Laplace Intelligence stress-tests robots before real-world deployment

Laplace Intelligence stress-tests robots before real-world deployment
Ann Arbor startup Laplace Intelligence builds automated red-teaming tools for physical AI, including robots and autonomous vehicles. Founded a year ago by Howie Sun and Henry Liu out of the University of Michigan, the company uses high-fidelity simulations built on Nvidia's Cosmos to recreate real-world environments and stress-test AI systems before deployment. AI agents within the platform automatically generate rare, tricky scenarios using sources like news and police collision reports. The goal is to help companies find and fix problems before launch, and build a trusted evidence base for customers, insurers, and regulators.
Read the original article →