So…you know if you’re getting a mid-week AI For Humans newsletter there’s something big happening.

Last night, I got hands on with Seedance 2.0 and reader, I am gobsmacked.

When I woke up this morning, I had a revelation about Bytedance’s new state-of-the-art (truly) AI video model Seedance 2.0 and realized I should at least make sure that y’all are aware of what this could be.

It’s not just that it’s exceptionally good quality, multi-cut, 15 second AI video clips that have that same sort of creative juice that Sora 2 did on arrival.

It’s that, like Sora 2 when it first arrived, there are no IP restrictions at all and, because this comes from China, we may never see them.

That is a massive thing and has huge implications for the creator economy, creative agency work and, most of all, Hollywood.

What happens next?

Please support AI For Humans by learning about our sponsors below:

Great ads don’t happen by accident. And in a world flooded with AI-generated content, the difference between “nice idea” and “real impact” matters more than ever.

Join award-winning creative strategist Babak Behrad and Neurons CEO Thomas Z. Ramsøy for a practical, science-backed webinar on what actually drives performance in modern advertising.

They’ll break down how top campaigns earn attention, stick in your target’s memory, and build brands people remember.

You’ll see how to:

In 2026, you have to earn attention. This webinar will show you exactly how to do it.

Save you spot

Here are the simple facts (right now) about Seedance 2.0.

My basic takeaway is that this model does the sorts of stuff that others promise. It makes creative choices that you’d like to see. It figures out vague prompts in ways that are both surprising and kind of shocking. It handles native audio generation better than any model other than Sora 2.

And the consistency and quality of what is possible in a prompt has never been crazier.

Watch this example from Ethan Mollick and tell me that doesn’t stand up very close to TV production quality.

I called Seedance 2.0 the ‘Deep Seek’ moment for AI video because, in case you don’t remember, last year around this time the stock market plunged due to the launch of the Chinese AI model and proved that America wasn’t alone in the AI race.

The timing of this release and last year’s drop are not inconsequential. Chinese companies often prepare big release right around Chinese New Year and that’s about a week away. They know a ton of Chinese people will be on their phones during the holidays and ready to generate all sorts of stuff.

Deepseek shocked the world (and the larger frontier AI companies) with a model that was very close to the cutting edge and proving that it was possible to train a top-shelf LLM on the lesser AI chips that China has official access to.

This year, with Seedance 2.0, I think the impact could be even bigger.

Just like with the original release of Sora 2, it’s clear that Seedance 2.0 was trained on all video. Like everything including all Hollywood content.

Famous celebrities, huge IP and many, many more things that are owned by other people reside within the Seedance 2 model.

Want Sylvester Stallone as Rocky working a fast food job with Optimus Prime? Seedance 2 will make it for you.

This is going to be a huge deal when clips like this start circulating in Hollywood.

Remember, it only took two days for Sam Altman to back down off of what was prompt-able in Sora 2 because of the uproar it caused in Hollywood writ large. And, of course, because American IP laws are generally pretty strong the studios and unions had a good argument.

But Seedance 2.0 is a Chinese model. They run on a very different sort of IP structure and large American companies, particularly Hollywood studios, don’t have nearly the influence there they used to.

Love this newsletter? Forward it to one curious friend. They can join in one click.

There will be panic both in Hollywood and on Wall Street. It’s entirely possible that this model hits Disney and Netflix stock the same way that the Deepseek panic hit NVIDIA.

But outside of the panic, the larger question looms:

Is IP protection over? What about personal likeness rights? How can you corral this if the company that makes it doesn’t have an incentive to do so?

I don’t know. But I bet there are a ton of lawyers waking up to ALL CAPS emails this morning.

The good news here is that there’s a new state-of-the-art video model coming soon for everyone to use. You’ll be able to make nearly whatever your heart desires.

I enjoyed this longer piece from PJ Ace (friend of the pod and ultimate AI video guru) about how the Seedance 2.0 moment means it’s going to be less about single prompts and more about what truly creative people can put together.

To me, this moment is about new IP and what people might create if given these sorts of tools. The toolbox just got a lot bigger and I’m excited to see what people will make.

But, of course, that doesn’t stop me from fretting about the future of what I’ve done my whole life either. Hollywood… you may not be cooked but the water is getting hotter.

See you Friday for a new episode!

-Gavin

One small caveat here: With every new AI model, there are bad actors who use their exceptional AI and editing skills to make it look like something is an output of the model when it might not be.

For instance, I’m not sure this Brad Pitt / Tom Cruise clip is a Seedance 2.0 output but if it is… good lord.

Below, I’ve collected some of my favorite Seedance 2.0 examples so far.