According to this article, they are doing some of that already. Presumably it will improve:

> Waymos are getting more assertive. Why the driverless taxis are learning to drive like humans

https://www.sfchronicle.com/sf/article/waymo-robotaxis-drivi...

Oh very cool. My only complaint is that it didn't lay on the horn for the last example of the reckless driver nearly causing an accident.

I wonder if this is related to the Foundation Model: https://waymo.com/blog/2024/10/ai-and-ml-at-waymo