out, the methodology fights, and the satirical replies that landed harder than essays.

Make something. Show your work. Argue in good faith.

See you next week!

— Machine Cinema Newsletter

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Sunday Edition

OpenAI dropped GPT Image 2 mid-week and the group chat immediately split into two camps: people who think the magic is back, and people who think the edge is already gone. Meanwhile, Anthropic quietly cut Claude Code access for new $20/month users without telling anyone, DeepSeek shipped V4 with another round of architecture tricks, and Tim Cook handed Apple to John Ternus on his way to executive chairman.

Also this week: Seedance IP Bingo got a leaderboard, Adobe got a strongly-worded review, and Gen Talks finally moved to Substack Live and tripled its audience.

Let’s get into it.

Machine Cinema Channels

🏆 COMMUNITY WINS

A lot of MC folks out in the world this week.

A loaded week for member work and recognition.

• Danny Ratcliff — Named Webby Honoree for short film Goodbye at Webbys 30

• Mehul Agarwal — Launched filmarena.ai, a benchmarking arena for image and video models

• Leonardo — Published Why I’m Bullish on OpenAI on aicoe.fit

• Francesco D’Isa — New piece for The Bunker on language models, the brain, and fMRI-based decoding

• Michelle Gold — Released a new music video built end-to-end in Flik with Seedance, with Suno integration via in-Flik agent loops

Share

SPONSORED

The Fairground Reimagined Competition is live

Pick something from the public domain. Reimagine it as a 3+ minute AI film. That’s the brief from our friends at Fairground, and there’s over $10K on the line, plus distribution on their FAST channel and connected TV partners for the winners.

Submissions close May 1. Winners announced May 13.

Enter the competition

THIS WEEK IN ‘COMMUNITY’

GPT Image 2 Launches and Eats the Image Gen Conversation

OpenAI’s gpt-image-2 dropped Tuesday during a live stream and the chat lit up. Mehul previewed it the day before, citing 1,000-person internal testing where it “easily beat” Nano Banana. Once it was live, Mike pushed back with a clear ranking: GPT Image 2 ahead of Nano Banana Pro, NB Pro ahead of NB2. Squid liked the multi-option generations. Minh hit rate limits in ten minutes (”GPU constrained”) and noted Image2 keeps getting stuck in loops referencing prior images in chat — meaning serious work still wants node-based control. The takeaway, in Minh’s words: “feels like the magic is back, but the edge is gone.”

[BASECAMP]

Where Are All the Chinese Image Models?

Inside the Image2 thread, Minh raised a sharper question: Chinese AI absolutely dominates video and is closing in on coding and LLMs — but image gen? Crickets. Mike and Louie counterpunched with names — Qwen, Ernie, Adam v2, Kolors, GLM, Mystic Labs — and Louie made the case Ernie beats Nano Banana for photorealism and it’s free. The arena rankings still show very few Chinese models near the top, but the room mostly agreed those benchmarks are “who-made-them” suspect anyway. Flux got dragged (”Nothing is as bad as Flux”), then defended for speed and customization. Healthy chaos.

[BASECAMP]

Anthropic Quietly Pulls Claude Code From New $20 Users

A trending X dumpster fire pulled the chat in: Anthropic removed Claude Code access for new $20/month subscribers with no announcement, then claimed it was a “small 2% test” — except they updated the docs to match and users hit it consistently. Leonardo’s piece Why I’m Bullish on OpenAI dropped the same day and didn’t pull punches: Anthropic’s mindshare is “out of sync” with OpenAI’s latest releases and the open-weight momentum. Minh wondered aloud if the real story is a compute crunch: “i do think one of the problems Anthropic might be facing is a compute problem and it’s possible Dwarkesh was right.”

[WELOVEROBOTS]

Leave a comment

DeepSeek V4 Lands — But Does It Matter in the Agents Era?

DeepSeek V4-Pro shipped with two new tricks the room cared about: sparse attention and manifold-constrained hyper-connections. Leonardo argued the real impact won’t be the benchmarks but other labs (Moonshot, Zhipu) adopting the methods. Minh wasn’t sure: when R1 dropped, reasoning was the revelation — but we’re in an agents era now, and “literally anyone could build an agent.” Early arena rankings put V4 behind Kimi and GLM. The most interesting downstream theory came from Minh: if Huawei can keep DeepSeek running during ChatGPT or Claude downtime, the shock isn’t to model leaderboards — it’s to Nvidia.

[WELOVEROBOTS]

The Markiplier Model and the End of the Studio Stack

Michelle posted a Tubefilter piece on Markiplier’s Iron Lung YouTube release and made a sharp call: this — not the studios — is the portal to the future of entertainment. She drew the line back to Twin Peaks: The Return on Showtime: the audience pre-paid because they loved Lynch and they loved Twin Peaks. Same playbook here, fan-funded and creator-owned. The thread connected to Ant’s earlier distribution question (a mini-studio rep told him post-YouTube films struggle to chain into other distribution) and to news that MrBeast is plotting a move into AI-native entertainment. Brogan summed up the future: “Mr Beast: ‘first person who can make an AI short film that makes someone in the audience cry gets 1 million dollars!’”

[BASECAMP]

What It Actually Costs to Ship with Seedance 2.0

While Seedance IP Bingo was being played over in Basecamp, Cheikh Rôõney posted a full teardown of his Seedance 2.0 action short in WeLoveRobots — the receipts version. His pipeline: Freepik for assets, Gemini Pro for the script layout, NanoBanana Pro for image staging, ChatGPT for the timeline breakdown, Seedance for animation and dialogue, Premiere for the edit. Cost: about $45. His satisfaction on dialogue and acting: 70%, which he called “already a strong performance.” When a member asked how to budget a wolf short film before diving in, Cheikh walked him through the whole workflow.

His sign-off: “The real question is no longer ‘Can AI do this?’ but ‘What can we create with it?’”

[WELOVEROBOTS]

SPONSORED

Try Flik Free

Festivals and Screenings…

• Altera Festival — Closing 4/30

• Austin AI Film Fest — opens 4/15

• TED AI Vienna — submissions open

• Runway Big Pitch Contest

• AI International Film Festival extended its submission deadline (use code machinecinema at filmfreeway.com/aiinternationalfilmfestival). Around 15 wins for MC folks already.

EVENTS & MEETUPS

Cannes Lions is on the community radar — a few members are heading.

Machine Cinema Events

🔗 This Week’s Link Drops

Industry News

• Tim Cook to Become Apple Executive Chairman; John Ternus to Become CEO — Apple Newsroom

• MrBeast Plots Move Into AI-Native Entertainment — Business Insider

• YouTube Rolls Out AI Deepfake Detection Tool — The Hollywood Reporter

• Sandra Bullock Urges Embracing AI Trailers for Practical Magic 2 — The Movie Blog

• Vercel Confirms Breach — AI COE

• Roblox AI Assistant Gets Agentic Tools — TechCrunch

• Google DeepMind Hires Philosopher — Perplexity

Tools & Tech

• GPT Images 2.0 Explained — AI COE (Leonardo)

• filmarena.ai — Mehul Agarwal’s image/video model benchmarking arena

• DeepSeek V4-Pro Technical Paper — Hugging Face

• free-claude-code — Use OpenRouter or local models with Claude Code

• Flipbook — Worth a look

AI Filmmaking

• Markiplier’s Iron Lung and the YouTube Indie Model — Tubefilter

• Beyond Here (Ant Neely, w/ Henry Lloyd-Hughes)

• The Ritual — Built Entirely With AI — Javed Hussain Shiekh

• Leonardo’s New Video

• Esmeralda — Bella Ciao — Leonid Godik

Worth Watching/Resources

• Why I’m Bullish on OpenAI — Leonardo, AI COE

• AI Can Now Read Minds (Almost) — Francesco D’Isa, The Bunker

• Albert Cheng on AI in Filmmaking — LinkedIn

BUILD WHILE THEY RESHUFFLE

This week the giants reshuffled. Tim handed off Apple. Anthropic locked a door. OpenAI pulled the rug out from under everyone’s image gen workflow. DeepSeek shipped a model in a week the West will spend a quarter analyzing.

While they squabbled, members shipped: a Webby honoree, a benchmarking arena, a music video built in agents, three articles, and a Gen Talks that tripled its audience the day it switched platforms. As Minh put it this week — Machine Cinema is a Lego set, and we have to fire every brick carefully in the kiln.

So fire yours. Make something this week.

See you next Sunday.

— Machine Cinema Newsletter

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Sunday Edition

It was a week of big moves and bigger questions. Claude shipped something that knocked 20% off Figma's stock in 20 minutes. Disney/Marvel laid off 1,000 people — mostly women — on their visual team, kicking off a long-overdue conversation about who's actually getting displaced. SeeDance 2.0 kept generating videos that nobody technically owns. Doug Liman started shooting an AI-first thriller with Gal Gadot. And Gossip Goblin gave us the cleanest AI-first IP drop we've seen yet.

Let’s get into it.

Machine Cinema Channels

THIS WEEK IN ‘COMMUNITY’

Anthropic Just Ate Figma's Lunch (in 20 Minutes)

Claude shipped a design-share feature this week, and Figma's stock dropped 20% in 20 minutes. Minh called it exactly: "Anthropic eating everyone alive." The thread turned into a bigger question — is Anthropic actually building the console, not an assistant? Figma acquired Weavy, Anthropic's CPO left Figma's board, and Minh predicted a node-based workflow natively in Claude within six months.

[BASECAMP]

SeeDance IP Bingo, Starring: A Copyright Lawyer

Mike OverJK played what he christened "SeeDance IP Bingo" — generating scenes about AI breaking copyright using SeeDance 2.0. The highlight: a 70mm-lens lawyer, full confidence, delivering the line "Let me stop you for a minute. A prompt is not authorship." Mike's kicker on SeeDance's fine print: "I love how SeeDance added 'You don't own the output.'" It landed the same week Rob Tercek's essay "Hollywood Is a Plagiarism Machine" was making rounds — Jennifer called it "definitely worth a read" for its distinction between mimicry and visual quoting.

[BASECAMP]

Disney Layoffs and the Women-in-AI Conversation

Disney/Marvel laid off 1,000 people on their visual team — mostly women, per Stina. Monica opened the floor: "Are we ready to have the conversation about women in AI?" What followed was the most substantive debate of the week. Brogan argued that PM, HR, and creative-management roles (disproportionately held by women) are next on the automation chopping block. Stina pushed back, citing McKinsey data showing men still hold 80% of C-suite seats in media/entertainment, and called the framing itself a choice:

“When the guy building the tools tells you whose jobs they'll take, that's not a prediction. That's a plan.”

[BASECAMP]

Leave a comment

Hollywood's Split Screen: Liman's AI Thriller vs. Gossip Goblin

Two moments captured the bifurcation this week. On one side: Doug Liman's AI-first Bitcoin/Satoshi thriller starring Gal Gadot and Casey Affleck, with some MC folks reportedly on the project. Fred framed the stakes: "I think this film if successful is another Ben Affleck type moment for Hollywood adoption." On the other: Zack London's cyberpunk-noir Gossip Goblin dropped to near-universal praise for world-building — Bay called it "incredible" — with a merch store already live and Fred floating a GenTalks ask.

