Waymo self driving head to head

[Legend]

It's pretty hilarious if you compare it to how "the media" and "experts" think.


(https://www.businessinsider.com/tesla-sef-driving-not-in-top-ranked-autonomy-guidehouse-research-2023-3)

Waymo and others don't have a driver in the front seat but they still hit things and have issues that require human drivers to take over.

[Legend]

Cruise stuck in the Middle of the road in SF.  pic.twitter.com/0XOcfC3Q2M

— Tesla Owners Silicon Valley (@teslaownersSV) June 4, 2023

It's funny how this approach to self driving is pretty much a dead end. The cars rely heavily on HD maps so when the HD maps don't match reality, they often fail.

Will be interesting to see if they gradually shift their focus, or if they just ignore this stuff since it's rare.

[SWORDF1SH]

Why do they have Tesla at the back of the pack for FSD? I thought it was the opposite.

[Legend]

Why do they have Tesla at the back of the pack for FSD? I thought it was the opposite.

It's a dumb chart made by "experts."

Tesla is by far the leader for regular cars. Close for robotaxis with the potontial to leapfrog them.

[the-pi-guy]

Does Waymo have advantages?

Like is it better for mapped roads, but garbage otherwise?

[Legend]

Does Waymo have advantages?

Like is it better for mapped roads, but garbage otherwise?

Yes Waymo is fully driverless with no one in the front seat, but it's only available in three cities. Every route the car will ever drive has already been tested extensively. The car only has to adapt to traffic, pedestrians, and other dynamic obstacles. If the car ever gets confused, a remote human operator will advise it or a human driver will show up and finish the drive like a regular taxi. Old example of this happening

Waymo cars won't drive outside their zones. They also currently avoid highways and other roads that are difficult.


Essentially the approach used by Waymo and most other companies is to fully solve self driving in a hyper limited case and then expand the envelope. Tesla's approach is to solve universal self driving and then improve reliability.

At this point it's really hard to judge progress as an external observer. Teslas can drive the same roads as Waymo but they'll fail and require human intervention more often. Without internal data or people taking hundreds of identical trips, we don't know their relative reliabilities.