Google Genie 4 + Gemini Is Insane: Let’s Build 5 AI Worlds Live

Peter Yang · · 13 min

Peter Yang tests Google's Genie 4 world model early, building five interactive worlds live, from a kaiju attack in San Francisco to aliens in New York, and weighs what it means for movies and gaming.

Watch the episode 13 min · YouTube

The gist

  1. Yang explains that a world model predicts the next frame from your input, like pressing W to move forward, much as LLMs predict words.
  2. He says Genie 4 adds three things: generated sound, live prompt changes like a shooting star, and Gemini 4 planning multi-step events.

What they recommend · 1 pick

Genie Main topic

DeepMind's interactive world model system that can generate and simulate interactive environments.

"I think world models are the next big thing in AI after LLMs, and I was lucky enough to test Google's Genie 4 early... As you can see, it's not perfect. I don't know why the bowls are stacked on each other, but this is the beginning of something that's real..."
Peter Yang ▶ Play at 0:06

What Peter uses it for: Generating interactive 3D worlds and scenes from prompts, with live prompt-based changes and sound

3 creators recommend Genie, including CJ Hess, Ali Ghodsi.

Peter, Riley, Claire and Liam agree on 39 things

See what they all use, in their own words. Paid ads are left out.

See what they all swear by Or pick your own people

Also mentioned

2:17 Gemini · "And third, Gemini 4 works behind the scenes to plan multi-step events like this robot arm working through a..."
12:15 ChatGPT · "So, just like how everyone is using AI to build personal apps, for example, in ChatGPT, we might soon be..."

Ads in this episode

Wispr Flow. They're paid, so we keep them out of the picks above.

See what Peter recommends without being paid →

◎

Related episodes

Episodes covering similar topics

What do Peter and experts like them all swear by?

39 things in common

See