I posted about this on Threads and Bluesky and immediately was asked for a blog version. Please enjoy, share, comment and follow me for more (as this is an ongoing topic I keep exploring).
0:00
-21:43
In early 2024 a friend told me he was beginning to produce Vertical Short Dramas as “the solution to everything Quibi and early YouTube got wrong.” As an *early adopter* of new content formats, I’ve shadowed his journey and many others… and now it seems everyone I know in the entertainment industry is discussing this “gold mine.”
But something doesn’t feel right.
Literally everyone I know praising the format is also judging and disparaging the content. “They are making so much money on terrible scripts - they should pay me to make a good one.”
*About the author: I was early to YouTube…I’ve literally published, patented and pioneered blockchain and virtual reality storytelling concepts…and you might remember my 🍍 pineapple hat from the 2023 strike.
The trend is filming roughly 100 one-minute “episodes” of a melodrama story, releasing the first 10 episodes for free on social media and a proprietary app to “hook viewers.” Audiences can then purchase in-app “coins” or watch ads to pay-per-swipe to continue the story, resulting in the audience spending roughly $20 per feature-length storytelling experience.
The storylines and dialogue are often criticized for being cliche, AI-generated scripts with very low production value and often over-animated performances by unknown actors. This is often due to filming the entire series within 7 days with a budget of $50,000-$200,000 using non-union casts and crews, funded by international platforms who prioritize volume over perceived quality. In fact they film so quickly and at such a high volume that it’s common to suddenly recast if there is literally any delay or negotiation.
When you look at these same companies and series on traditional social media platforms the views are *fine* but not of a volume that would indicate they are converting an audience that would be meaningfully monetized.
I want to be clear that I don’t actually have any judgement against the content. I’ve made indie films for 20 years and I actually enjoy a little bit of elevated theatricality, but from a business thesis: You’re telling me there is this audience paying millions for low-fi content at a time when streaming services and theatres are struggling with retention? Sorry, the math just doesn’t math. (But I’m happy to be wrong!)
In the last 6 months I’ve starred in two feature films, both of which have been critically acclaimed and are available on most platforms to watch free-with-ads or to rent for $5… and we are STRUGGLING to get viewership.
Side note, the best way to support me is by watching one of these:
The Devil and the Daylong Brothers: A southern gothic musical about three brothers on a bloody quest through Appalachia.
Succubus: A screen-style body horror film about a man joining a dating app, only to swipe right on what may be an inhuman presence.
Every Hollywood publication, agency and golf course is talking non-stop about this growing format and “all the money” it’s making.
But according to who?!? Has anyone actually seen this money?
The budget of these projects mean the day rates are micro-budget terms, usually a few hundred dollars for supporting roles while a few “stars” can negotiate around $1k per day, but…
That’s the LOW end of what you’d be paid on a proper film or tv show… and you would then normally get residuals, insurance, etc.
The “profit” is all internally earned and self-reported. This doesn’t pass the smell test. Especially when you consider the 2023 strike revealed the “cord cutting” hype from streamers actually resulted is decreased earnings and unsustainable models for traditional film and TV.
And now the hype is forcing everyone to shrug and consider, “well I guess this is the future.”
Sound familiar?
We’ve recently been told repeatedly to adopt marketing buzz without verifiable results demanding the world invest endless capital to “adapt or die.”
Who would have deep pockets…
need a vast library of content…
made as quickly as possible…
not care about storytelling…
work under non-union contracts…
and be primarily based overseas?
Y’all… this sounds like A.I.
For anyone criticizing the “quality” of the output: There is always a reason. I don’t think anyone is “accidentally” lighting money on fire. They clearly don’t want/need quality. It’s a library build.
It’s growing a content pipeline as fast as possible.
But for who?
I think these actors are about to become base characters for a generative model to make the REAL movies and tv shows… using the likeness in perpetuity that they signed away for a few thousand dollars.
They’re not building hits.
They’re building a massive, tagged, pre-structured library of:
Faces
Voices
Emotions
Scenes
Contexts
That’s the perfect training set for multimodal AI. These “vertical drama companies” aren’t studios…
They’re AI training farms in disguise. And the tokens and apps themselves could be part of the “data washing.”
I can’t shake the feeling. The “audience” is AI training. And this is about to be the cheapest and fastest acquisition of mass IP in human history.
But aren’t there actors beginning to break out and go viral in this medium. Sure! That’s how social media already works - heck, that’s how celebrity has always worked. But that doesn’t reveal anything about the business. Tom Cruise is jumping out of airplanes and we still can’t get people to go to the movies.
