Laplace Intelligence red-teams robots and self-driving cars in simulation
Laplace Intelligence, an Ann Arbor startup founded a year ago out of the University of Michigan, is building automated red-teaming for robots and other physical AI, using a high-fidelity simulator post-trained from Nvidia's Cosmos world model and AI agents that design corner cases on their own. CEO and co-founder Howie Sun said the environments are rebuilt to scale from real map data with live traffic, and that other actors in the scene react to the vehicle under test, so problems can be found and fixed before deployment. Customers include Japanese truck maker Isuzu, Texas-based autonomous trucking firm Bot Auto and Nvidia, and the company is working with T-Mobile on a free smart-city traffic pilot for San Jose while starting early work with warehouse humanoid robotics partners.
As robots and self-driving systems move into the physical world, AI-generated adversarial testing could become a shared evidence base for safety, insurers and regulators.