Date: Monday, May 11
Start Time: 2:05 pm
End Time: 2:35 pm
The era of “software-defined” machines is giving way to AI-defined architectures, where behavior is shaped less by deterministic code and more by learned, agentic systems that improve through data. In this talk, we’ll explain how automotive autonomy has become a proving ground for this transition and why the same architectural patterns are now appearing in humanoid robots and industrial automation. We’ll unpack the technical drivers behind this convergence—foundation-model-based perception and planning, shared world representations, simulation and scenario-based validation and deployment pipelines that enable continuous updates at the edge. We’ll also share lessons learned on what must change when moving from siloed stacks to a unified AI stack: how to structure the data flywheel, manage safety and regression risk and measure ROI as models and policies evolve. Attendees will leave with a practical mental model of AI-defined systems and a road map for applying it across physical AI domains.

