Waymo's self-driving is not as good as they'd like people to think. Neither is anyone else's. They're all behind schedule. This acqui-hire means nothing. I said the same thing to the TSLAQ cult when Tesla bought DeepScale. Talent is in high demand, so you grab it when you can.
Waymo vans with empty driver's seats cruise around Phoenix suburbs every day. Mostly with members of the public, but sometimes
with journalists. They can do the same around Mountain View, CA. Do you think they do this without a license?
If Tesla wants to acqu-hire someone, they should look at Voyage who is working on a
super-AEB. What makes this interesting, IMHO, is that it's independent of the regular self-driving system. Independent, uncorrelated systems can dramatically reduce failure rates. Voyage uses an expensive 32 line LIDAR, but the concept should work with the cheap, low res LIDARs beginning to appear. This assumes they've done meaningful work vs. mere demoware, of course, especially on the false positive problem. But if so I think this approach could solve one of Tesla's biggest problems with relatively little cost.