If you were online yesterday and you have any presence in Hollywood, someone probably texted you a viral AI clip of Tom Cruise and Brad Pitt fighting on a rooftop.
This sparked a firestorm at X, with discussions ranging from “it’s over” to “this is name and likeness infringement” to “let me know when you can do it at scale.”
And everything in between.
Let’s dive in.
Viral fight (in more ways than one)
The clip originated from a Chinese company and was tweeted by filmmaker Ruairi Robinson.
The 15-second clip depicts two A-list stars exchanging blows in a desolate concrete environment. The footage isn’t a deepfake in the traditional sense of pasting faces onto stuntmen, but rather a fully generated sequence of two famous actors fighting.
Again, since this is from a Chinese company, Brad and Tom can’t easily sue. Because if an American company did this, it would be illegal and there would be a lot of backlash and precedent for it being violated.
So is this a big deal?
Give credit where credit is due
Despite the Uncanny Valley elements, the fluidity and lighting of the fight sequences are surprisingly close to the quality of a big-budget studio action movie.
It’s great that multiple angles are available and the wide range of cameras is dynamic.
But 15 seconds is too short to be anything more than a proof of concept that shows how rapidly generative video technology is evolving.
It’s interesting to watch technology grow, but I have to reiterate that 15 seconds isn’t all that impressive. I’m nearby Not the quality we expect There is The quality we expect. And while it’s interesting to see technology doing 90% of the work, with a lot of AI, closing that last 10% gap is actually quite difficult.
I have never seen AI scale as fast as predicted. Also, the computing power that would be needed to scale AI is very expensive to obtain.
What you posted was just a reel you were cool scrolling in the bathroom.
Pleasuring your ass in the bathroom is fine, but it’s a far cry from having to sit in a theater seat.
This technology has a long way to go to make that happen.
So it’s not over yet?
If I worked at the Post, I would be concerned when I see videos like this. The idea of using two sentences to generate a fight scene between stars is alarming.
But when you break it down, this is two celebrities dressed all in black fighting on a boring rooftop in a nondescript city.
My suspicion is that if you bring in something like naturalistic costumes or easily identifiable landmarks and have to scale it to a theatrical quality, something like this falls apart.
Also, the combat felt more like TV than a movie. That means it’s not professionally choreographed and has too much editing.
And…it also looked like they were rolling with the punches, which is kind of funny when there aren’t any actual people involved.
Again, it seems like it’s that last 10% of things that really matter to the audience that they get it right, and it takes hundreds of millions of dollars to get even close to that.
This is a technology worth looking at, and probably worth learning more about, but I’m not sure I believe it will be adopted any time soon and take away a ton of jobs.
bring it all together
These are just my thoughts. It was interesting to see people fussing over it, but again, I’m not going to sweat it until I see a story that lasts at least five minutes and action that doesn’t feel like someone cut it with some clean 80’s coke.
Let us know what you think in the comments.
