特斯拉称 FSD 使碰撞风险减少 7 倍、非公路事故减少 5 倍，按比例推算的“每年可避开 3.2 万人死亡、190 万人受伤”听着震撼。理性看两点：一，统计口径和样本选择决定了结论力度；二，哪些情景 AI 真能覆盖，哪怕还得靠法规与基础设施配合。总体评价：技术潜力巨大，但我们也需要问清楚“怎么测”和“测给谁看”。🔎 Numbers are striking: Tesla claims 7x reduction in crashes, 5x in off-road incidents, translating to ~32k lives and 1.9M injuries avoided annually. Two sober points: (1) stats hinge on methodology and sample; (2) AI’s real-world coverage depends on scenarios, regs, and infrastructure. Bottom line: big promise, but ask how it’s measured. 🔎