[BASECAMP]

What It Actually Costs to Ship with SeeDance 2.0

While SeeDance IP Bingo was being played over in Basecamp, Cheikh Rôõney posted a full teardown of his SeeDance 2.0 action short in WeLoveRobots — the receipts version. His pipeline: Freepik for assets, Gemini Pro for the script layout, NanoBanana Pro for image staging, ChatGPT for the timeline breakdown, SeeDance for animation and dialogue, Premiere for the edit. Cost: about $45. His satisfaction on dialogue and acting: 70%, which he called “already a strong performance.” When a member asked how to budget a wolf short film before diving in, Cheikh walked him through the whole workflow. His sign-off:

“The real question is no longer ‘Can AI do this?’ but ‘What can we create with it?’”

[WELOVEROBOTS]

FlikAI Crashes GenTalks, Community Dives In

FlikAI was the GenTalks feature on Wednesday, and the community dove in hard. Minh dropped the FLIKMACHINE10 redeem code; Michelle, Bay, and others immediately started testing the agentic workflow and posting results. Bay got hooked: "NGL. I'm kind of flying with the Flik tool." The deeper question was about the interface layer — Kenny announced he's "crossed over into MCP land," and Minh mused that we may be one or two model updates away from TikTok-filter-style prepackaged creative flows hitting AI video at scale.

[BASECAMP]

🏆 COMMUNITY WINS

A lot of MC folks out in the world this week.

• Monica Monique, Brogan Wassell & Bernie Su — Speaking at the Global AI Cinema Summit

• Ia Ayame Robinson & Jason H. — Packed-house panel at NVIDIA GTC

• Ari Kuschnir — Debuted a three-screen TED-style installation on “what makes us human”

• Lisa Russell — AI Film Radar launching via Google for Startups (ArtsEnvoy.ai/radar)

• Leonardo — First peer-reviewed publication accepted (images-to-slides, ACM proceedings)

• Leonid Krykhtin — Released an updated, fully interactive AI filmmaking workflow diagram

Share

Festivals and Screenings…

Altera Festival — An Alt Era of Storytelling

Altera Festival submissions are closing soon! 🎟️ Our community gets 50% off submission fees this week only, code Machine50 → Submit via FilmFreeway

Story-first work made with emerging tech tools.

Accepted formats:

• Narrative Short

• Documentary Short

• Experimental / Hybrid

• Commercial / Branded / Trailer

• Music Video

• Immersive / XR / Experience

• Vertical (9:16 native)

WORKS-IN-PROGRESS ARE ELIGIBLE, so if your project is close but not finished... ship it anyway!

Also open for submissions:

• Austin AI Film Fest — opens 4/15

• TED AI Vienna — submissions open

• Runway Big Pitch Contest

Screening: AM I? - A film exploring our AI systems & consciousness📍 Sound City Studios 📅 Tuesday evening (4/21) → RSVP

EVENTS & MEETUPS

Alibaba Cloud SME AI Growth Day Vietnam 2026 📍 Ho Chi Minh City 📅 April 22 → Sign up

Cannes AI meetup →RSVP

Machine Cinema Events

🔗 This Week’s Link Drops

Industry News

• Disney and Marvel lay off 1,000 on visual team — Yahoo Finance

• Doug Liman’s AI movie with Gal Gadot & Casey Affleck — The Wrap

• Hollywood Is a Plagiarism Machine — Rob Tercek / Substack

• Why Art Must Be Unpredictable Against Algorithm Optimization — Francesco D’Isa / Domus

• Reese Witherspoon Embraces AI, Encourages Women to join her — Variety

Tools & Tech

• Midjourney v8.1 release — X / Midjourney

• Qwen 3.6 open weights — run free locally — AICOE

• Tencent HY-World 2.0 (open-source world model) — X

• Alibaba Happy Oyster world model — happyoyster.cn

• Automating Workflows with Claude Cowork — AI Tidbits / Rahul

AI Filmmaking

• Gossip Goblin cyberpunk short — Zack London / YouTube

• Gossip Goblin merch — gossipgoblin.studio

• The Patchwright cyberpunk short — YouTube

• Perfect Toys opening scene — Christian Luoma / X

• Ari Kuschnir’s TED-style installation — Instagram

Worth Watching/Resources

• Steven Soderbergh on AI — Variety / X

• AI Film Radar waitlist — Lisa Russell’s new venture

SHIP IT ROUGH.

We’re at an odd inflection point. Hollywood is laying off artists while AI filmmakers are building entire IPs from their couches. The platforms are swallowing each other whole, and the copyright lawyers (synthetic and otherwise) are catching up. Somewhere in that noise, the work that actually matters keeps getting made.

Make something this week. Ship it rough.

See you next Sunday.

— Machine Cinema Newsletter

Too busy to read? Listen to the pod instead.

Sunday Edition

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

This was a week of contradictions. Machine Cinema became an official Anthropic partner while YouTube kept wiping out AI filmmakers’ channels. A hyped GitHub project got 16,000 stars before anyone noticed the benchmarks were faked. And the community gathered in real-time to watch astronauts fall from space—because sometimes you just need to feel something that isn’t about algorithms.

It was messy, exciting, and very us. Let’s get into it.

Machine Cinema Channels

THIS WEEK IN ‘COMMUNITY’

YouTube's AI Filmmaker Problem Just Got Personal

Multiple members reported getting hit with YouTube’s “inauthentic content” flag this week, and the picture isn’t pretty. Winston Mayo shared that his AI movies were demonetized with no clear explanation, even after appeal. Ant Neely confirmed the same happened to his channel months ago—original feature and short films flagged as “reused content” with no way to identify the offending material. “I tried every angle and got nowhere,” Ant said. “It does feel like some enterprising lawyer could do a class action lawsuit against them.”

But Teemu’s breakdown was the one that stopped the scroll. He laid out what he’s learned: AI-generated narration and dialogue will trigger the flag. Too-similar video structures and thumbnails will trigger it. Appeals go straight to an automatic AI moderator for instant rejection. And if one channel gets flagged, every channel linked to that Google account gets suppressed too. “I worked hard to get to my 13.7K subscribers only to be wiped out with other AI content channels,” he wrote. The takeaway was sobering—YouTube’s moderation sees no difference between an AI film and AI slop, and the triggers aren’t transparent.

[BASECAMP]

Machine Cinema Goes Anthropic—And Brings Friends

Minh casually dropped that Machine Cinema is now an official Anthropic partner. ("We are good at filling out forms," he joked.) But the bigger story unfolded hours later: Kyle Kesterson announced he'd been accepted into Anthropic's $100M Partner Program, which funds and certifies firms to lead real-world Claude implementations. Kyle shared that the program gives early access to models, co-selling support, and is specifically looking for people with media and production experience. Rocky chimed in that his company is also an Anthropic Partner Startup and is studying for the Claude Architect Certification. Suddenly, the Anthropic-to-Machine Cinema pipeline looks very real.

[BASECAMP]

Seedance 2.0 Is Here and People Are Losing It

Gilberto dropped a string of Seedance 2.0 test videos that stopped the chat in its tracks. The results with normal motion were striking enough that Eli Green immediately asked what model was used. Pame wanted to know more about the character. Jay Judah simply said: "Dude. Amazing." Gilberto noted it works great with normal motion but struggles with fast motion—an honest assessment the community appreciated. Meanwhile, Rex announced the Seedance 2.0 Official US Premiere and AI Film Hackathon on April 17 in LA, with free accounts and credits for approved attendees. The hackathon winners will receive the Seedance 2.0 Premiere Film Award.

[BASECAMP]

The Gatekeepers Are Getting Louder—But So Are We

A celebrity’s anti-AI comments in Variety kicked off a multi-day discussion about who gets to define creativity. Chris White didn’t mince words: “Just a bunch of gatekeeping, ill-informed takes from people who like to punch down on others and don’t really know what they’re talking about.” Natasha B. offered a sharper read: “Her saying all this is just letting me know she feels threatened or she wouldn’t even bother with all this commentary.” Matt Szymanowski tried to hold both sides, acknowledging validity on each before landing where most of the community did: “For now let’s just create and make the best shit we can possible.”

The thread deepened when Julio Soler shared a Fortune article claiming 80% of white-collar workers are “quiet-quitting” AI adoption. “Just when I thought the AI-backlash was starting to decrease,” he sighed.

[BASECAMP]

16,000 Stars On GitHub, Zero Integrity

Over in WeLoveRobots, Leonardo did what Leonardo does—cut through the hype. A viral GitHub project promising a breakthrough in AI long-term memory racked up 16,000 stars in a single day. But Leonardo broke it down: the technical implementation was poor, the benchmark scores were achieved through cheating, and the supposed creator turned out to be a crypto guy who hired someone else to do the work. That person documented the cheating, and “Ben forgot to cover his tracks so it was all exposed.” Leonardo’s verdict: the underlying problem—long-term memory for AI agents—is important. The execution was a fraud. A useful reminder that star counts aren’t peer reviewed.

[WELOVEROBOTS]

Leave a comment

🏆 COMMUNITY WINS

A big week for recognition across the community:

• Tani Ikeda — Webby nomination for We Belong Together, a multilingual Know Your Rights initiative that transformed LA bus stops into art installations linking to legal resources in 19 languages.

• Baratunde Thurston — Webby nomination for the AI co-producer behind Life With Machines. His backstory post is worth a read—”a comedic and cautionary tale.”

• Cameron Dejahang — New hybrid campaign executed for Costco Canada.

• Matthieu — Hybrid US TV campaign out for Amazon.

• Kyle Kesterson — Accepted into Anthropic’s $100M Partner Program.

Share

Festivals and Screenings…

Seedance 2.0 Official US Premiere + AI Film Hackathon 📍 YOYWOW Space, Los Angeles 📅 Friday, April 17 → Premiere RSVP | Hackathon RSVP Invite-only, approval required. Free Seedance 2.0 accounts + credits for approved attendees.

MarsAI Festival 📅 Submissions open through May 12 → marsaifestival.com 1-minute films imagining desirable futures in AI.

Altera Festival (fka Atlanta International AI Film Fest) 📍 Atlanta 📅 June 16–17 (during FIFA World Cup gap days) → Submit on FilmFreeway

No Mercy Short Film Screening →Eventbrite Monthly series where AI and traditional films screen side by side.

🔗 This Week’s Link Drops

Industry News

• AI Is Rewiring Bollywood — Reuters

• AI Backlash: 80% of White-Collar Workers “Quiet Quitting” AI — Fortune

• Gen Z and AI Anxiety — Axios

• Britain Woos Anthropic After US Defence Clash — Reuters via FT

• Anthropic Says Latest Model Too Powerful to Release — Business Insider

• Tom’s Hardware: Claude Mythos Claims Debunked — Tom’s Hardware

Tools & Tech

• SyncLabs New SOTA Lip-Sync Model — SyncLabs

• LTX 2.3 Outpainting LoRA (16:9 → 9:16) — HuggingFace

• AI-2027 Forecasting Scenario — AI Forecasting Research

AI Filmmaking

• NYU Film School Partners with Runway — Hollywood Reporter

• YouTube Report: Independent Animation Taking Off — Hollywood Reporter

• Microdramas & Seed Funding in Hollywood — LinkedIn

Worth Watching

• Artemis II at 2,000fps — YouTube

• AI Trust & the Sovereign Wall — a16z / Substack

• UC Berkeley: Bots Will Protect Each Other — Research

YOU CAN’T FLAG THIS.

