Google’s new text-to-video model, Veo 3, is barely out in the wild—and already, the internet is lighting up with experiments that feel more like film festival entries than tech demos.

Announced at Google I/O 2025, Veo 3 creates videos with high realism and fidelity--including 4k output and real world physics and audio--all from simple text prompts. It’s Google’s answer to Sora, Runway, and the swelling tide of generative video tools. But while the model is technically impressive, what’s truly turning heads is what people are doing with it.

Across social platforms, creators, animators, filmmakers, and even a few brands have begun showing off the strange, stylized, and sometimes downright stunning videos they’ve made with Veo. The results hint at a future where storytelling becomes less about cameras and more about ideas—and the prompts that spark them.

Here are some of the crazy examples showing what Veo can do:

A giraffe popping a wheelie in NYC

A comic telling an actually funny joke

A soothing baking video

A realistic gaming livestream

Bigfoot as a vlogger

A Broadway musical number about garlic bread

And of course...Will Smith eating spaghetti

It’s early days for Veo 3, and Google has so far restricted access to select creators through its VideoFX waitlist. But if these first glimpses are anything to go by, Veo 3 isn’t just about generating video—it’s about expanding the very definition of what video can be.

As with other generative media, questions around originality, rights, and ethical use are already percolating. But for now, the creative energy is undeniable. We’re watching the birth of a new visual language, one frame at a time.

Got a favorite Veo clip or made one yourself? Tag us @glitchwiremedia on X—your AI video masterpiece might be next on our list.