Sterling Anderson doesn't think humanoid robots should be building cars (www.theverge.com)

General Motors recently unveiled an ambitious tech roadmap focused on integrating AI and advanced computing in vehicle development, aiming to redefine automotive design with its “intelligence-defined vehicle” paradigm. The company intends to streamline the development process, reducing timelines from four to five years to just over two by leveraging AI-driven virtual testing. Sterling Anderson, GM's chief product officer, emphasized the shift from traditional methods to probabilistic AI that utilizes extensive in-house data for generative design, contributing to improved safety and faster iterations in vehicle prototyping. This initiative comes at a critical juncture for GM, as it faces fierce competition from efficient Chinese EV makers and shifting demand for electric vehicles in the US. The reliance on untested AI architectures and software could pose significant risks; any failure might set GM back against more agile rivals like Tesla. However, by embedding safety-critical functions on-board and utilizing cloud resources for non-essential AI interactions, GM aims to ensure high performance while managing technological complexities. The strategic pivot not only represents GM's response to evolving market dynamics but also highlights the growing importance of AI and data-driven methodologies in automotive innovation.