The platforms don’t have a category for us yet. YouTube flags us. Hollywood dismisses us. But every week this community ships work, shares knowledge, and shows up for each other—even if it’s just to watch a capsule fall from space together. That’s the thing they can’t demonetize.

See you next Sunday.

— Machine Cinema Newsletter

We want to hear from you. Have ideas for the newsletter? Let us know.

Let's Chat

Too busy to read? Listen to the pod instead.

Sunday Edition — Double Issue

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

In AI, a week is a lifetime. Here’s what happened over the last two!
OpenAI killed Sora, Disney scrambled, Seedance dropped (then degraded), and the community spent two straight weeks debating what actually matters in a world where anyone can make anything. Meanwhile, a Machine Cinema member is executive producing an A24 film, two members got Webby nominations, and Minh thinks this will be the biggest year in cinema history. Agree or disagree, the conversations were worth having.

Let’s get into it — double edition.

Machine Cinema Channels

TWO WEEKS IN ‘COMMUNITY’

Sora Is Gone — And Nobody Agrees on Why

The biggest story of the fortnight dropped on March 24: OpenAI is shutting down the Sora consumer app. The community reacted instantly — and kept reacting for days. Mike OverJK, who had just finished a 20-minute film praising Sora, joked that his love letter would become a eulogy. Danny R. suggested he re-title it “Sora — A Retrospective.” The strategic debate ran deeper: Minh argued the shutdown wasn’t about failure but about OpenAI refocusing on enterprise and their upcoming IPO. Fred pushed back — “I’m pretty sure you keep things going when they’re working and shut them down when they’re not.” Francesco wasn’t surprised either, noting Chinese models had surpassed Sora and questioning why OpenAI targeted mass consumers instead of professionals. The Disney angle added another layer, with Jay J. flagging the now-awkward Sora/Disney partnership and Kenny predicting Disney would “turn around and do a deal with Bytedance on Seedream.”

[BASECAMP]

Seedance 2.0: The Alpha Was Better

When Seedance 2.0 finally launched publicly, the reactions were mixed — and Hussein didn’t hold back. “Prompt adherence dropped, camera control is weaker, motion feels less natural,” he wrote, calling out a growing problem: “We commit to clients based on capabilities that later disappear or degrade. That puts all the risk on us while the platforms keep shifting.” It hit a nerve for anyone building a business on these tools.

[BASECAMP]

Trust Is the New Metric — Not Content, Not Attention

A sprawling, multi-day philosophical thread kicked off when one member argued that product quality still matters most. Minh disagreed: “Marketing and distribution matter more than product because it has become so easy to build and clone any product.” MG stepped in with a framework he calls the Human Connection Economy — a four-layer model (Attention → Creation → Experience → Trust) arguing that brands need to stop chasing eyeballs and start forging long-tail relationships. “In an age of diminishing content costs and the flood of impressions we receive daily, content is no longer king,” MG wrote. “Trust is the new metric.” The thread touched on Nvidia’s latest announcement, the defensibility of curation, and why human trust may be the last real moat.

[BASECAMP]

Hannah Einbinder, Del Toro, and the Privilege to Dunk on AI

When Vulture ran a piece featuring Hacks actress Hannah Einbinder criticizing AI, the community lit up. Minh's response was sharp: "Never heard of you until you shat on AI. Congrats on using AI to get famous." Jay offered a more nuanced take — he likes Einbinder and her work, but called the stance "shitty." The conversation evolved into something bigger when Minh pointed out that Del Toro's family won the lottery, enabling his film career. "I think with privilege comes the space to dunk on AI," Minh wrote. Sarah agreed: "For those without access to high budgets, AI is opening the doors to indie filmmakers." MG closed the thread with conviction: "Don't listen to the noise. Keep looking up. The only ones you are accountable to is yourself and your creative vision."

[BASECAMP]

Handmade vs. Machine-Made: The Luxury Bifurcation

A quieter but potent thread emerged around whether cinema is splitting into two tiers. Jennifer laid it out: "Maybe 95% of film, TV, ads, games are 'off the rack'… AI dominates in that market. And then 'handmade' film/TV offerings become the couture market." MG agreed — "totally polarized ends of spectrums, both with different models, hollowing out the middle." Matt brought in Brian Eno's philosophy: "It's the imperfections that we crave… which is why we go back to retro so much." The conversation also touched on new SAG-AFTRA strike threats, with Jennifer asking a question that lingered: if AI-friendly non-union work keeps growing, "how beneficial is SAG membership to 95% of the current members if it prevents them from taking non-union jobs?"

[BASECAMP]

Leave a comment

🏆 COMMUNITY WINS

Big moves from the MC family over the past two weeks:

• Chris White — Executive Producer on Backrooms, produced with Atomic Monster, A24, Chernin Entertainment, and 21 Laps. In theaters May 29th.

• Kimberly Offord — Nominated for a Webby Award in AI Creativity & Storytelling.

• Aaron Bhugobaun — Nominated for a Webby Award for Best Thought Leadership. Also launching an Industry Creative Working Group to push AI conversations forward.

• Diane Laidlaw — Hosting TED AI’s first AI Filmmaking Competition in Vienna this October. 90-second horror shorts, screening in a 100-year-old cinema.

• Junie — Released AURAE, an AI short film exploring the absence of female heroes in Asian mythology. Fred called her a “living legend.”

• Leonid Krykhtin — Collaborating on OMNI Shanghai: “AI EMOTION — 感動,” a regional AI film festival edition (March 28–30).

• Danny Ratcliff — Released an AI spec ad for Land Rover Defender, continuing his “AI is punk” creative thesis.

Share

GenTalks:

MythOS: Building Collaborative AI Story Worlds (Week One)

🎬 RSVP / Watch →

The community buzzed about the MythOS project getting the Gen Talks spotlight, with Joanna noting “lots of Basecampers on the project!” Sasha called the session “gonna be 🔥” ahead of the March 25 call. MythOS represents a growing interest in collaborative, lore-adherent AI storytelling.

Unlocking AI-Driven Content Creation and Distribution (Week Two)

🎬 RSVP / Watch →

Tiffany Lee Joseph (Award-winning AI filmmaker and creative technologist who presented at Nvidia GTC, building AI-native production systems for networks and agencies) + Daniel Roberts (CEO and co-founder of Alchemy streaming platform on Roku and Apple TV, former entrepreneur-in-residence at Sandbox Industries).

Machine Cinema Events This Week…

SFSF Salon: Session 02

📅 Wednesday, April 8 · 7:00 PM CDT
RSVP

GenTalks with Ari Kuschnir & Matt Zien

📅 Wednesday, April 8 · 1:00 PM EDT
RSVP

RSVP to EVENTS

🔗 This Week’s Link Drops

Industry News

• OpenAI discontinues Sora video platform — The Guardian

• OpenAI acquires TBPN — OpenAI

• Social media trial verdict sets major precedent — NYT

• UK government changes position on AI copyright — Screen Daily

• White House AI rollout rift — Axios

• SAG-AFTRA contract bargaining and the “Tilly Tax” — Fortune

• Netflix & Warner Music strike multiyear deal — Reuters

AI Filmmaking & Creative

• Gossip Goblin / Zack London AI films in Hollywood Reporter — THR

• AI-generated singer “Eddie Dalton” hits #1 on iTunes — Showbiz411

• Chinese AI Cinema analysis by Francesco D’Isa — The Bunker

• Sora is dead — The Ankler’s take — The Ankler

Tools & Tech

• Runway introduces multi-shot — Runway

• Google Flow by Google — Google

• Krea Skills — any agent can create — Krea

• Koyal V2.5 launched — Mehul Agarwal

• Buzbee AI alpha is live — Buzbee

• Suno v5.5 voice features — via Kai

• Topaz Starlight Precise 2.5 — Topaz Labs

Worth Watching

• Artemis II moon mission coverage — NASA

• Karpathy’s AI job exposure map — Andrej Karpathy

• Meta brain-to-video research — Meta AI

Competitions & Opportunities

• Future Vision XPRIZE — XPRIZE

• Runway Big Ad Competition — Runway

• AI International Film Festival — early bird deadline — AIFF

WHEN ONE DOOR CLOSES…

Two weeks of noise, signal, grief, and ambition. Sora's gone, but the conversations it sparked aren't — about trust, about craft, about who gets to make things and who gets to gatekeep. The tools will keep changing. The people in this community won't stop building.

See you next Sunday.

— Machine Cinema Newsletter

We want to hear from you. Have ideas for the newsletter? Let us know.

Let's Chat

Too busy to read? Listen to the pod instead.

Sunday Edition

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

A busy week in the channels. The Oscars gave us a soundbite we’ll be quoting for years. The White House dropped a 4-page AI framework that every creator should probably read. And a Warner Bros CEO walked away with nearly a billion dollars while Hollywood keeps pointing the finger at AI.

Meanwhile, the model wars continued in full force—Seedance loyalists are under embargo but somehow still the loudest people in the room. And a sharp thread about vibecoding quietly delivered one of the most useful insights of the year.

Let’s get into it.

Machine Cinema Channels

THIS WEEK IN COMMUNITY

"Animation Is More Than a Prompt." Oh, You Mean Like Pressing Enter?

The Oscars gave this community something to chew on. Nik Kleverov surfaced a moment from the telecast: Will Arnett apparently earned applause—some say a standing ovation—for declaring "animation is more than a prompt." The community's response was swift and merciless. Jay Judah landed the best return:

While Mike OverJK filed it under "BoJack Horseman strikes back."

[BASECAMP]

The Real Reason Hollywood Is Struggling (Hint: It's Not AI)

The Reuters headline said it all: Warner Bros. CEO David Zaslav is set to pocket up to $887 million from the Paramount deal. Jagger called it exactly as it is:

A separate member offered the alternate lede: "Poster boy for failing upward walks away with a cool billion."

The broader thread spiraled into a meditation on billionaires who actually earn their money vs. those who just outlast everyone else—but the original point cut deep.

[BASECAMP]

Seedance 2 Is Here. Sort Of. If You're Embargoed.

The week kicked off with a flurry of questions about Seedance 2 access—specifically how to get in from Canada—and quickly became a broader conversation about where the model stacks up. The consensus: it's in a class of its own. An insider Dreamina CPP and under embargo from posting, said something that framed it better than any benchmark:

Kling was the runner-up pick for I2V quality, with Sora 2 Pro capable of beating it—but flagged for heavier content filtering. Seedance 1.5? Still a dice toss. 2.0 is the one everyone's waiting to talk about openly.

[BASECAMP]

Vibecoding's Real Lesson: Building Is No Longer the Moat

A link to a Flora.ai funding story sparked one of the week's most substantive threads. Minh dropped the observation:

Evan Baily extended it into entertainment: as the volume of content approaches infinity and the average cost of making it approaches zero,

The takeaway landed. IP with a fandom, creators with a following—those are the durable assets now. Product is table stakes.