One of the biggest problems with the tech industry is that every product misrepresents itself as “new” and “innovative.” But most are just updated applications to systems and problems we already solved. (I could do a whole separate post on how most of A.I. buzz is just re-purposing of cloud computing and machine learning we’ve had for decades).
But audiences have over a decade with digital media engagement:
We know social media numbers can be “heated.”
We know “active users” have been misreported by tech companies.
We know viewership and revenue has been on the honor system any time a new streaming service comes to town.
We know once money becomes “tokens” it’s not traceable.
So we know hype can sometimes just be… hype. And when it comes to AI and celebrity: We’ve already seen massive franchises like Star Wars use replication posthumously to create new products with actor likeness previously captured for use in new movies, video games, merchandise.
It’s not implausible. And ask yourself…
Would you rather spend years negotiating likeness rights from SAG actors…
…or get thousands of people to sign away everything for a few hundred bucks?
Voice. Face. Body. Performance.
In perpetuity.
They don’t need “hits.” They need volume. Traditional film has focused on building *story* IP but the IP that has always hindered Hollywood has been *likeness.* The biometric data of celebrities is the holy grail - especially for A.I. - when we look at the “highlights” of generative content from the last year:
The formula for the 2025 internet seems to be Credibility + Personal Library = Monetization.
Now what if you could own both sides of that equation? To create “stars” you owned and could control forever? Because once the model is trained…
It can replace the actors.
It can recreate the scenes.
It can write, perform, and deliver new content—without humans.
Furthermore the dirty secret of AI systems is they need a steady stream of new REFINING DATA to keep training the module or it becomes stale and regurgitive.
This theory would supply that volume. It establishes that pipeline. And it builds that credibility.
Again, it’s not implausible. And ask yourself…
Is it more or less plausible than millions of people you don’t know spending seemingly limitless money on swipe-able vignettes you think are slop but that apparently cannot be made fast enough to keep up with the demand?
Now go a step further…
This content is made for these companies’ proprietary apps. They control the views, monetization, performance, uploading, backend… everything.
Using very simple automation, A.I. could be simulating user behavior: Watching scenes…Unlocking tokens. Creating fake engagement that trains the system while boosting hype.
This already happens at engagement farms where people build views, likes, comments.
This already happens with automation of bot accounts posting and engaging.
This can already be done with early “AI agents” or GPTs/APIs that allow the AI perform multiple recurring task and assess its own result according to specific parameters.
It’s actually an elegant and achievable solution for the “views” to track the “instancing” of it “training” on the episodes. And it would be a really clever way to build an internal, private ledger of the training and tagging pipeline.
Have you heard of ‘The Dead Internet’ theory? AI accounts talking and cross posting with other AI accounts at such a high volume it makes human interaction nearly impossible and imperceivable.
Automation like this would turn this theory into a unstoppable business model for data and money laundering.
So maybe they’re not making “bad” scripts by accident. They’re just feeding the machine.
And the audience? Is the machine.
If they use A.I. to write the scripts and match it to the final filmed episodes, it would even more elegantly scale this model. Once companies set a standard for lower quality of camera, plot, dialogue, performance, etc… the bar is lower to generate a matching result. With current text-to-video solutions, trained on enough of these episodes, it could be relatively easy to generate a bonus episode or spin-off series.
Right now the criticism of AI video characters is the stoic, robotic lack of expression. So the solution would be to train a library on the opposite with tagged stock emotions, character and scenarios.
We’ve been here before. It’s Commedia dell'arte for data farms. It’s the original Studio System for virtual influencers.
And look… if it’s not profitable or sustainable they will run out money (most tech companies do). And what happens then? Well, they do what every Hollywood studio is doing… They sell the library.
Except this library now includes a refined generative model of pseudo-celebrities whose digital replicas and personas are completely owned and operated by the company. They make money on both the IP and the underlying technology to generate from that IP.
I know this is starting to sound like science fiction, but let’s assume you wanted to use this method for data harvesting. Here’s how I’d design it…
The vertical shorts series launches on the app.
The app shows: “1,000 views and 300 coin unlocks!”
But what if those aren’t people?
What if that activity is a training model disguised as an audience? So “1,000 views” is a ledger of the ingestion, the number of times it’s fed the video into the system. Meanwhile “300 coin unlocks” tracks the successful returns of tagging and training. More coins tells they system it’s a more refined model.
Suddenly, the same act of AI ingestion becomes a gamified creator economy. And this doesn’t mean there are not human views as well. Plenty of social media influencers buy followers and views while still having authentic engagement from real people. But this form of botting could hide the true intention of the platform, let the data refine itself, and keep a ledger of the best quality outputs.
To the outside eye, it’s just “growth” and “easy money” (without royalties, unions, or consent). But under the hood…
This is data laundering.
