Strange Lab

We're building world models for business, a new class of foundation models to simulate every decision across the economy.

World models learn the dynamics of a world to predict possible actions, counterfactuals and outcomes. And once you have a world model, you can ask all manner of questions you couldn't before. What actions should be taken? What outcomes could we expect from those actions? What should we pay attention to?

We have seen these questions answered across worlds, from cybersecurity to open source software to pharma to fiction. And we're starting to see generalisation and benefits of scale.

We think interactive world models are essential to an AGI future because they make open-world exploration and continuous learning possible. Every company and enterprise will have its own, to help humans and LLMs operate with it and within it.

Our ultimate goal is to simulate every decision, across the economy. To make the causal understanding of our world explicit. And with it, to get to true general intelligence!