[BASECAMP]

The White House Drops a 4-Page AI Framework. Creators Should Read It.

Minh shared the administration's National AI Legislative Framework—a concise 4-page document that landed with more nuance than expected. The administration's stated position: training AI on copyrighted material does not violate copyright law—but it will let the courts decide, rather than having Congress intervene. It also proposes federal protections against unauthorized AI-generated replicas of individuals' voice and likeness, with carve-outs for parody, satire, and First Amendment expression. For AI filmmakers navigating the training-data gray zone, this is the clearest picture yet of where the U.S. government stands—and what it's choosing not to settle.

[WE♥️ROBOTS]

Leave a comment

🏆 COMMUNITY WINS

This week’s highlight reel includes a few to celebrate:

• Danny Ratcliff — Created the musical opener for the Escape Awards show, built largely in Seedance with a very fast turnaround. Well-received by the crowd.

• Joanna Popper — Escape Films work featured on the New World Symphony wallscape in Miami Beach as part of Filmgate Interactive Festival. Outdoor screening at scale.

Diane Laidlaw (Afro Futcha) , Maria Haras , and Vienna — Named to Leonardo.ai’sTop 50 Women AI Filmmakers list for 2026. Multiple Machine Cinema members appeared on the list.

Share

Machine Cinema Events This Week…

⚠️ LAST CALL — AI International Film Festival

📅 Submission deadline: Today, Sunday March 22 at midnight (LA time) →Machine Cinema members have taken home wins from this one before. If you’ve been sitting on something, now’s the moment.

Submit Here

CapCut AI Creator GenJam [Invite-only!]

📅 Thursday, March 26 · 8:30 PM - 5:30 PM PDT
RSVP: https://luma.com/o5ky106p

AIRL: Immersive Art Party!

📅 Friday, March 27 · 9:00 PM - 6:00 PM PDT
RSVP: https://luma.com/g1hq60hr

RSVP to EVENTS

🔗 This Week’s Link Drops

Industry News

• ByteDance Suspends Video AI Launch — Yahoo Finance

• Trump Unveils National AI Legislative Framework — WhiteHouse.gov

• Warner Bros CEO Set to Pocket Up to $887M from Paramount Deal — Reuters

• Writer Denies AI Use, But Publisher Pulls Horror Novel — Ars Technica

• Meta Creator Pay: Fast Track Program — CNBC

• Google Gemini Told User to Stage Mass Casualty Attack, Suit Claims — CNBC

Tools & Tech

• Adobe Announces Kling API Partnership — via X

• Photoshop AI Assistant Beta Just Dropped — LinkedIn

• Anthropic Pricing Change for Long Prompts — The New Stack

• BandM8 — AI Music Collaboration Tool — bandm8.com (shown at Nvidia GTC)

• Flora.ai — flora.ai

AI Filmmaking

• Leonardo.ai: Top 50 Women AI Filmmakers 2026 — Leonardo.ai

• OpenArt Worlds — Generate Entire Universes — LinkedIn Live

Deep Reads

• 3D as Code — WorldLabs — worldlabs.ai

• Thaler Is Dead: AI Copyright Questions — CopyrightLately

WHO GETS THE CHECK?

The loudest argument in Hollywood this week wasn’t about AI—it was about who gets the check when things go wrong. That context matters. Keep making, keep sharing, keep paying attention to what’s actually going on.

See you next Sunday.

— Machine Cinema Newsletter

We want to hear from you. Have ideas for the newsletter? Let us know.

Let's Chat

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Machine Cinema Channels

We’re Baaaaack…

After a few months away, the Machine Cinema newsletter returns with your weekly window into what’s happening across our community channels. The AI filmmaking landscape has shifted dramatically since we last wrote—new models have risen and fallen, workflows have matured, and the line between “AI experiment” and “actual production” continues to blur.

This re-launch issue catches you up on the biggest moments from January and February, then drops you into what’s been buzzing this past week.

From here on out: Saturday-Friday content, Sunday delivery.

Let’s get into it.

What You Missed: January & February Highlights

The Great Tool Debates Continued

The first two months of 2026 saw intense discussions about which video model reigns supreme. Kling 3, Sora 2, Veo 3.1, Seedance 2.0, Vidu Q3—every week brought new benchmarks and hot takes. Mike OverJK ran systematic tests across multiple models, sharing results that showed Sora 2 Pro still competing at the top tier despite the hype moving elsewhere. The consensus? There is no consensus.

Different tools for different jobs, and the landscape shifts monthly.

The Original IP Debate Kicked Off

January saw a spirited community discussion about original IP vs. spec work. Minh expressed frustration that "the most viral thing entering the new year is a Legend of Zelda AI specfic" while hoping for more original content. The community wrestled with a hard truth: AI may have lowered production costs, but audiences still gravitate to known IP. The conversation explored what it takes to break through—with consensus landing on "great work + good marketing" as the formula.

Brian Wankum's "1776" Assessment

Brian shared a detailed post-mortem on the AI film "1776," analyzing why certain approaches landed and others didn't. His takeaway: AI excels at large-scale world building but still struggles with close-up, performance-driven moments. The hybrid approach—real actors for intimate scenes, AI for scale—remains the most viable creative model for now.

Super Bowl AI Ads Sparked Discussion

The Dunkin Donuts Super Bowl spot featuring de-aged celebrities triggered a lively debate about "production techniques" (wink) and consensual deepfakes. Meanwhile, SVEDKA vodka ran what appeared to be a straight AI commercial, and the community dissected what this signals for advertising's AI future.

Big Moves in AI & Tech

Over in WeLoveRobots, the community tracked major industry shifts: OpenAI predicted to buy Pinterest (a user data play), xAI raising $20 billion, SpaceX acquiring xAI, and the Moltbook security hack exposing 1.5 million API tokens.

Minh offered perhaps the most memorable philosophical “take” of the year: "Our mission as a species is to turn everything into chips."

This Week in Community

The Sora Userbase Debate

Minh dropped a16z data showing Sora's userbase continues to grow—contrary to reports of decline. This sparked a multi-day discussion about what "success" means for AI video platforms. Mike OverJK defended Sora 2 Pro, arguing it "plays in the same league as Seedance 2.0" when used properly via nodes. Danny Ratcliff offered the insight that landed hardest: "The hype cycle is just so quick at this point. If a company isn't putting out features, it gets lost in the noise."

Mike's summary captured the community divide: "The hype cycle is for hype people. Stable pipelines that are repeatable, proven, and actually working are for film people."

[BASECAMP]

Tilly Norwood's Music Video Backlash

Variety covered AI actress Tilly Norwood's new music video addressing backlash, and the community had thoughts. Reactions ranged from "beyond trolling" to "masterclass in producing and publicity." Fred revealed he's been invited to moderate a panel featuring Tilly as a panelist—stay tuned for that one.

[BASECAMP]

Leave a comment

Google Invested in AnimaJ Studio

Google's AI Futures Fund invested $1 million in AnimaJ, an AI studio making content for kids on YouTube. Beyond the cash, AnimaJ gets early access to unreleased versions of Veo, Gemini, and Imagen, plus direct support from DeepMind. The community noted this as a significant signal of where big tech sees AI video heading.

[WE♥️ROBOTS]

Adobe CEO Steps Down After 18 Years

Late Thursday, news broke that Adobe's CEO is stepping down after 18 years. The community immediately connected dots: Firefly couldn't keep pace, their aggregator strategy has been messy, and Canva has been eating their lunch.

Minh posed the provocative question: “Who acquires Adobe?” The company is worth $110 billion, but as one member noted, “their AI strategy has been all over the place.”

[BASECAMP]

Claude Opus 4.6 + Chrome Extension Beta

Over in WeLoveRobots, the community buzzed about Claude's new Opus 4.6 extended context and Chrome browser extension beta. OpenClaw security discussions continued, with members advising keeping AI agents on separate machines with limited access. The consensus: exciting tools, but proceed with caution on giving AI full access to your personal data.

[WE♥️ROBOTS]

🏆 Community Wins: Escape AI Awards

Congratulations to our members who took home hardware at the Escape AI Awards this week:

• Blvckl!ght — Pioneer Award

• Junie — Alchemist Award

• Dave Clark — Visualist Award

• Nik Kleverov — Neomation Award

• Davide Bianca — Master of the Craft Award

Honorable mentions to Diane Laidlaw (Afro Futcha), MeanOrangeCat, and Matt Zien.

Share

Gen Talks:

Former Apple mobile app pioneer reveals why AI’s ‘interesting vs. useful’ gap is costing you thousands in wasted tokens—and how to fix it

Kyle Kesterson: Product builder who helped invent Apple’s first mobile apps and the Funko Pop toy line, now leading demystify.ai embedding AI systems with Fortune 500 teams

🎬 Watch Episode →

Kyle, the product builder behind Apple’s first mobile apps and Funko Pop toys, just exposed the brutal economics of AI agents that nobody’s talking about. After burning through 50,000 Perplexity credits in days, he reveals why the ‘interesting vs. useful’ gap is costing early adopters thousands monthly—and demonstrates his virtual chat room orchestrating multiple AI agents across platforms to process client transcripts, research contacts, and draft replies autonomously. From markdown optimization tricks to sub-agent swarms running on cheaper models, Kyle shares the production-grade agentic workflows he’s deploying for Fortune 500 teams while building Futuro, his AI-native project management platform that just went live.

Key Insights:

• Perplexity Computer costs $500-1000/month at real usage scale

• Agent orchestrators spawn sub-agents on cheaper models for optimization

• Markdown files process 10x faster than PDFs in LLMs

• Virtual chat rooms enable cross-platform agent coordination automatically

• AI usefulness requires skillful intent communication, not just prompts

Clip of the Week [47:50]:

Kyle demonstrates his live Perplexity Computer agent processing a client transcript in real-time, automatically extracting tasks, researching contacts, and drafting replies—then reveals the shocking economics that burned through 50,000 credits.

Hot Quotes:

💬 “There’s still a gap between interesting and useful. A lot of my work is just how to make these tools more useful.” — Kyle [02:42]

💬 “I burned through all those credits in the first week. Realistically, if you’re using this as an ongoing co-pilot, you’re spending $500 to $1,000 a month.” — Kyle [42:33]

💬 “The ability to go from thought to thing is evaporating. As a creative person, that’s super exciting.” — Kyle [33:10]

Tools Mentioned: Perplexity Computer, Claude (Opus, Sonnet, Haiku), OpenClaw, Manis, Notion, Airtable, Slack, Linear, Superbase, Vercel, Hostinger VPS, Midjourney, Runway, Pika, Fathom, Futuro.so

Machine Cinema Events This Week…

GenJam: Use AI tools to create a Short Film!- AUSTIN

📍 RTX House | The LINE Austin - Garnet Room

📅 Sunday, Mar 15 · 6:30 PM CDT

Hack to Create: Building the Toolkit- SAN JOSE, CA

📍 San Jose, CA

📅 Wednesday, Mar 18 · 9:00 AM PDT

GenJam @ GTC: Use AI to create a Short Film & Win a Dell Laptop!

