Computer scientist Louis Castricato, who was studying large language models, felt like he was hitting a dead end and decided to start a new company called Overworld. The company’s goal is to develop AI that can understand and navigate a world, not just words.

What are World Models?

World models are AI systems that teach robots how to react in a physical environment. They are considered the next frontier in AI development. Prominent scientists, such as Fei-Fei Li, are dedicating themselves to this field. Li describes the concept of a world model as ‘one of the most important and most overloaded terms in AI today.’

World models learn the statistical structure of space and time, including how light falls on a surface and how objects respond to force and follow the laws of physics. They are different from language models, which learn the statistical structure of text.

Applications of World Models

World models have various applications, including building video game worlds where a scene can adapt as a virtual character moves through it and interacts with objects. They can also be used to build AI with a broad awareness of its environment, similar to a robot’s brain.

Investors are showing interest in world model makers, with venture capitalists like Steve Jang investing in companies like Overworld and Causal Labs. The future of AI is expected to involve multiple types of models with different philosophies and architectures.

Original reporting: Texarkana Gazette — read the source article.