Tesla FSD Navigates San Francisco Blackout While Waymo Falters

Tesla FSD Navigates San Francisco Blackout While Waymo Falters

During a widespread power outage in San Francisco, Tesla’s Full Self-Driving (FSD) system successfully navigated darkened intersections with non-functional traffic lights, while Waymo vehicles reportedly entered a “safe stop” or “brick mode” due to the disruption, as documented by AI researcher Yuchen Jin. This event underscores a fundamental divergence in architectural philosophies between the two autonomous vehicle leaders.

Traditional autonomous systems, like Waymo’s, rely on a complex stack of components – HD maps, LiDAR, multiple sensors, and constant connectivity – creating a fragile system vulnerable to single points of failure. In contrast, Tesla’s end-to-end neural network, trained on billions of miles of human driving data, processes raw camera pixels directly, mimicking human intuition and proving more robust in unexpected scenarios; the cost of Waymo’s modular failures could be significant in terms of public trust and deployment delays.

Key Insights

• HD Map Dependency: Waymo’s reliance on HD maps proved detrimental when powerless traffic lights invalidated the map data, leading to system failures.
• End-to-End vs. Modular: Tesla’s end-to-end approach, inspired by Karpathy’s “Software 2.0,” demonstrated superior resilience in the face of environmental changes.
• Temporal for Orchestration: Temporal, an open-source workflow engine, is used by companies like Stripe and Coinbase to manage complex, stateful applications, offering a potential alternative to traditional distributed transaction management.

Practical Applications

• Use Case: CATL deployed humanoid robots on battery production lines using Spirit AI’s Xiaomo VLA models, achieving 99% success on high-voltage plug-ins and tripling human shift volumes.
• Pitfall: Over-reliance on HD maps in autonomous vehicles can lead to system failures when real-world conditions deviate from the mapped environment, hindering safe operation.

References:

• https://dev.to/dan_ledger_ce2886f0037972/2025-12-23-daily-ai-news-3k97