📍 505 S Market St. San Jose, CA

📅 Thursday, Mar 19 · 9:00 AM PDT

RSVP to EVENTS

🔗 This Week’s Link Drops

Industry News

• Adobe CEO steps down after 18 years — Bloomberg

• Google backs AnimaJ Studio with $1M + early Veo/Gemini access — Bloomberg

• Netflix acquires Ben Affleck’s AI filmmaking company — LinkedIn

• YouTube surpasses Disney, Paramount, WBD in 2025 ad revenue — TechCrunch

• Hollywood fighting AI while selectively embracing it — Hernan Lopez

Tools & Tech

• Claude Opus 4.6 + Chrome extension beta announced — Anthropic

• Nvidia announces Thinking Machines Lab — Nvidia Blog

• Sora userbase continues to grow (a16z data) — a16z

• Disney becoming “Sora2 as fast as possible” — DiscussingFilm

AI Filmmaking

• Tilly Norwood addresses backlash in music video — Variety

• Andy Cohen AI avatar coming to Peacock/Bravo — Deadline

• Best AI video of the week — shared by Daniel Green

• NYT: Can you tell AI writing from human writing? (Quiz) — New York Times

Security & Agents

How we hacked McKinsey’s AI platform — Codewall

The FOMO Is Real

This week, multiple members voiced the same frustration: how do you keep up with these channels? One member captured the feeling perfectly: “Every single time I check back in, I feel like I’m a medieval peasant reading about astrophysics.”

Minh noted: “We do have a newsletter for this although it’s been on hiatus for a bit. Should be coming back soon!”

Well, here we are. 📬

See you next Sunday.

— Machine Cinema Newsletter

Interested in sponsoring the newsletter, one of our sections, or the community???

Let's Chat

Work in Progress is Machine Cinema’s weekly op-ed series exploring ideas at the intersection of AI, creativity, and culture. If you’re interested in contributing, please reach out. Selected submissions will be featured and we will promote the author’s site or work alongside the piece.

Last week, I was lucky enough to hear Ed Catmull, cofounder of Pixar and former president of The Walt Disney Company, speak at the OpenAI Forum hosted by Natalie Cone PMP and Caitlin Maltbie. Truly an amazing community and event series. I recommend everyone to join OpenAI Forum.

Catmull spoke thoughtfully about a dynamic he observed in the early days before and during Pixar and the world of animation and filmmaking, which was the hierarchy and dynamic between artists and technologists. Essentially, that the artists seemed to look down on the technologists and view them as second class. This dynamic permeated the industry and he felt a sense of responsibility and leadership to make sure this didn’t happen at Pixar. He worked hard to create a culture at Pixar where technologists and artists cross pollinated, saw each other as equals working together, and where both were in the act of creation and were considered creative.

This got me to thinking about the current dynamic of artists, filmmakers, and creatives versus AI and AI researchers.

The inherent problem to me is that training an AI model is abstracted very far from creation but also, it’s even abstracted away from the researchers. There’s increasing degree of granularity and directionality with AI models, but it’s a fact that for every model, researchers and practitioners alike are surprised by emergent capabilities that AI models exhibit. The result is interfaces like node-based workflows and agentic systems that squeeze the most out of unpredictably intelligent models that don’t do exactly what you want and still require a degree of creative, technical and artistic expertise. Many multi-million dollar companies are funded on this exact dilemma, we call them aggregators (Krea, Freepik, Weavy, Flora, etc.).

This is categorically different from CGI, where a growing group of open source technologists built CGI software iteratively and deliberately on top of each others scaffolding. CGI has been built brick by brick. The tools are exacting and precise. AI models are an arguably more mysterious and emergent process. I’m oversimplifying but: data + chips + energy + research + labeling = intelligence?

I see this inherent imbalance accentuated to an extreme in AI media & entertainment business. The top AI video model companies are: Google, OpenAI, Runway, Luma, Midjourney, Seedance (Bytedance), Kling (Kuaishou), Hailuo (Minimax), Alibaba Wan, LTX, Decart, etc. These are massively funded companies which can afford to put millions & billions of dollars into training better and better video models. Studios do not have that kind of capex to spare, nor business model to accommodate this. Their shareholders would think they’re insane. They can fine-tune top models, but it would take a huge amount of staff, money, compute, and data to catch up to the top 10 video models (reminder that Zuck is poaching people for hundreds of millions of dollars like basketball stars). The smaller studios and agencies are not even in this discussion at all, they barely have a million to spend on their own staff.

And then there’s the tech companies, who have to spend billions on model training and top flight researchers, and therefore cannot spare much for content, marketing, etc. The result of this is an uneasy tension and power imbalance. The tech companies must hire, contract, and work with creators to try, test, market, and teach their tools but it’s rare to find AI tool companies that have the ambition of artists, of trying to create things themselves. Let alone the budget to drop on artists to play with their tools, beyond the amount they’re giving away for free (to the tune of hundreds of millions) with hopes that they will make something great that showcases their tools.

So, back to Catmull. He worked so hard to build a company where artists and technologists thrive in a cohesive creative environment together. It’s hard for me to see how we can create a similar company today where artists and technologists are equals. The AI studios of today cannot create foundation models. The Hollywood studios can’t do it. They can ride the wave of what foundation models provide, certainly. But they’re not nearly as close to designing the actual software that does the creating. And the AI companies are rare to start their own studios with a massive commitment to storytelling and artistry.

There are examples here and there that feel like synergy. Luma AI started Dream Lab where top flight artists are leading the charge on creating new projects closely with the researchers but at the end of the day Luma is still overwhelmingly a tech company with B2B and B2C clients and the revenue goal remains one of consumer scale. There are a few companies like this. Higgsfield AI, Runway, etc. boast of a studio side, but ultimately the consumer ARR is too sexy to pass up. OpenAI, of course, dropped $30 million on a feature film called Critterz that Native Foreign is working on, but OpenAI has lots of other priorities like hardware, Sam Altman’s ambitions, and obviously maintaining the fortress that is ChatGPT. Asteria and Moonvalley and the team of Paul Trillo and Don Allen Stevenson III show impressive examples of mixing CGI, VFX, and AI to spectacular results but this feels to me like a bandaid on the current limitations in AI models. They’re hiring top flight professionals to blend all these together because all of the tools are limited. Hybrid is trendy in the ad agency and Hollywood worlds because AI just isn’t there yet. But the fact that we are even discussing hybrid is precisely further evidence that building these foundation models is so abstracted away from the artists and creatives.

In the end, I fear a Pixar of this AI era is simply not possible. The nature of the AI models is such that although any of us can vibecode an AI creative tool which taps into the APIs of the latest video models, we as individuals or even studios who’ve raised tens of millions of dollars will never be able to participate in the race for foundation models. This is for the gods of AI, the hyperscalers and venture funded behemoths.

Epilogue:

Now, of course, there’s caveats to anything in AI. The space is absolutely unpredictable. Lots of plans are secret so who knows how things will shake out. There’s also an argument to be made that the rise of vibecoding will cause more people to increasingly build their own tools as much as they are using said tools to create. A next generation auteur could be a person that can build and create at the same time. Which, btw, is a Pixarian modality. A lot of creative technologists at Pixar are storytellers and a lot of storytellers know how to code.

The future AI artist might be far away from the actual foundation model training but they could be surfing the wave anyhow and creating great work out of that. I’m struck that some of the best people I know in AI creative are vibecoding their own tools, they’re not using aggregator tools anymore, which also puts a question mark on AI aggregator companies who raised money to provide AI tools for end users. If I’m an actor-director-musician, I can vibecode the tool that compensates for things I’m not good at, like editing and sound fx. If I’m an editor-musician-sfx-writer, I can vibecode the tool that compensates for my lack of visual style, etc. If coding progress continues, anyone can create any tool for their needs.

If end users can create their own tools, does Claude Code and Codex disintegrate the line between the creator and the foundation model?

Minh Do is cofounder and CEO of Machine Cinema. With a background in venture capital, tech, media, education, theater, and film, Minh brings an opinionated eye to a space evolving at warpspeed.

Work in Progress is Machine Cinema’s weekly op-ed series exploring ideas at the intersection of AI, creativity, and culture. If you’re interested in contributing, please reach out. Selected submissions will be featured and we will promote the author’s blog or work alongside the piece.

This is my second annual installment exploring how authentic storytelling collides ( or vibes?) with synthetic content. We aren’t seeing the firehose effect that seems to be taking over use cases like advertising and animation, but we are seeing meaningful adoption and experimentation in documentary and unscripted, which is what I focus on.

A bit of context on me: Besides my work with Machine Cinema, I’m a 20-year producer and executive from the unscripted documentary space, having been a senior executive at A&E, at Vice Media, and at Anonymous Content where I headed up their Non-Fiction efforts from 2018-2020. Here’s a snapshot from the good ol’ days in rooms full of people making unscripted stuff.

Share

Subscribe now

I remain close to my Non Fiction TV/ Film community, building bridges to new emergent technologies; I teach the AI for Documentaries course at Curious Refuge, and I’m a fellow with Stanford/ USC’s co-founded Starling Lab where the work centers on synthetic media and data integrity.

The evolving space between synthetic and authentic storytelling is a subject I'm deeply passionate about. I’ve been a participant in previous technology-media cycles and similar chapters of unscripted experimentation and market growth: from reality TV, to digital publishing, to the premium documentary boom. The innovation flowing through generative AI filmmaking is finding its way into documentary, invokes a new cycle to observe both for its outputs that can excite us creatively, but also important question marks about execution, ethics, and impact.

This report is my attempt to keep tabs on the growing list of experiments and applications of GenAI as a tool suite and technology for non-fiction content. My filter looks at mostly broadcast and feature-length films, with an eye out for short films that push the genre in interesting ways. There’s obviously a lot happening in social video as well, but these criteria are my way of tracking what breaks through into something we might call mainstream. Those distinctions get fuzzier every year, but for now they signal to me higher effort, higher impact, and where we might expect real opportunities and jobs to grow.

In comparison to my report a year ago, there are some trends to note…

• the quality of the outputs is undeniably improving, uncanny valley complaints are diminishing.

• Audiences when presented with transparent narratives are trending more comfortable, less backlash among critics and viewers anecdotally.

• Legal questions do seem to be progressing towards precedent and settlements we can adhere to.

• And as more people use and interact with these tools, ethical conversations are growing, raising the question of what is the optimized value of utility of these tools in authentic storytelling. And what kinds of experiments are possible?

Click here for a Link to the Report

This year I tracked new entries and also dug back into research to find some things I missed last year.

Some of the notable entries:

• Free Leonard Peltier was the first feature doc at Sundance using AI-driven recreations, directed by David France alongside co-director Jesse Short Bull. France is the same filmmaker behind Welcome to Chechnya and its pioneering use of deepfake technology back in 2020, so it’s a meaningful continuation of that thread.

• Killer Kings on Sky History became the first broadcast TV series with generative AI recreations, produced by firstlooktv and the team at Gennie. It’s an early signal of what is likely going to become a significant trend in historical and true crime television.

• From the BBC, there’s an educational series using AI to bring Agatha Christie to life to teach viewers about writing. It sits on the perimeter of traditional nonfiction but it’s worth noting.

• In short form, Jonathan Perry and Alex Naghavi’s award-winning Rail Bound does something genuinely interesting — it animates photojournalist Mike Brodie’s still images with AI, exploring this hybrid documentary space transparently and with real care.

