Alex Kendall· @alexgkendall · X·· 2 小时前
事实摘要
自 2017 年以来，我一直反复被告知端到端 AI 不可能安全，自动驾驶必须依赖传统的基于规则的系统。 但我们坚持了下来。今天我们发布 Learned Safety 策略，这是十年研究的成果。 https://wayve.ai/technology/learned-safety/
正文 · 原文
Since 2017, I've been repeatedly told that end-to-end AI can't be safe, and traditional rule-based systems are required for autonomous driving.
But we persevered. Today we are releasing our strategy for Learned Safety, the result of a decade of research.
https://wayve.ai/technology/learned-safety/
来源：Alex Kendall · x.com