In Is Traditional VFX Dying?ErikDoesVFX went head to head with Higgsfield, pitting modern VFX against GAIV. His wife claims victory, but then he reveals Higgsfield’s second, more creative version:
My take: the genie is out of the bottle and GAIV is just another tool that creators now have in their toolkit. I predict Hollywood will quickly become its biggest user simply because their schedules and budgets will render traditional VFX unfeasible.
Of course, Tim also posted another video all about his production pipeline. He implores you to:
“Don’t buy (or sub to) an AI film studio — build one! This is the future workflow of AI filmmaking: a complete production office that lives in one folder on your desktop, and you can download the whole thing free.”
The production office model uses an LLM, an MCP and martini.film to organize the project, cast it, storyboard it and generate shots. The filmmaker in Tim then edits the footage and mixes the sound and music traditionally.
The main difference is that this method is more hands on and avoids MCP automation.
Note, both pipelines work best with a great idea — and script — first!
My take: this protocol is the state of the art for AI-generated content in June 2026: use Claude (with a Skill) to design the look of the film, the characters, prompts, locations and storyboards. Use these images to generate shots in Seedance. Edit as usual. I predict that as more and more folks discover this method, the quality of your ideas, script and editing will become much more important.
“The 95-minute feature cost under $500,000, with most of the budget going on computer costs. The first 25-minute segment required 16,181 video generations to produce 253 final shots, a 64:1 curation ratio. The film was produced by a team of 15 directors, cinematographers and editors, most of them working in person from Almaty, Kazakhstan with some collaborators joining remotely. Alex Mashrabov, CEO and co-founder of Higgsfield, explained that the execution phase took around two weeks, although the underlying idea and script had been developed by the filmmakers over several years.”
Back to the WSJ for some technical details:
“Every prompt had to be extremely long and detailed. Each one would typically start with a prefix that defined requirements like style (8k IMAX, photorealistic), lighting (natural light only, “contre-jour” backlight, camera on shadow side) and the type of camera it should look like it was being shot on (“cine lens,” 180-degree shutter motion blur). The lighting was key to avoiding the AI sheen that typically gets branded as “slop”. AI-generated video tends to over-light scenes in an unnatural way. That prefix would also have to remind the AI to obey the laws of physics with wording like: “gravity and inertia respected—mass has real weight, correct contact shadows, no floating props.” The individual prompts were, on average, 3,000 words each.”
“disruptive toolkit designed to level the playing field for creators who don’t have Hollywood-sized bank accounts.”
The nine strategies, from the article, are:
Training your “Digital Eye”: Cinematography Knowledge Still Matters
Because AI doesn’t know the emotional difference between a high-angle and a low-angle shot your knowledge about filmmaking is your greatest asset.
The “Micro-Problem” Strategy: Build Your Own Pipeline
Look at AI as a series of specialized assistants, rather than a single “creative partner”, and use it to solve niche problems such as Audio Issues, Localization and VFX Generative Fill.
AI-Native vs. AI-Assisted: The Gap in Decision-Making
The panel highlighted the growing divide between “AI Native” creators (who often lack traditional film training) and “Filmmakers using AI.” Filmmakers don’t just prompt; they curate.
Personalization vs. The Shared Experience
One panelist floated “Dynamic Content“; the idea that a screen could generate a version of a film tailored to the viewer’s preferences (e.g., skipping gore for a sensitive viewer or emphasizing a certain subplot).
The “Dishes” Principle: Using Tech to Buy Back Time
The ultimate goal for the indie filmmaker isn’t to let AI do the “fun” part (the writing and directing). It’s to let it do the “chores” like organizing a scene list, budgeting with predictive models, or building a character-tracking app.
The “15-Minute Wall”: Understanding AI’s Memory Problem
One of the most practical “craft” takeaways for filmmakers is the current technical limitation — AI models have a “short-term memory” problem. So don’t try to “generate” your whole movie yet. Use AI for shots, textures, or cleanup, but rely on your principal photography to maintain the “scaffolding” and continuity of your characters.
The “Signal Through the Noise”: Curation as a Creative Act
With the barrier to entry dropping, the sheer volume of “content” is exploding and the role of the filmmaker shifts toward being a human gatekeeper. Your job is now curating.
Directing for “Dynamic Viewing”
Personalized Exhibition: AI could allow for versions of a film that adapt to the viewer so when shooting, consider capturing “excess” coverage.
AI as Your “Marketing & Advice” Department AI can act as your business consultant or Creative Producer. Feed it your script and ask: “What is the most marketable 30-second hook for a Gen Z audience?” or “What film festivals have a history of programming films with these specific themes?”
The article concludes with a warning:
“The intersection of AI and indie film isn’t about the technology—it’s about curiosity over fear. If we leave the tools to Silicon Valley, they will build a “push-button” industry. If we engage with it as filmmakers, we can build a more accessible, disruptive, and human-led future.”
My take: art is always an abstraction of reality. Cinema, and by extension TV and screen-based media, use a vocabulary and grammar that is barely 130-years old. I predict most viewers will accept the new tools in time, like they did animation and Computer Generated Imagery.
He used Claude Cowork as his “Production Office” (subscribing to the Pro plan). The key here is to grant it access to one folder on your computer and give it instructions in a .md markdown file.
He used Luma’s Agent Canvas as his “Studio”. Watch the masterclass to see his process.
And sign up for his newsletter to download the SKILL.md instructions markdown file.
My take: wow! Such a great short film! Such a generous man! Way to go, Tim!
“InterPositive’s mission — to use emerging technology in ways that protect and expand creative choice — is deeply aligned with Netflix’s long-standing belief that innovation should serve storytellers and the creative process.”
“The InterPositive system builds an AI model based on an existing production’s dailies, then lets a filmmaker introduce that model into the postproduction process to provide the ability to do things like mix and color, relight shots, and add visual effects.”
Note that InterPositive owns a patent on technology that Ben invented called “Method, system, and computer-readable medium for training a captioner model to generate captions for video content by analyzing and predicting cinematic elements”. It describes systems designed for enhancing video content analysis and generation through cinematic element recognition and metadata utilization.
Price has not been disclosed.
My take: is this the beginning of Netflix turning into a “dream factory”? Imagine sitting down on the couch and prompting the movie you’d like to see. Or a spin-off with some of your favourite characters. Or — and I want this so much — “Yeah, this movie, but make it 90 minutes instead of two hours and forty-five minutes.”
StoryPeer.com is a new platform developed by Gabriel Dimilo to help writers get free feedback on their work by giving feedback on other writers’ works anonymously.
The platform uses tokens to facilitate reviews: offer them to other writers to give notes and hence earn them by reviewing work.
Everyone starts off with seven tokens.
Readers will pick your project based on its title, logline, genre, and length. And how many tokens you offer. A rule of thumb I suggest is to offer one token for every 30 pages.
My take: beware, this can become addictive! I’ve already reviewed two features and received notes on one of my shorts.
Text to Video (Google Veo 3.1 can even voice dialogue.)
Image to Video (Probably the best way to ensure consistent characters.)
Video to Video (More work, but worth it!)
Lip Sync (The weakest link IMHO.)
Ingredients to Video (This hints at the future.)
Chat Edit (Sort of combines Video to Video and Text Prompting.)
Tao’s insights are very educational!
My take: really nice to have this summary of the various approaches. Note that you most definitely will use some combination of each; don’t just fixate on one tool.