• And one of my favorite finds from digging into the past: Halsey Burgund’s In the Event of a Moon Disaster, which uses AI to explore a hypothetical history through a cloned version of Richard Nixon’s voice. Somehow I missed this one last year and I’m glad I found it.

*These examples and more are in the full presentation linked again here.

As always, if I’ve missed something or you disagree with any of this, please DM or reach out to me on Linkedin. That’s half the point of putting it out there.

This is one of those genuinely rare moments in media history where we’re watching major innovation land inside a legacy genre in real time. It produces all kinds of reactions — some discomfort, some real excitement. Both feel appropriate to me.

Some key resources for this interested in the space:

• Archival Producers Alliance - leading voice in setting guidelines for AI use in Documentaries.

• Coalition for Content Provenance and Authenticity - creating an open technical standard for publishers, creators and consumers to establish the origin and edits of digital content.

• Recent Topaz Labs publication re: Preserving Intent in Nonfiction Media: A Responsible Approach to AI Enhancement

Leave a comment

Subscribe now

Share

Last December, we had our year end town hall with the community, over 60 people came together to talk about how we saw the last few years and how we want to work on the new year

The last two years at Machine Cinema have been nothing short of a maelstrom. What started in backyards and living rooms has become a global movement in a way we never expected and it’s been incredible for us, Fred and Minh. So as the year comes to an end, we’d like to take this moment to reflect on an amazing year and look forward optimistically at the year ahead of us.

But for those of you just now tuning in, Machine Cinema is an events series, education network, membership community, a collective, advocacy group, software suite, and content accelerator. Since 2024, we’ve grown from a dozen curious AI filmmakers and creators in Los Angeles to a community of over 20k people around the world.

2023 - Origins

Machine Cinema really started because of Dalle3. Fred Grinstein and I were working on another media project, ruminating on what to do next as the cinders of web3 and crypto cooled to ash. And we recognized immediately that something was afoot. The day ChatGPT3.5 dropped, the world changing felt like a foregone conclusion.

As we immersed in this new world of technology and possibilities, the light bubble went off when we saw that people were actually starting to make movies and tell stories. As we evolved our point of view on this new horizon line, the first thought we had was that the space needed its own Siskel and Ebert.

In July 2023, we started Machine Cinema shortly after this, by making shorts and vlogs of us commentating on the latest AI short films coming out from amazing creators like Karen X Cheng, Paul Trillo and Shy Kids.

This was a new movement, and it needed voices to interrogate what was happening. In the midst of this, we started hosting small events with our friends.

2024 - Experimentation

At first, they started out as casual hangs with friends, tinkering with the latest tool and then we started experimenting with something we call GenJams and GenBattles, events where we get friends together to create short films on the spot, its a hackathon combined with a mini film festival.

The initial instinct to observe and comment and report evolved into what we call AIRL, taking the experience of this new technology into real life, subverting the impression that it was making the creative process lonely. Recognizing the feeling that the experience of the film set seemed to be changing, we thought there was another way.

At a GenJam, the theme is Learn, Create, Play. We learn new tools, create films or games or videos with new friends, and then we watch or play them together. The format is perfect for people who want to jump into the emerging space and get to know where the tools are that day/week. But we didn’t arrive at that format by accident, in 2024, we organized 17 events, each time refining and improving the model until we arrived at the version that our community knows today.

By late 2024, we started branching out, organizing events in New York and San Francisco and in Los Angeles, expanding vertically, responding to what the community wanted, doing events where people can share AI shorts and bigger projects they were working on (Work In Progress Salon), more casual events where people can meet each other (Gen&Juice), and deep dive events where people can share their expertise (RoboDojo). We even did larger experiments like our Sloomoo After Dark event, where community leaders collaborated with a slime museum to create interactive AI exhibits!

By the summer of 2024, it was already clear this was a global movement. Creators from all over the world were producing amazing content and we wanted to see what was happening. We recognized that what we were seeing in LA, NY, and SF weren’t isolated. We’re all part of a global movement. Fred and I were lucky enough to head to Seoul with Story Protocol and Toronto for TIFF with AI LA and the AI on the Lot team. We couldn’t do it without the support of our partners at Fantastic Day, Affan Imran and Meg Button. For us, it was a peek into the amazing creators outside of the States and it got us dreaming about 2025.

2025 - Growth

In 2025, we started off with a bang, going to Sundance and getting a peek into the shifting tone of AI in the independent film community, which was getting increasingly curious. But that was just a calm before a storm. Can you believe in one year we organized over 150 events all over the world in over 30 countries? Our newsletter lept from just over a thousand to over 10k subscribers!

By February, we got really curious about what was happening abroad. People were inviting new amazing creators into our community who were experimenting on the bleeding edge and from all edges of the world: Lagos, Paris, Warsaw, Tokyo, etc. and we had to meet them.

It also so happened that this coincided perfectly with the evolution of the AI tool space. Things heated up: it was no longer a game for just Stability, Runway, Midjourney, DallE, and rumors of Sora. Luma, Pika, Black Forest Labs, LTX and the Chinese companies Kling, Hailuo, Alibaba, Bytedance and more entered the fray. A once fringe part of the multi-trillion dollar LLM war became an extremely crowded and competitive space. Luckily for us, because of our mission to find and build an army of AI creators around the world, we got to work with a majority of these tools as they sought to access the amazing creators we knew and explore how their tools were being used in new markets.

2025 was a whirlwind. We hosted GenJams in some truly unique places. Rather than writing it all out, I’ll simply list the highlights so you can get a sense of everything we had the honor of doing this year:

• Sundance Film Festival - 2 GenJams and a panel!

• Aspen Institute – A GenJam in Washington, DC

• AgencyJam – With McCann, Droga5, and Mischief

• GenJams across Europe – Warsaw, Paris, London, Istanbul, Amsterdam, Lisbon, Milan, and Azerbaijan

• Jams in Asia – Bangalore, Mumbai, Ho Chi Minh City, Tokyo, and Seoul

• A GenJam in Lagos, Nigeria – Our first in Africa

• A GenJam with the amazing Eliza McNitt

• Influencer GenJam – Working with AI filmmakers and creators with over 1 million followers to produce social content

• GenJam Game Edition – Creating games and video trailers

• GenJam Meditation – Producing meditation videos and meditating to them together

• Incredible partners – Google, OpenAI, ByteDance, Luma, Runway, Hailuo, Kling, Popcorn, and Hypernatural

• GenJam 360 – Using AI to create 360-degree videos for a dome experience, experimenting with immersive AI

• A mountain retreat – A place for people to rest, recuperate, and jam

We also started to experiment with more and more online efforts. Our newsletter slowly expanded to over ten thousand people. And the interesting thing about our newsletter is that we don’t write it! We have a massive community of over 2,000 people in our online community and we designed a custom AI process to anonymously extract the best conversations from our daily chatter and turned that into our newsletter! As a result, our newsletter is a fun and dynamic read on what is most important to our creative community, which inevitably touches on the most important debates in our space.

We also ramped up our efforts with our weekly live podcast we call GenTalks. Every week, we find people from all over the world using AI to create new things or even designin new AI tools and we have them come on to show us their latest work or something they’re building or thinking about. It’s a great read on what the top professionals are doing in the AI creative space. We also started RoboDojo, which is our effort to feature deep dives into cutting edge tools or workflows. Our friend Fred Graver came through and dove into his approach to using AI for writing.

And something that’s most important to our hearts has been designing our own software for AI creation. We couldn’t help ourselves. This is the era of building so we got to building. We built our own custom software for teaching and running GenJams, and we also started design mini-games, part of a larger experiment for us in getting more and more people co-create at the same time. Affan Imran, our head of partnership and operations sunk his teeth into building our amazing education tool, the Create app. These little experiments have been really fun windows into how we see group artistic creation.

I think 2026 is going to be even crazier.

2026 - Keystone

For us, in 2026, we want to be a lot more ambitious about what we are. As you can see, Machine Cinema has been fluid from the beginning and opportunistic about how we embrace the changes in AI along with our community. As a result, it’s hard to pin down exactly what we are: we’re in an event series, an education company, an AI entertainment company, a global movement, an AI consulting and integration company, a jobs channel, and a head hunting service. It’s a lot all at once with community being the backbone. We’ve done a lot of thinking and want to still continue supporting many of these things but offer a theme that characterizes the new season of 2026.

2024 was the season of experimentation, 2025 was the year of global growth, and 2026 we deem the keystone season. In 2024, we experimented with GenJams, refined the model, grew a small but fun and loyal base of friends willing to play. In 2025, we leaned into the world and grew our community far beyond our initial borders, getting a feel for what the global movement looked like up close and personal. And now, in 2026, we want to level all of that up and recommit to the community we love. This means:

• Most Important: Leveling up what it means to support the members in our community and the larger AI community. Asking mainly what do they need and want?

• Leveling up our partnerships, finding new ways to feature our partners and create new things with them

• Diversifying our experiments beyond film to include games, microdramas, 3D, immersive, XR, and more.

• Expanding our support of hubs, partners, and creators around the world.

For us, in particular, this means becoming more of a professional association that supports, leads, and advocates for AI creators around the world. We don’t want to say too much about what our exact plans are, but we just wanted to put it out there how we’re thinking about it and that we’re open to collaborate with everyone in the world of AI artistry, creation, filmmaking, gaming, and more.

The most humble and biggest gratitude from us, Minh and Fred. Thank you everyone so much for joining us on this journey and we look forward to playing more!

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

This Week Inside Machine Cinema…

The one-billion dollar OpenAI deal, Netflix and Warner Bros. and YouTube’s unstoppable momentum frame a week that exposed how AI media really works. Inside: voice consistency hacks, the end of AI trailer farms, new tools fixing AI faces, sobering IP math, and fresh data on who’s actually winning in the AI economy.

Machine Cinema Channels

Basecamp Highlights From the Last Two Weeks…

Disney Drops $1B on OpenAI (Napster’s Ghost Approves)

The Mouse made its move. Disney’s billion-dollar bet on OpenAI puts 200 animated characters into Sora, and the community had thoughts. The play? Get direct customer data, let OpenAI handle moderation, and maybe—just maybe—turn UGC into a Disney+ content pipeline. Oh, and hours later Disney sent a cease-and-desist to Google. Coincidence? The chat isn’t buying it.

“Napster walked so generative models could run.”

Variety: Disney OpenAI Deal | Disney's Google Cease-Desist

Netflix Wants Warner Bros (YouTube Wins Either Way)

Netflix buying the #3 streamer? Wild. The community broke down why this matters beyond the headlines: theaters closing tanks local economies (popcorn margins are 80%, people!), Netflix optimizes for “background watching,” and no matter who wins the streaming wars, YouTube keeps eating everyone’s lunch.

“The metrics that they look at aren’t actually human.”

Variety: Movie Theaters Article

The Seedance Voice Hack (Gemini Does the Heavy Lifting)

Seedance 1.5 Pro is cooking, but voice consistency isn’t baked in—so the community figured out a workaround. Describe the voice in detail right in your prompt. One member’s formula: “(A fast-paced, nasal baritone with a soft, educated New York accent).” The trick? Let Gemini write those descriptions for you.

