We’ve officially moved from the AI image era into the AI video era, and it’s already far more advanced than people realize.
In a conversation with Hard Fork hosts Kevin Roose and Casey Newton, Google DeepMind CEO Demis Hassabis called Veo 3—Google’s cutting-edge video model—“the biggest announcement of the [I/O 2025] show.”
He added that it “seems to be viral now.” It’s true. I’ve seen people share Veo 3-generated AI videos on X and Reddit, but also on TikTok and LinkedIn, where this kind of news spreads more slowly.
In this short post, I’ll show you the most impressive examples I’ve gathered—they are very impressive. I’ll also share 10 reflections on the implications of this technology.
Prompt theory
Content warning: emotional distress. (Sources: 1, 2)
Fake news
You can tell by the teeth, for now. (Sources: 1, 2)
Stand-up comedy
Jokes not included in the prompt (I guess AI is funny now?). (Sources: 1, 2)
ASMR
ASMR is the most searched term on YouTube globally. (Sources: 1, 2)
TV and advertisement
How do I get some Puppramin?? (Sources: 1, 2)
Eating spaghetti
It’s not been two years!! (Sources: 1, 2)
Sports
Complex movement remains elusive (AI physics is quirky). Below, veo 2 vs veo 3. (Sources: 1, 2)
Filmmaking and shorts
Complex videos usually require a combination of AI tools. (Sources: 1, 2)
Flow, a filmmaking tool built on top of Veo 3, is better adapted for longer, more complex videos. The two examples below were made with Flow. (Sources: 1, 2)
The implications of Veo 3 are massive. Below are 10 profound ways this technology could reshape culture, creativity, and business.