The Sphere's Oz AI debate

In partnership with

Welcome to VP Land! The fight scene between Adam Driver and Scarlett Johansson in Marriage Story has found new life: The US Department of Agriculture is using the audio to scare off wolves...blasting the argument from drones with speakers.

In the last poll, we asked if you think people will be interested in generating their own TV shows, like Showrunner AI featured. Most of you said maybe, but mainly for niche or personal projects. Check out today’s poll below.

In today's edition:

Sphere's AI-powered Wizard of Oz adaptation sparks debate
Scenario launches AI tool for editable 3D models
Netflix scales up YouTube-originated content
Filmmaker shares AI tutorial for seamless sequence shots

Sphere's AI-Enhanced Wizard of Oz Ignites Industry Debate

There is an AI storm brewing in Oz - but it might be the usual knee-jerk, ill-informed internet mob overreaction.

Here's a quick rundown of how we got here:

We've already known that the Sphere is spending $80 million to adapt the 1939 Wizard of Oz, shot in 4:3 aspect ratio, for its immersive 16K theater, utilizing AI through a partnership with Google to assist with the transformation.
CBS Sunday Morning aired a behind-the-scenes feature that showed footage from the adaptation inside the Sphere for the first time (we shared this video last week).
The internet had some thoughts on how the classic was being modified and reframed to fit the immersive landscape, and it quickly turned into a meme format.

Some of the changes the CBS video featured:

Super Resolution: An AI model was built to generate super-resolution images of characters like Judy Garland, enhancing grainy close-ups.
Outpainting: AI (or AI assistance) extends the traditional frame of the movie to fill the Sphere's wide screen, showing more of the scene, such as the full scarecrow, yellow brick road, and mountains of Oz.
New Character Performances: AI generated new performances from original actors to ensure characters like Uncle Henry remain in the new wider, outpainted backgrounds, prompting these performances based on the script and directorial context.

But this wasn't all just 'AI': a thread on the VFX subreddit from some of the VFX artists who worked on the project noted that traditional visual effects work went into creating every scene, with many of them arguing the project relies more on conventional techniques than generative AI.

Our take: The angles shown in the CBS Sunday Morning special (super wide and high) likely did the project a disservice; the adaptation is meant to be seen closer to the floor, in a seat. The real verdict will be known when audiences get to check this out later this month.

What do you think? Share your thoughts in the poll later in the newsletter.

SPONSOR MESSAGE

Training cutting edge AI? Unlock the data advantage today.

If you’re building or fine-tuning generative AI models, this guide is your shortcut to smarter AI model training. Learn how Shutterstock’s multimodal datasets—grounded in measurable user behavior—can help you reduce legal risk, boost creative diversity, and improve model reliability.

Inside, you’ll uncover why scraped data and aesthetic proxies often fall short—and how to use clustering methods and semantic evaluation to refine your dataset and your outputs. Designed for AI leaders, product teams, and ML engineers, this guide walks through how to identify refinement-worthy data, align with generative preferences, and validate progress with confidence.

Whether you're optimizing alignment, output quality, or time-to-value, this playbook gives you a data advantage. Download the guide and train your models with data built for performance.

Download the guide

Scenario's Generate in Parts Transforms Image-to-3D Workflows

AI platform Scenario has launched Generate in Parts, a feature that automatically transforms any image into a structured 3D model with editable components in a single click. The announcement from CEO Emmanuel de Maistre highlights a step forward in the development of AI-supported 3D technology workflows.

The feature uses PartCrafter, the first open-source AI model that intelligently splits objects into multiple 3D parts like arms, wheels, and panels, each with clean geometry.
Unlike traditional monolithic AI meshes, Generate in Parts creates modular 3D assets that creators can animate, edit, or prepare for 3D printing without additional processing.
The tool addresses a major pain point in current image-to-3D workflows where AI-generated models typically arrive as single, uneditable objects that require manual segmentation.
This advancement aligns with the broader industry trend toward AI tools that enhance rather than replace creative workflows, giving artists more control over their digital assets.
The feature is currently available to Scenario users and represents a move toward more professional-grade AI-powered 3D creation tools.

YouTube Stars Get Netflix Production Upgrades

Netflix is increasingly acquiring YouTube-originated shows and scaling them up with bigger budgets, professional production teams, and global marketing campaigns. The streaming giant views YouTube as a "farm league" for identifying promising creator-led content that can be transformed into Netflix Originals.

Netflix Co-CEO Ted Sarandos described YouTube as useful for creators to "cut their teeth on" while noting that YouTube doesn't provide upfront funding, leaving creators to shoulder all production risks themselves.
The strategy has already produced notable transitions including reality series Inside from The Sidemen and dating show Pop The Balloon, with Netflix providing significantly larger budgets for enhanced production values and international exposure.
Netflix executive Jeff Gaspin emphasized that YouTube serves as a valuable development tool, allowing the platform to evaluate proven concepts rather than relying solely on paper pitches or sizzle reels.
The approach mirrors MrBeast's transition to Amazon's Beast Games, where YouTube content gets transformed into episodic series with production budgets reaching $250 million for enhanced storytelling and contestant development.
Early results show mixed success, with Inside averaging 17.5M hours viewed on Netflix compared to 15.4M YouTube views for its first episode, while Pop The Balloon Live struggled with only 1M hours viewed.

Creating Cinematic Sequences with Smooth Transitions Using AI

Filmmaker Benjamin Muñoz Alonso shared a tutorial demonstrating how to create seamless cinematic sequence shots using AI tools.

Step 1: Create a detailed close-up image. Generate a pre-shot and post-shot from it, making sure the movement between them is smooth.
Step 2: Design a pre-shot frame and add fluid movement. Add effects (like smoke) for more natural transitions.
Step 3: Use AI mistakes if they look good. Pick frames with clear, simple movement.
Step 4: Use the Extend feature to lengthen shots, but never more than twice to avoid losing quality.
Step 5: Let something move in front of the camera before the final frame for a dynamic transition.
Step 6: Use in-shot elements (like reflections) to make transitions more interesting.

VFX expert Alden Peters breaks down a complex shot from episode 3 of Friends of Sophia, where two characters overlook a sprawling cyberpunk cityscape—created using a blend of Blender, Lightcraft Jetset, After Effects, and iClone.

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

🎥 The new InScene LoRA model enables precise generation of new coverage shots directly from within an input image’s scene for consistent visual storytelling.

🛰️ Reddit user developed a FLUX Kontext LoRA that transforms low-resolution Google Earth screenshots into high-resolution drone-style photos.

🚀 Disney has announced the 2025 lineup of growth-stage startups chosen for its Accelerator program.

🤝 Animation studio LAIKA has joined the Academy Software Foundation to collaborate on open-source software initiatives for the film industry and media production.

⏱️ Meta recommends 20-40 minute "Goldilocks" session lengths for optimal time in VR headsets.

🎞️ Santa Monica has launched its own film office to streamline permitting and support film production services, separate from FilmLA.

Addy and Joey break down the latest AI tools changing post-production, including Runway Aleph’s text-based video editing, Luma’s Modify with Instructions, Wan 2.2, Ideogram’s character generator, Adobe Harmonize, and more.

Read the show notes or watch the full episode.

Watch/Listen & Subscribe

Spotify | Apple Podcast | YouTube