写了一篇英文Tesla FSD vs Waymo

The Ultimate Stress Test: What San Franciscos Blackout Taught Us About Autonomy

The recent mass power outage in San Francisco wasnt just a municipal headacheit was a definitive moment for the future of autonomous driving.

While the city went dark, we saw two very different futures play out:

The Geo-fenced Approach: Dozens of Waymos reportedly stalled at intersections, bricking when traffic lights went dark and remote assistance became overloaded. When the infrastructure failed, the systems reliance on predictable environments became a bottleneck.

The First Principles Approach: Meanwhile, Teslas FSD (Supervised) continued to navigate the chaos. Why? Because it doesnt rely on a pre-solved map of the world. It uses vision and neural nets to interpret the world in real-timejust like a human does.

Why First Principles is Winning
Elon Musks insistence on Vision-Only was mocked for years as impossible. Critics said you need LiDAR. They said you need HD Maps.
But from a First Principles perspective:

1. The World is built for Vision: Our entire road system (signs, lights, lane markings) is designed for visual processing.

2. Generalization Specialization: A car that can only drive where there is a 5G signal and a perfect map isnt truly autonomous; its a high-tech train on invisible tracks.

Seeing FSD handle unmapped, blacked-out intersections while others waited for a server to tell them what to do proves that solving the hard problem of general vision is the only path to true scalability.

Its not always the easiest path, but its the right one.