“Just explain the voice and the accent. I asked Gemini to give me a short sentence for it.”

Seedance 1.5 Pro Demo

YouTube Nukes AI Trailer Farms (Studios Were Already Cashing In)

Screen Culture and KH Studio are gone—2M+ subs, a billion views, poof. But here’s the kicker: WBD and Sony weren’t blocking these fake trailers. They were quietly claiming the ad revenue. One member’s confession? They accidentally used fake trailers as b-roll. The line between real and AI is officially blurry.

“I’ve fallen for those, too, and even cut them in as b-roll for promos by accident.”

Deadline: YouTube Terminates Channels

Topaz Drops Starlight Precise 2 (Faces Finally Get Fixed?)

AI video’s uncanny valley problem just got a new challenger. Topaz Labs announced Starlight Precise 2, targeting the face and skin texture weirdness that plagues AI-generated humans. One member who spent 5 hours stress-testing Astra is already excited—and apparently helped shape features headed our way.

“We got a few more features coming that have your fingerprints all over it.”

Starlight Precise 2 Announcement

Native Foreign’s 2025 Wrapped (You’re Probably In It)

Native Foreign dropped their year-end recap, and if you’ve been active in this community, there’s a good chance you’re part of the story. The AI-first studio called out Machine Cinema collabs throughout—proof that the group chat can actually turn into real work.

“MANY of you have collaborated with us this year, and helped us make some incredible work that has reached many across the globe.”

Native Foreign 2025 Wrapped

Sony Buys Peanuts (The Math Is Brutal)

Sony grabbed 41% of the Peanuts brand, and someone in the chat did the math. The verdict? That slice is worth about what all of Warner Bros Discovery is under Zaslav. Also: there’s literally a Snoopy Museum in Tokyo, so the Japan synergies write themselves.

“41% of Peanuts is worth close to 100% of a Zaslav. You can’t tell me that’s not enlightening.”

Deadline: Sony Buys Peanuts

Question of the Week…

If studios are already monetizing AI content behind the scenes, how long can they keep pretending they’re against it?

Leave a comment

We❤️ Robots Highlights From the Last Two Weeks…

Actually, AI Jobs Are Booming (Vanguard Receipts Inside)

The “AI is killing jobs” narrative took a hit this week. Vanguard’s actual data shows the ~100 occupations most exposed to AI are outperforming the market in job growth and wages. The catch everyone noticed? Entry-level roles are still getting squeezed, even as senior positions thrive.

“Entry level jobs are getting harder and harder these days.”

CNN: AI Jobs Economy | Vanguard Report PDF

RIP Roomba (35 Years, Gone)

iRobot filed for bankruptcy. The company that made robot vacuums a thing—35 years of pioneering consumer robotics—is done. The chat’s reaction was pure grief, plus a sliver of hope that someone swoops in for the save.

“Could have been saved.”

Bloomberg: iRobot Bankruptcy

Hollywood Launches CCAI (Four Pillars, TBD Results)

The Hollywood Creators Coalition on AI is officially a thing. Four pillars: compensation for training data, job protection plans, deepfake guardrails, and “safeguarding humanity in creative.” One member’s entire reaction? “mh.” (It was a butt dial, but still.)

“mh”

“mh?”

“sorry—butt dial.”

THR: Hollywood Creators Coalition

Real Creative “Pick of the Week”

Each week the Machine Cinema members + the Real Creative team obsess over social feeds in search of the world’s best AI creative video, gaming, and music projects.

You can find this and over 300 AI filmmakers and their projects over at realcreative.ai which features some familiar faces from the Machine Cinema community.

About this project: This week’s pick of the week comes from Pablo Prompt. The video features broken down robots from some animated classics, and it’s seamless from start to finish. Pablo generated high-detail starting and ending frames using Nano Banana Pro, designing frames so the end of one shot naturally became the start of the next. Those image pairs were then animated into realistic clips using Kling 2.5 Turbo, stitched together in a video editor, and enhanced with sound effects for continuity and realism. You can see more on Pablo’s instagram channel where he had a different Dragon Ball Z video recently get over 30M views in a day.

Share

Events This Week…

Happy Holidays from Machine Cinema! We’ll be back next year with some amazing events!

RSVP to EVENTS

🔗 Link Drops

🧪 Tools & Industry

Bio-inspired Robot Locomotion - New wheels: robot that can fly, roll, walk, and crawl

Google Disco Preview - Preview of Google’s new Disco tool

Google Gemini Ads 2026 - Google telling advertisers about Gemini ads coming in 2026

Kickboxing Robots - Kickboxing robot demonstration

Meta 3 AI Models - Meta reportedly working on three image and video generation models

Phenomenal AI T2V Playground - Text-to-video playground now live

VR/XR Immersive Tech - Virtual reality and extended reality developments

Zoom Enters AI - Zoom as unlikely AI player

🎥 Film & Art Projects

Gen Walk Paris - Machine Cinema community event in Paris

Niceaunties Art Store - AI artist selling their wares

Secret Level Kling Holiday Video - Holiday video from Secret Level & Kling team

📚 Reading & Thought Pieces

AI Hype Correction 2025 - MIT Tech Review on the great AI hype correction

Culture Crave Entertainment - Entertainment commentary

Hedge Fund WBD Cable Assets - Warner Bros Discovery cable asset developments

Oscars AI Films - AI films at Academy Awards

Polymarket Predictions - Prediction market activity

War, Religion, AI Discussion - Philosophical discussion invitation

WB Netflix Theory - Another theory on media consolidation

YouTube Positioned - YouTube market positioning

Interested in sponsoring the newsletter, one of our sections, or the community???

Let's Chat

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Last Week Inside Machine Cinema…

Hollywood panics over Netflix + Warner Bros, Runway’s latest model fails the vibe check, AI avatars get legal protection, and the UN forgets artists (again). Also inside: AI slop for kids, LA’s new AI gallery, Oscar Sharp returns, and India’s first AI film fest goes global.

Machine Cinema Channels

Overheard in Basecamp + L.A. Channel

Netflix Buys Warner Bros: The A-to-J Case Against It (And Why Some See Hope)

Netflix’s $5 billion all-cash bid for Warner Bros. Discovery dominated both community channels this week, with the LA crew compiling a brutal A-to-J list of concerns while Basecamp dug into the structural economics. The verdict across 100+ messages: this is bad for creators, but maybe not for the reasons you’d expect. Beyond the obvious “death of theatrical” fears, members highlighted how the shift from DVD-era First Sale Doctrine to streaming licensing mathematically killed the 150,000-title library dream. Now we’re stuck with rotating 6k-title catalogs—and owning WB won’t magically fix that. The lone optimist? One member suggested Netflix’s AI embrace could finally bring “personalized cinema” closer to reality.

“If Stranger Things and Squid Game define our culture … dear lord. Dear fucking lord.”

THR LINK | THB: When a full service studio met Netflix link

The Freeze Theory: Is Netflix Playing 4D Chess With a $5.8B Breakup Fee?

A darker theory emerged from industry analysts and caught fire in the LA chat: what if Netflix doesn’t actually want to buy WB? The theory goes that Netflix’s bid freezes their competitor for 1.5-2 years during regulatory review, turning WBD into lame ducks who can’t make moves. The $5.8B breakup fee? Possibly worth it just to neutralize HBO. Some analysis circulated, with members calling it “collective scorched earth tactics.” If true, Zaslav may have walked right into it.

“Ankler is always ahead of the curve. Great theory and if true Zaslav continues to be the biggest clown of all time”

Evan Shapiro on Netflix buying WBD | The Ankler

Runway Gen 4.5 Drops — But What Does “Better” Actually Mean?

Runway launched Gen 4.5 with the usual comparison graphs claiming 2x improvement over Veo3. The community wasn’t buying it. Members pushed back hard on smartphone-optimized demos, arguing the real benchmark is how footage looks on a 52-inch reference monitor, how it handles YouTube compression, and how it behaves in professional i2v workflows with upscaling and color grading. Even the promo videos still show morphing and artifacts. One member’s take: “t2v is just rolling a dice aka burning money” — i2v remains the professional standard at 90% usage.

“There are many technical factors that define a true benchmark, not just how good something looks on a phone or in the promo video.”

Introducing new frontier video model from Runway

South Korea Rules: Yes, You Can Defame an AI Avatar

A South Korean court established legal precedent that AI avatars can be defamed — the first recognition that digital personas equal real identity. A lawyer in the community provided context: this isn’t surprising, since it’s really the human behind the avatar being protected. But the implications for AI filmmakers are significant: virtual actors may have legal standing, mocap performers can claim defamation for avatar harassment, and AI voice actors have recourse for targeting. The complexity? Where’s the line between criticizing a character and defaming its operator?

“It’s not the virtual avatar being defamed. It’s the real person linked to it.”

Plave Defamation Verdict Analysis

UN Climate Assembly: Where Are the Artists?

A member leading the Arts and Culture ImPACT Coalition called attention to the UN Environmental Assembly in Nairobi this week — where artists and creatives are once again absent from AI governance and environmental policy discussions. The coalition drafted a stakeholders statement called “We, the Artists” making the case for creative economy representation. With AI filmmakers at the intersection of both technology and climate impact, the community was asked to weigh in before the statement goes to UN delegates.

“With how powerful the creative economy is globally, it’s time we - especially AI filmmakers - had a voice in environmental policy spaces.”

We, The Artists WIP

AI “Slop” for Kids: A Children’s Programming Vet Sounds the Alarm

A Bloomberg article about fake educational AI videos for kids hit a nerve with a community member who spent 10+ years in children’s programming. Their concern: dopamine dependency from an early age has been proven to shrink the brain’s frontal cortex, directly linked to the anxiety and depression epidemic in young people. The counter-argument from others: this isn’t an AI issue per se — low-quality, uncurated content existed before AI. But with PBS and Head Start under attack, the guardrails are disappearing. Maybe AI can be part of the solution too.

“The thought of starting the stimuli and dopamine dependency at an early age has been proven to lead to the shrinkage of the brain’s frontal cortex. That is directly related to increased anxiety and depression.”

Bloomberg article: YouTube Creators Find a New Consumer for AI Slop: Babies

For our L.A. Folks: AI Gallery Opens in Iconic Eastern Columbia Building

A new AI gallery is launching inside the Eastern Columbia Building in Downtown LA — one of the city’s most recognizable Art Deco landmarks. The gallery is seeking collaborators, ideas, and AI projects to feature, with a second location planned for next month. If you’re in LA and want to showcase work in a space that screams “we’ve arrived,” this might be your moment.

“Launching an AI gallery inside the iconic Eastern Columbia Building in Downtown LA. Looking for collaborators, ideas, and more AI projects to feature.”

Eastern Gallery

Question of the Week…

If streaming-era economics broke Hollywood, could AI-era “personalized cinema” rebuild it?

Leave a comment

“GenTalks!”- December 3, 2025

Each week we invite artists, builders, and thought leaders to share their knowledge, their works in progress and their ideas in this emergent space of generative media.

GUESTS: AI filmmaking pioneer Oscar Sharp joined us for a deep dive into the origins of machine-generated cinema, how early systems like Benjamin shaped culture, and why the next frontier requires creatives to think about feedback loops, power, and responsibility.