It hides the true source and purpose of each interaction:
“View” = AI observation
“Like” or “Comment” = Internal Tagging
“Unlock” = Rewarding valuable data
“Earnings” = Internal tokens, perhaps even to quantify how much training has been invested on a particular clip.
Also remember, the app doesn’t need to pay creators real money. It pays in tokens it created. That, like TikTok, creators have to figure out how to withdraw only after reaching a certain threshold.
Meanwhile, their face, voice, and choices are being ingested by models to build:
Performative AI
Emotion synthesis
Synthetic dialogue
Virtual avatars / Digital Replicas
Because raw scraped data (like YouTube or Twitter) increasingly comes with legal baggage. But if a creator signs away their own work, through TOS and nonunion agreements… The platform gets clean, labeled, perpetual-use data…on the cheap.
And they can order more volume of their most performant stories and characters. It’s not about stories. It’s opt-in, indemnified, monetized AI training (without calling it that).
And once you’re in the system, the model is learning from you:
Your rhythms. Your timing. Your expressions.
Forever.
What if this isn’t a creator economy.
What if it’s a synthetic engagement shell wrapped around an AI ingestion pipeline.
If I was building this, I wouldn’t need your fans. I wouldn’t need your ideas.
I’d want your face, your voice, and patterns so I controlled a you who doesn’t need you (or your permission).
Meanwhile all your own online posting and fame generates even more value for MY version of you. It doesn’t stop you from growing your own brand. It’s just free marketing for mine.
Not because anyone was tricked.
Because these companies moved fast, under the radar, with plausible hype and cheap contracts during a time of desperation within the industry.
And whether you agree with me or not… The theory that A.I. companies could be mass commissioning a library of performance data is as plausible as a sudden flood of paying audiences for underproduced content during the threat of a global recession.
The beauty of a non-union contract is that it’s - by design - not an “industry standard.” It’s just a napkin with scribbles for everyone to decide what works… for everyone.
These companies notoriously will avoid negotiating and threaten to find “someone else.” But when they say, “a dime a dozen” …show me the dozen. Breakdown Services is flooded with these projects right now so I suspect they still need a high volume of TRAINING data. And they are willing to negotiate for recurring “stars” who have appeared in other Vertical Dramas (likely because those are proven personas that can provide even more valuable REFINING data).
This is just negotiation 101. Hold out for what you want and know your worth.
Yes, there are many creative people who want to make movies… for a living. There is ultimately a finite number of people who have the ability, geography, and flexibility to work on low budget projects - especially when filmed with little-to-no notice in a specific region. Now take into account the number of moderately talented and compelling performers who can memorize 15+ pages of dialogue per day and have that on-camera “it” factor… There just won’t always be “someone else.”
Then take into account that if they can produce cheap looking feature films in 7 days and sell it for $20 per viewer to an untapped audience of millions… so can you. So many tradespeople defer compensation (or even invest their money) in career building experiences. If I’m wrong and what these companies are really buying is content… you can make content and sell it to them.
I’m not here to shame or stop anyone from working in a speculative, emerging industry. These types of opportunities are literally how I’ve made my career. And I am looking for the right project/partners to get involved in this trend.
But I beg you to do it with your eyes open and know your rights. In fact AB 2602 included legislation that an individual could not sign away their rights to A.I. without including a lawyer or collective bargaining entity.
Asking for Most Favored Nations (MFN) to be compensated equally with your peers.
Asking for A.I. protections.
Asking for profit participation (residuals) and accurate reporting.
Asking to be tagged / cross-posted / verified to grow your own socials.
Asking to pitch your own scripts and collaborators.
Asking for First Right of Refusal/Negotiation on derivatives, sequels, spin-offs.
Asking to work under SAG-AFTRA’s contracts or waivers.
Asking to work with WGA, IATSE, Teamster peers.
If there’s money and an audience… we should be able to see it.
And if there is something nefarious happening behind the scenes, you should get as much value out of the exchange as possible. If I’m right, A.I. companies historically light a lot of money on fire… you might as well get some warmth.
If I’m wrong, it’s still a good idea to conduct your business… like a business; and ask for these protections.
And if you’re a producer reading this and worry your talent might be empowered to negotiate better terms, remember…
The value of a creative property is derived entirely from the combination of its individual components. By championing the rights of your team, you increase the pie…so your slice will also be bigger. If your business thesis is, “Do it for as cheap as possible,” the easiest salary to cut it your own - because your labor is not on screen.
Let’s protect each other. Rather than racing to the bottom, we should be pushing everyone upwards.
After all, aren’t these called Verticals.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.