This week, I moderated a panel on world models at the All In conference (no relation to the podcast), and it gave me a chance to dig into one of the most mysterious corners of the AI world. The big players in the space are Yann LeCun’s AMI Labs and Fei-Fei Li’s World Labs — and while both time as both have gathered a lot of buzz and funding, they also rank pretty low the trying-to-make-money scale.
At their core, global models are about automating spatial intelligence, so the field could head in lots of exciting and lucrative directions, from robotics to interactive video to more complex self-driving systems.
But after I commenced to press on where we would surely see the tech commercialized, things got foggy. The closest things I found to an authority was Michael Rabbat, a co-founder of AMI Labs and the company’s VP of World Models, who joined me on the panel. But when I pressed him on exactly what the corporation was working on, he was cagey. “We’ll talk about it when we’re ready to talk about it.” Over e-mail, he explained, “We’re still in a research and building phase, so we’re not talking publicly about any product plans or timeline.”
To be honest, AMI is less than a year old, so it’s fair sufficient to keep quiet. But this type of caginess extends to the complete world-modeling space. World Labs’ Marble is probably the most fully developed product in the space, and its demos range from straightforward media creation, constructing explorable environments for video games, or CGI effects. There are robotics use cases too, but the whole platform seems more designed to show capabilities.
That secrecy even extends to these companies’ suppliers. On the sidelines of the same conference, I spoke to Alex de Vigan, CEO of Physicl — a data supplier for the burgeoning world model business. He told he is aware of Physicl’s data has been useful for whatever they’re building, but he’s still in the dark about what precisely that is. “I wish they could inform us more. We could build more useful data if we knew what they have been working on,” de Vigan told me.
Part of the mystery comes from how versatile world models are as an idea. The simplest version is a navigable map of the world, same as the AI models that power self-riding cars. But the same modeling approach that supports a Waymo weave via traffic traffic could also support a humanoid robot carry boxes, or turn a few minutes of video footage into an explorable environment. AMI has already dipped its toe in production, biomedicine, robotics, and even AI software for doctors by its Nabia partnership. Surely it won’t pursue all of those — but maybe 1 or 2 of them are standing out?
No one doubts that there are lots of possible businesses to be built on model tech — and as long because it’s easy to fundraise, there’s no precise pressure to focus on one. In fact, there’s good reason not to. If AMI declared tomorrow that they built a humanoid OpenClaw or a next-generation Hollywood rendering system, a lot of other labs would unexpectedly be very interested in the space. Soon, the lab would face potential competition from the other world model companies, the neolabs, or even OpenAI and Anthropic.
In some approaches, it’s the flip side of all that easy fundraising. Your competition can fundraise, too — and the same money that helps you to build under the radar is also funding a lots of potential rivals as soon as the path to market will become clear. But even if that competition is inevitable, it’s best if you delay it for so long as viable, this means that keeping quiet about precisely what you’re building.
Cixin Liu fans will recognize this as a dark forest scenario: If you don’t realize who else is in the woods, it’s best not to attract attention.
