World Labs Ups The Ante In World Generation

Forbes

11-06-2025

Red cup Mushroom, Wild Fly Agaric Amanita Muscaria toxic with poison and hallucinogenic properties. ... More Autumn season in the forest with orange, red and brown leaves from the trees and saturated fall nature colors in Stramproy area in the province of Limburg in the Netherlands, near the Belgian borders on October 24, 2020. (Photo by Nicolas Economou/NurPhoto via Getty Images)
A majestic red and black checkerboard pathway, framed by overgrown shrubbery. A manorial structure with bright red windows, on a craggy cliff at the end of a long and winding road. A close pathway through a proliferation of neon mushrooms.
What does all of this have in common? It's been generated by groundbreaking technology from World Labs, where the model can extrapolate an entire world based on one single image.
It's important that people understand the technology at work here. Just a couple of decades ago, our state-of-the-art technique was to combine thousands of images into a frame-by-frame virtual exploration of a digital space. People used this technology to sell homes with virtual real estate tours, and for other kinds of use cases.
This new thing is quite different. The technology is understanding how light and shadow affect surfaces. It's understanding how structures change when viewed from different angles. And most importantly, it's dreaming up a world of color and light, the likes of which we've never seen!
World Labs, by the way, is a new player in the industry. It was created by Fei-Fei Li of Stanford fame just this year.
FEATURED | Frase ByForbes™
Unscramble The Anagram To Reveal The Phrase
Pinpoint By Linkedin
Guess The Category
Queens By Linkedin
Crown Each Region
Crossclimb By Linkedin
Unlock A Trivia Ladder
One of the most prominent points of these public demos that I've been looking at online is that they're bright, bold and beautiful. These larger-than-life scenes are rendered capably in 3-D, and you can move around a little bit, although billing it as a digital 'world' is a little bit misleading. It's not hard to generate an out of bounds error, just by taking a few steps forward. On a side note, the page also caused my browser to crash more than once.
Having said all that, this is extremely illuminating, some examples of capability far beyond what we would have thought possible just a few years ago.
The closest thing we have to compare it to is the rendered game environments that were until quite recently mostly hand-coded. A Techcrunch article compares the World Labs project to Minecraft, which might appeal to a ten year old player more than a VC. But the ramifications of this are evident, or at least they should be. This isn't just a kid's game. Recent applications like a 'dress-anybody' garment simulator are just the beginning.
Further down in the same demo page, you have a presentation attributed to Brittani Natali, who apparently put together a short film using different elements of today's new cutting-edge tools. The full stack included the World Labs generator as well as ElevenLabs, presumably for speech, and other tools like Suno and Blender and CapCut.
On another side note, it seems that in addition to a dearth of information for the creator (Brittani Natali) in a google search, ChatGPT was unable to locate any info on this person either. It seems the individual has a pretty thin web footprint!
Anyway, as for the film, the result is striking - the viewer of the film experiences walking through a good number of these digital environments quickly, and then centers in on a dilapidated house that's empty and abandoned.
Our love affair with abandoned spaces has always been there in the human imagination. We love to explore – and see inside something that no one has seen for years. But what if we're seeing inside something that nobody has ever seen in the history of humanity?
These colorful liminal spaces aren't just abandoned – they're brand new. This is the frontier of our world where we set out through uncharted waters to see what lies ahead.
Is it scary? Is it exciting? You decide. But the bottom line is that it's here, and it's going to start popping up in unexpected places.
I've covered a lot of big headlines this month, from corporate strategies pivoting quickly, to hardware wars that are determining who will benefit from the next generation of models. But it's also very interesting to keep an eye on the leapfrogging that AI is doing in multimedia – from Stable Diffusion and Dall-E, to Sora, to this new thing - where pretty soon, metaverse environments are going to create themselves.