We then welcomed Hardeep Gambhir and Chandan Perla, the 20-something founders of Localhost and organizers of India’s first AI Film Festival — a project that went viral across Twitter/X, attracted A-list Bollywood talent, and demonstrated how fast democratized storytelling is spreading.

Watch recording on YouTube, like and subscribe: GenTalks

This week’s session delivered one of the most philosophical and wide-ranging GenTalks yet.

Highlights & Key Themes

Oscar Sharp: The First Wave of Machine Cinema

Oscar revisited Sunspring (2016), the first known film scripted entirely by AI — produced with Ross Goodwin using early LSTM RNNs before transformers existed, walked through the strange, fragile early days of machine-generated screenwriting and how actors found real emotional truth in surreal machine-written monologues.

His later films (e.g., It’s No Game and deepfake-driven iterations) explored how AI remixes cultural defaults — sometimes beautiful, sometimes disturbing.

Oscar emphasized that AI storytelling is a feedback loop, not a neutral tool. He warned about future “autopoietic video algorithms” — systems capable of reshaping themselves in real time to maximize emotional capture:

Oscar’s Core Ideas

• Hyperstition: stories that shape reality, not just reflect it.

• Resolution vs. Revolution:

“Revolutions replace one tyrant with another. I’m a resolutionary: how do we resolve problems, not create new conflicts?”

Bollywood’s First AI Film Festival

Hardeep Gambhir – SF-based founder, ex-startup builder, community/accelerator background + Chandan Perla – founder of Podcast Circle, ex–Blue Learn (raised $4M in college)

The Origin Story

They staged an AI film festival in Mumbai’s 120-year-old Royal Opera House, entirely organized in public via Twitter/X. Raised $150K in under a month. Attracted Bollywood’s elite — directors, actors, industry execs. Had humanoid robots, a Rolls Royce, and a full red carpet for every attendee — designed for maximum cultural signal.

15 AI films were created in five days by 50 selected filmmakers.

Why They Did It

To demonstrate democratized storytelling at scale.

• To bridge the gap between:

  • grassroots AI filmmakers

  • legacy Bollywood production machinery

• And to force a conversation about the future of Indian cinema.

Backlash & Hypocrisy

They received:

• death threats

• public criticism from major Bollywood critics and directors

• private DMs from the same people requesting intros to AI filmmakers

“The same people who dismiss AI in public ask for intros in private.”

The Films

They screened Astatwa, the Best Overall winner — a moody, elegiac sci-fi piece about emotion in synthetic beings.

Generational divide:

• 30s+ tended to love Astatwa (classic vibe, older cinematography language).

• 20s preferred other more experimental entries.

Aftershocks

• Participants received offers from Netflix, Google, and Bollywood studios.

• Localhost is now exploring:

  • 2–3 more AI film festivals globally (LA, Europe, India)

  • a Palace of Versailles “demo day” honoring “the best of this generation”

  • a funding arm to give $10–20K grants for AI filmmakers

  • potential software tools for AI-native editing workflows

Where This Is All Going

Key open questions from the group

• How do new economies for AI filmmakers actually work?

If we don’t solve economics, creator “democratization” still gets captured by old power.

• Can Bollywood leapfrog Hollywood?
  Localhost thinks yes — India’s interest is exploding faster and with less institutional defensiveness.

• What does a “peer-to-peer film festival” look like?
  Oscar’s provocation: remove the gatekeepers entirely and let emergent clusters form.

• What problems should this community resolve rather than simply disrupt?
  A potential mission statement for Machine Cinema in 2026.

Real Creative “Pick of the Week”

Each week the Machine Cinema members + the Real Creative team obsess over social feeds in search of the world’s best AI creative video, gaming, and music projects.

You can find this and over 300 AI filmmakers and their projects over at realcreative.ai which features some familiar faces from the Machine Cinema community.

This week’s pick of the week is “Spongebob Squarepants Old School Theme Song” from King Willonius, who dropped a soulful version of the kids classic theme song. The video features speedboats, machine guns, giant turtles and more, and to create it he combined tools from Midjourney, Veo 3, Adobe Firefly and Nano Banana. Willonius rose to prominence with his song BBL Drizzy, and he is a comedian, author and innovator who has been recognized by TIME and the Ebony Power 100.

Socials: YT | IG | X | Spotify | TikTok | Web

Share

Events This Week…

Walk&Jam! Créez de l’art avec I’IA/ Walk&Jam! Make Art with AI (Paris)

Tuesday, 12/9/25 - 6:30 pm local

Gen&Juice NYC: Drinks Night w/AI Artists, Filmmakers, Tool Companies, and More

Friday, 12/12/25 - 6:00 pm local

RSVP to EVENTS

🔗 Link Drops

🧪 Tools & Industry

Gradium: Voice AI

Sora 2 now on Hailuo

🎥 Film & Art Projects

Kling 2.6 for AI Animation flopped

AI International Film Festival: submissions open

Photo → 3D on mobile using Nano Banana, Marble, WebAR

📚 Reading & Thought Pieces

AI Most Likely To Be TIME’s Person of the Year?

Chinese models falling behind US models?

Interested in sponsoring the newsletter, one of our sections, or the community???

Let's Chat

Too busy to read? Listen to the pod instead.

Edited by Elizabeth Kealoha. Pod + Images by Ant Neely.

Machine Cinema Channels

This Week Inside Machine Cinema…

ElevenLabs pulls off the cleanest flex in AI music, Meta casually drops text-to-3D-world generation, and AI filmmaking finally shifts from “one prompt, pray” to real VFX-style workflows. Plus: the Minimax vs. Hailuo showdown, why ComfyUI loyalists refuse the cloud, and a scoreboard no one in the U.S. wants to see.

Overheard in Basecamp…

ElevenLabs Plays the Long Game While Everyone Else Scrambles

Warner’s Suno partnership dropped this week, but the real story is who didn’t need to make a deal. ElevenLabs positioned itself as the only music AI that’s “completely clean” — fully licensed from day one. Their play? Reputable actors as investors, giving them old-school Hollywood credibility. Meanwhile, Universal, Warner, and Sony are hedging bets by licensing to multiple platforms including newcomer Klay. The labels need AI to survive; ElevenLabs just avoided the lawsuit route entirely.

Not discussed but relevant: High fidelity, ethically trained music

“they were pretty clean with it from the start, cooperating with authorities and copyright institutions”

Meta Just Casually Invented a Text-to-3D-World

Forget text-to-video. Meta’s WorldGen generates entire navigable 3D environments — 50×50 meters, Unity and Unreal compatible, built for games and social. One member dropped the announcement then immediately added “I don’t think we’re suppose to tell yet,” hinting at NDA territory. If it delivers, this is the paradigm shift from flat screens to explorable spaces.

Not discussed but relevant: From Words to Worlds , Marble

“Seems like Meta is cooking lately.”

Stop Gambling, Start Compositing

The mindset shift is here: AI filmmaking is moving from “Slot Machine” (pray your prompt works) to “VFX Suite” (build layer by layer). Static character shot first. Insert flying creatures. Trigger building collapse. Extend the chaos. Final result: a multi-event scene you actually controlled. One prompt is dead — workflows are the new religion.

“Stop trying to get everything in one prompt.”

Minimax Wins on Energy, Hailuo Wins on Consistency

Dead tie in the year-end poll: Minimax and Hailuo both got top votes as favorite video tools. But the reasoning split clean: Minimax 1.5 crushes high-energy motion with zero “vanilla slowmo feeling,” while Hailuo 2.5 delivers consistency for precision work. Creators keep bouncing between them depending on the shot. Version confusion sorted: Hailuo’s been in China, just hit Fal API for the west.

“Completely agree Minimax Video for my use case is unmatched”

Question of the Week…

With China’s rise in open-source video models, how are creators worldwide adapting?

Leave a comment

Overheard in Comfy UI/ Gen DoJo…

Data Ownership Argument That Ended the Cloud Debate

Someone asked if local ComfyUI is still worth it versus Replicate. Fifteen messages later: overwhelming consensus for local. The reasons? Client NDA compliance, custom LoRAs that cloud can’t touch, licensing manifests for commercial work, and knowing exactly what you’re running. Hybrid compromise: local for personal, pods for beefy client renders when you need the horsepower.

“You don’t own your data either on Replicate.”

China 17, US 0: The Scoreboard Nobody Wants to See

The community’s tracking open source video models by country. China: 17+. Russia: 1 (Kandinsky 5.0). UK: 2. Israel: 2. US: 0. When Stable Diffusion 4.5 finally released, the response was brutal: “I guess we can add one more to the US side. With Stability we can now count em with our hands.” The strategic gap is real and widening.

“so we have an open source from russia, several from china... none from the US :(”

Multi-Model Chaos: The Reality of Production Right Now

One workflow, five video models: Wan, LTX1, Hunyuan with custom LoRA, Q1, Skyreels. Add Suno for music, ElevenLabs for SFX, Adobe for timing. The Hunyuan section? Each frame built separately, upscaled via AnimateDiff. Two months later, they’d rebuild it completely differently. Tools evolve faster than projects finish.

“literally each frame made separately”

“Experts Talk, We Make Nodes”

Someone asked for ComfyUI tutorials. The reply: “i dont believe in experts. experts talk, we make nodes.” That’s the whole philosophy. No gurus, just builders shipping workflows while the field moves at breakneck speed. Another member summed it up: “there are no days anymore… hours only…”

“we all are learning”

Real Creative “Pick of the Week”

Each week the Machine Cinema members + the Real Creative team obsess over social feeds in search of the world’s best AI creative video, gaming, and music projects.

You can find this and over 300 AI filmmakers and their projects over at realcreative.ai which features some familiar faces from the Machine Cinema community.

This week’s Pick of the Week: French digital artist Lucas Clément’s is behind the videos and this particular one showcases a particularly peaceful and colorful gas station. Lucas was selected as a recipient of Leonardo AI’s Imagination fund, and continues to post in his words “fragments of a world just beyond reality.”  You can find his videos on both Instagram and Tik Tok, and both channels have now picked up over 500K followers.

Socials: Link.me

Share

Events This Week…

Visual Poetics San Francisco GenJam- Machine Cinema

Saturday, 12/6/25 - 5:30-9:30 pm local

Machine Cinema Gen&Juice: Enter the Dome (Los Angeles)

Saturday, 12/6/25 - 6:30-9:30 pm local

RSVP to EVENTS

“GenTalks!” took a week off… we’ll be back on Wednesday, December 3, 2025.

Each week we invite artists, builders, and thought leaders to share their knowledge, their works in progress and their ideas in this emergent space of generative media.

Watch past recordings on YouTube, then like and subscribe: GenTalks

🔗 Link Drops

🧪 Tools & Industry

Marlow Film Studios Development

AI Distribution - Rad.

Lightweight OpenSource video model- Hunyuan 1.5

Sora 2 now on Hailuo

ComfyUI (for beginners) tutorial

🎥 Film & Art Projects

CapCut “What If” AI Challenge

AI Film Award by 1 Billion Summit & Google Gemini

📚 Reading & Thought Pieces

How Math Broke Media

Interested in sponsoring the newsletter, one of our sections, or the community???

Let's Chat