Agentic Engineering, Not Just Better Models

Peter Ludwig, co-founder and CTO of Applied Intuition, argues that the next major advantage in physical AI will come from making engineering systems — validation, integration, deployment, and operations — as agentic and intelligent as the models themselves. He contends that model quality alone cannot close the large gap between capability and certified, deployed machines; instead, industry progress depends on agentic platforms grounded in simulation, telemetry, validation tooling, and governance. Applied Intuition says it built an internal agentic platform (Dana) that accelerated development cycles roughly 20x, increased deployment frequency from weekly to multiple times per day, and enabled over a thousand internal apps and agents. Ludwig calls for combining frontier models with faster, agentic engineering systems to create a compounding learning flywheel for physical AI.