If we've come this far in a year, then the near future is gonna be insane.
A year ago, we all laugh-cried about the AI model ModelScope creating video of Will Smith eating pasta:
It was horrible.
Still pretty cool, but horrible. No one would touch it unless they wanted to give viewers/customers a nightmare.
Fast forward to today, with Open AI's release of Sora, the newest model of AI video generation. 👇
Yes, that was a video made entirely from a simple text prompt without any editing. INSANE.
Hundreds of the videos have already been made (it can currently generate up to 60 seconds in length) and most of them are impossible to differentiate from real cinematography.
This is one made by Open AI CEO Sam Altman for YouTube star Mr. Beast:
The tech isn't perfect yet. Some of the videos are downright weird. Take this cat waking its hooman up, for example (watch the woman's face and the cat's paws):
BUT LOOK AT THIS OCTOPUS FIGHTING A CRAB!
AND THESE WOOLY MAMMOTHS!
PIRATES OF THE CARIBBEAN: WHOLE BEAN ROAST
MY NEW FAVORITE PODCAST:
And yes, it can show people eating without you realizing it's all fake:
You know this tech is going to be used to make disturbing, evil content. It will be used for propaganda. It will be used for pornography. We've now entered the age where nothing in the digital realm can be trusted unless it has an insane amount of vetting (people were sharing real videos as AI videos and I had to dig to figure out which was which). 😭
But I also think of the way it opens up new avenues for storytellers. I think of how it can be use to showcase beauty. I think of how it gets rid of the gatekeepers. It's the Wild West, but for all the disruption to current advertising and filmmaking industries, there will be a world of new opportunities out there.
I'll leave you with a compilation of additional videos, along with the text prompts used to create them:
P.S. Now check out our latest video 👇
