Why Moonvalley's Approach to AI Video Makes Sense | SOM Podcast

Most AI video companies are gambling with copyright law. Moonvalley made a different bet.

Real talk: Most AI video companies are doing it wrong.

They scrape everything they can get their hands on – YouTube, Vimeo, whatever – and feed it into their models. The results? Tools like Sora and Gemini are genuinely impressive. The quality is getting scary good. But here's the rub: Video producers are using these tools to create content that's trained on copyrighted work, while the original artists who made it all possible get zero credit or compensation. It's basically profiting off someone else's creativity without asking permission.

But what if there was a better way? What if AI could actually make creators more powerful instead of making them obsolete?

That's exactly what Naeem Talukdar, founder of Moonvalley, is betting on. And after listening to him break down his approach on the latest School of Motion podcast, we're convinced they might be onto something big.

‍
⚡️ The Plot Twist: Paying Artists is Actually Good Business

Moonvalley doesn't scrape content. They license it. They're actually paying filmmakers and production companies for their footage. Wild idea, we know. 🙃

"What you're really paying for is curated data," Naeem explains, "and that is worth its weight in gold."

This isn't just about being the good guys (though that's pretty cool). It's about building better tools and staying on the right side of the law. While most AI companies are betting that training on unlicensed content counts as fair use, that's still legally untested territory. With major lawsuits pending – publishers vs. OpenAI, artists vs. Stability AI, and more – the verdict is still out. Moonvalley's approach sidesteps all that uncertainty entirely.

Plus, when you train on high-quality, ethically sourced footage, you get better results...

⚡️ 3D Thinking in a 2D World

While most AI video models are stuck thinking in 2D, Moonvalley is teaching their models to understand 3D space. They've built a simulation environment that teaches their AI how cameras actually work – not just what a zoom looks like, but how it affects depth, composition, and storytelling.

This isn't just tech wizardry for the sake of it. It's the difference between getting a random AI-generated video and having actual control over camera movements, lighting, and spatial relationships—the stuff that actually matters when you're trying to make something that doesn't look like a bad hallucination.

⚡️ The Real Talk About AI and Creativity

Naeem doesn't blow smoke about AI replacing human creativity. He's clear that AI will never replicate what makes a Spielberg scene iconic – that's human taste, vision, and artistry.

What AI can do is remove the barriers that prevent talented creators from realizing their vision. As Naeem puts it: "This will unlock a new golden age where the people with real taste and artistry will no longer be blocked by budgets."

⚡️ Why Motion Designers Should Care

Whether you're pushing pixels in After Effects or crafting the next viral animation, this conversation touches on something crucial: The future of our industry. Not the dystopian "AI will take our jobs" future, but the reality of how these tools will reshape creative work.

This is just scratching the surface. The full conversation covers everything from the technical challenges of real-time video generation to why Naeem thinks "AI" is actually a terrible term for what these tools do. (Spoiler: He prefers "rendering engine," and once you hear his explanation, you'll get it.)

Watch the full episode above to get the complete picture of where AI video is heading – and why it might be the most exciting creative technology in years.