I'm not a filmmaker. I'm just someone who pays close attention to AI and last weekend I ended up spending about 14 hours going down a rabbit hole of AI video generation tools, specifically Seedance. What started as curiosity turned into one of those 2am moments where you're staring at the ceiling thinking about something you can't turn off.
I started running some rough math. Game of Thrones cost somewhere between $6 and $15 million per episode at its peak. The production crew alone was enormous, hundreds of VFX artists, 170 named cast members, location shoots across six countries. The revenue that show generated across HBO subscriptions, merchandise, licensing deals, and syndication rights has been estimated at over $10 billion over its lifetime.
That $10 billion was distributed across thousands of people. Unions, studios, distributors, residuals, network deals.
Now I'm watching Seedance generate 10-second cinematic clips from text prompts. It's not perfect. The motion artifacts are visible if you're looking for them and the consistency over longer sequences still breaks down. But here's the thing, that's where it is today. These models don't plateau. They iterate every few months.
Two or three generations from now, what does this look like? A team of 10 to 20 people with a good story, a strong visual direction, and a few hundred thousand dollars instead of a few hundred million. The rights stay with them. The royalties stay with them. Every dollar the IP earns compounds back to the same small group.
Everyone building in AI right now is either making SaaS tools or foundation models. The opportunity that almost nobody is talking about is IP. Building the next Disney or the next MAPPA with a fraction of the infrastructure.
I don't know if I'm early or just wrong. But I genuinely cannot stop thinking about it.
Has anyone else been looking at where AI video generation goes for entertainment specifically?
[link] [comments]