We Talked to Bryn Mooser About Moonvalley's Marey

In partnership with

Welcome to VP Land! Moonvalley's new AI video generation platform, Marey, has officially debuted, built specifically for filmmakers with impressive camera control tools and trained on licensed data.

We got to chat with Bryn Mooser, one of the co-founders of Moonvalley and Asteria, about their approach, his thoughts on AI’s role in Hollywood's future, and how they're using AI to help make Natasha Lyonne's directorial debut, Uncanny Valley, happen.

We've got the highlights of our conversation below, but check out the full videos (part 1 and part 2)

In the last edition, we asked for your thoughts on Freepik’s unlimited AI image generation, and most of you saw it as a green light to ramp up AI content (they’re also rolling out weekly unlimited video generation for different models). Check out today’s poll below.

In today's edition:

Moonvalley's 'ethical AI' model Marey launches
YouTube cracks down on repetitive video content
LTX Video launches 3 new LoRAs for precise AI control
Luma AI announces creative lab opening in LA

Moonvalley Launches Hollywood-Ready AI Model

Moonvalley has officially launched Marey, its AI video generation platform designed specifically for professional filmmakers and built exclusively on licensed data. The platform promises to address Hollywood's biggest concerns about AI video tools by offering frame-perfect control and commercially safe outputs.

VP Land recently sat down with Bryn Mooser, co-founder of Moonvalley, to discuss Marey, how he's adapting existing tools for Hollywood, and why he believes studio-quality filmmaking is about to open up to indie creators.

Ethical positioning sets Marey apart from competitors by using only licensed training data, eliminating copyright concerns that have caused hesitation for broader adoption of AI models in commercial projects.
The platform offers 3D-aware video generation that allows filmmakers to modify camera angles, character movement, and backgrounds after initial creation.
Advanced control features include pose transfer, camera control, motion transfer, and trajectory control, giving directors the same level of intentionality they expect from on-set production or high-end VFX studios.
At a $14.99 starting price point, Marey aims to make cinematic storytelling by making studio-grade AI video tools accessible to independent creators and small studios.
On launch, native ComfyUI API nodes were available to create workflows for both text-to-video and image-to-video generation
Industry feedback has been positive, with filmmaker Paul Trillo noting the platform gives him the precision control he's always wanted when directing AI-generated scenes (he also directed the launch video)

SPONSOR MESSAGE

Learn AI in 5 minutes a day

This is the easiest way for a busy person wanting to learn AI in as little time as possible:

Sign up for The Rundown AI newsletter
They send you 5-minute email updates on the latest AI news and how to use it
You learn how to become 2x more productive by leveraging AI

YouTube Clarifies AI Monetization Policy Changes

YouTube moves to demonetize "spammy AI slop" starting July 15, 2025, but clarifies that legitimate AI-assisted content creation remains monetization-eligible. The platform seeks to address creator confusion following backlash over policy update announcements that many interpreted as a broader crackdown on AI-generated videos.

YouTube emphasizes these aren't entirely new rules but rather improved enforcement of existing requirements for "original and authentic" content, with enhanced AI detection systems to identify mass-produced and repetitive uploads.
The policy specifically targets content that combines stolen clips with AI-generated voiceovers, reaction videos without meaningful commentary, and entire channels dedicated to pushing out lazily made AI spam at scale.
Creators using AI tools to enhance their content production—such as for voiceovers, scripts, or editing—remain eligible for monetization if they add clear value, commentary, or creative input to their videos.
YouTube's new detection systems can identify repeated use of popular AI voice models and flag video patterns characteristic of automated mass production, even when content is "repackaged" with new titles and thumbnails.
The enforcement approach takes an all-or-nothing stance, potentially removing monetization from entire channels rather than individual videos when violations are detected.

Form Studios Launches Virtual Film Challenge

Form Studios has launched the Virtual Film Challenge (VFC), a short film contest that encourages creators to showcase virtual production techniques in 60-second films. The contest opens for submissions on August 1, 2025, with a deadline of August 31, 2025.

Contest structure allows filmmakers from professionals to students to compete using any virtual production tools including green screens, motion capture, previsualization, and virtual cameras.
Form Studios brings advanced technology to the competition, including their SISU motion control robot that enables precise camera movements crucial for virtual environments and VFX composites.
The initiative positions Atlanta as a growing virtual production hub, complementing the region's existing reputation for film and television production.
Winners will be selected based on creativity, storytelling, technical skill, and effective use of virtual production methods, with results announced on September 26, 2025.
The contest partnership with Virtual Production Insider creates a platform for democratizing access to industry-grade virtual production workflows beyond traditional Hollywood circles.

LTX Studio Releases Three Control LoRAs

LTX Studio announced three new LoRAs for their AI video generation platform: Pose, Depth, and Canny LoRAs. These control modules give creators precise over character movement, spatial relationships, and visual clarity in AI-generated videos.

Pose LoRA captures body skeleton data from reference videos, allowing creators to choreograph precise character movements, dance sequences, and complex multi-character interactions rather than relying on random AI-generated motion.
Depth LoRA maintains consistent 3D spatial relationships throughout video sequences, ensuring foreground and background elements scale correctly during camera movements like dollies, pans, and zooms.
Canny LoRA extracts high-fidelity edge details using Canny edge detection, enabling sharp style translation and accurate preservation of facial features, outlines, and structural elements.
The platform released open-source training code alongside these LoRAs, allowing researchers and developers to train custom control modules on consumer-grade hardware and experiment with novel control types.
LTX Studio's LTX-Video model generates videos at 1216×704 resolution at 30fps using a DiT-based architecture, positioning it as one of the most accessible and customizable AI video platforms available.

Bryn Mooser explains how Natasha Lyonne’s $70M sci-fi epic Uncanny Valley is being made for just $10M using AI - and why the future of filmmaking lies in custom models, financial innovation, and creator ownership.

Stories, projects, and links that caught our attention from around the web:

🎨 Wand’s latest upgrade to its rendering abilities transforms rough drawings into detailed visuals.

🎥 Vidu AI has launched an upgraded Reference-to-Video feature that allows users to generate videos by uploading up to seven reference images.

🤖 Grok 4 was announced, introducing two new reasoning-only AI models, Grok 4 and Grok 4 Heavy, which delivered record-breaking benchmarks and major performance gains.

🚀 Film producer Andrea Iervolino has hired NASA veteran Bert Ulrich to spearhead his vision of making feature films shot in outer space.

🎬 Luma AI has announced the opening of a new lab in Hollywood, run by Verena Puhm and John Fingers