From the outside in this is how I see it (being in Austin which seems to be ground zero for a lot of the testing):

Waymo is betting that cutting zero corners and being rigorous will pay off down the line. They seem best in class.

Zoox seems to be betting that reducing time to get passengers in and out while providing a novel platform will be a differentiator. More sensors than Tesla but fewer cars than both the others.

Cybercab seems to be about cutting cost (and corners) wherever they thing they can get away from it and betting cheap will win the day (sort of like a discount airline). This assumes they're not doing something O(N) expensive behind the scenes, like having every ride monitored by a human operator in another country.

I remember some company dragging a person for awhile and the car not realizing. I'm worried tesla will find itself in a situation like that (or another edge case). I feel safer being around a Waymo as a pedestrian than a cybercab. I even had a Waymo stop for me at a pedestrian crossing when I was waiting for it to go by (I wasn't trying to cross).

In the long run Waymo will have to figure out how to have a cheaper platform and Tesla will have to toe the cheap/good-enough line. Zoox is still seemingly unlaunched so IDK about them. Time will tell.