The New Rules of Hollywood AI: Jonathan Yunger on Billion-Dollar Slates, Streamer-Approved AI, Ar...

Executive summary

Jonathan Yunger, president of Millennium Media and co-founder of Arcana Labs, argues that generative AI can help independent filmmakers maintain production value as traditional financing shrinks—provided it is used as an artist-directed production tool, not as a substitute for filmmakers, performers, or practical filmmaking.

Arcana is positioning itself as a secure, enterprise-focused B2B platform for film and media companies. Millennium, meanwhile, is pursuing affordable, high-concept genre projects that can be packaged and sold in a tighter independent market.

Business strategy and operating model

• Arcana Labs: enterprise B2B rather than consumer-first. The platform brings generative AI tools together, fine-tunes models for cinematic work, and has direct relationships with model providers. Yunger said the company is moving away from direct-to-consumer offerings and focusing on enterprise and B2B use.
• Privacy and compliance as a differentiator. Arcana’s enterprise environment uses isolated “sandboxes,” where a client’s production data is kept separate rather than fed into public models. Yunger said major streamers reviewed the system through intensive security processes lasting roughly five to six months, with review teams of around 40–50 people.
• Human-led creative process. The company describes its approach as “artist-driven AI”: filmmakers supply the ideas, source material, and creative direction; AI assists with selected tasks. Arcana says it does not replace actors, and Yunger favors traditional picture and sound post-production.
• Education as an adoption strategy. Arcana developed an academy to train artists and organizations, then expanded into film schools, universities, and secondary schools. Its stated curriculum starts with eight weeks of film theory before students use the tools.
• Millennium’s content strategy. Yunger said the company is looking for fresh, high-concept genre projects that can be made at controlled costs, packaged with one or two recognizable names, and often shot at its studio in Bulgaria. He mentioned a newly acquired sci-fi concept and a potential Aaron Eckhart project as examples.
• Practical production remains part of the proposition. Yunger emphasized that AI is not suitable for every film. He said John Rambo was shot 99% in camera, with practical action chosen because it best matched audience expectations for the franchise. Lionsgate acquired worldwide rights, with a theatrical release expected.

Frameworks and playbooks

Use AI selectively across the production pipeline

• Apply it in development and pre-production for concept exploration, storyboards, visual references, and pitch materials.
• Use it in production or post for appropriate inserts, establishing shots, background elements, or pickups that are difficult or costly to capture.
• Preserve live-action production and human creative work where they are central to the film, such as performances, key action, or emotionally important scenes.

Other production and financing approaches

• Reinvest production savings in creative priorities. On a constrained schedule, using AI for some scope-building shots could free directors to spend more time on performances, scares, or other essential scenes.
• Build in a safety net for production. Yunger’s example: when a location becomes unavailable or a needed shot is missed, production data may help create a pickup later. The goal is to make creative risk more manageable—not to eliminate traditional filmmaking.
• Lower costs to make risk more financeable. If a project costs less, a distributor may be willing to acquire an unconventional film for a lower price and still support it with marketing.
• Combine traditional film craft with iterative workflows. Yunger described production boundaries as increasingly fluid: development, production, and post can overlap, allowing teams to revise and refine ideas more quickly.
• Use short-form work to test demand. The hosts and Yunger discussed releasing shorts or episodes online, observing audience response, and using traction as evidence for a feature or series.
• Teach craft before tool use. Arcana’s education approach emphasizes storytelling and film theory first, on the premise that access to tools alone does not create strong work.

Metrics and figures cited

• Yunger corrected the host’s introduction to say the films he has produced have grossed more than $5 billion worldwide.
• He said independent-film minimum guarantees have fallen over roughly the past seven years, as distributors became less certain about revenues after streamers took control of the pay-one window.
• On Lucky Strike, an early Arcana use case helped with production-design iterations and period/family photographs. Yunger estimated savings of about $50,000 on a $7.5–8 million film.
• Yunger said some Arcana users produce 30–40 broadcast commercials per month, while others generate roughly 600–700 content pieces per week.
• He cited an online gaming company that hired 60 in-house AI artists, presenting the role as an emerging job category.
• He said Arcana expected to be in more than a dozen schools by August or September.
• For a hypothetical $1 million horror film with a 17–18-day shoot, AI-supported scope shots could potentially help the finished film feel closer to a $7–8 million production. This was an illustrative example, not a reported result.
• Yunger said a distributor might be more willing to take a chance at an acquisition price of $700,000–$800,000 rather than $2 million if lower production costs reduce the amount it needs to recoup.
• He estimated the creator economy at around $350 billion annually, compared with Hollywood at roughly $150–160 billion. These were broad figures cited in conversation, not independently substantiated in the episode.

Concrete examples and recommendations

• Production design: AI helped generate iterations and photographs needed for a film when actors arrived too late for a conventional photo shoot, saving time and money.
• Low-budget genre filmmaking: For a tightly scheduled horror film, AI-generated or AI-assisted inserts and establishing material could reduce pressure on the shoot while allowing more time for performances and scares.
• Pitching and proof of concept: Director Phil Hawkins described using AI to create sizzle reels for projects whose tone or visual concept was hard to communicate through references alone. He saw this as useful for pitches, not as a move toward making AI-only features.
• Access for emerging creators: Yunger described helping a young artist in India, who lacked a camera budget, create a horror short using his own writing and sketches. The artist later found production-company work and supported his family.
• Educational entrepreneurship: Arcana discussed a school model where students learn media and marketing skills by running an internal advertising agency and offering services to local small businesses.
• Practical recommendation: Yunger’s advice to aspiring filmmakers was to make work, use available channels such as social media to find audiences, and build craft rather than wait for permission or financing.

Sources and presenters

• Guest: Jonathan Yunger, president of Millennium Media and co-founder of Arcana Labs.
• Presenters/interviewers: Giles Alderson and Phil Hawkins.
• Source: The Filmmakers Podcast.