معلومات أساسية
Ann Arbor-based startup Laplace Intelligence is addressing the risks of deploying AI agents into the physical world by creating automated red-teaming tools. Founded by CEO Howie Sun and chief scientist Henry Liu, the company aims to stress-test robotics and autonomous systems to prevent failures before they occur in real-world environments.
The platform utilizes high-fidelity simulations based on Nvidia’s Cosmos, which reconstruct real-world locations like highways or city streets using map data. These environments are interactive, allowing developers to adjust conditions in real time. Crucially, the system employs AI agents to automatically generate rare corner cases, reducing the need for human intervention in designing test scenarios.
Laplace currently generates revenue through project-based contracts with clients including Isuzu, Bot Auto, and Nvidia. While the startup faces competition from well-funded firms like Applied Intuition, Sun asserts that their methodology of using AI to test AI represents a new paradigm for dynamic, interactive safety validation.
Laplace is participating in the Startup Battlefield competition at TechCrunch Disrupt in San Francisco. The company plans to use the event to help raise funds and scale its operations as it transitions from academic research toward broader commercial application.
