Some of the companies doing this use dual, independent LTE networks. I don't really understand the business model, though perhaps a Robotaxi company could use it for occasional trips outside their geofence or something.
Hopefully as networks become more robust, ie with 5g rollout, will probably become more safe if needed in distant future. I don’t see it as a bad idea to have on call drivers ready to navigate super edge cases as needed for autonomous as they transition from lv4 - 5.
Best video yet of Waymo driverless: 1 hour and 21 minutes of unedited rides, driverless with front view and FSD visualization view:
very impressive. the first ride was $12.41. I put in the same addresses on uber x it is $15.71. 2nd ride was $10.24 with Waymo and Uber x is $10.55.
So 3 months, or maybe 6? I'm particularly impressed with the assertiveness of the driving. The Waymo drive system isn't fooling around. I haven't seen Summon in a year, but it was like a blind man with a cane compared to this Waymo video. While I normally avoid the LIDAR debate, I think soon the "camera is good enough" side may need to fold. Waymo can likely move with so much certainty because it has a higher degree of confidence of what's in it's immediate environment. I assume Tesla needs camera movement to identify more difficult objects.
Tesla's strengths are in different dimensions. Does anyone think Waymo: 1. Will scale globally quicker? 2. Willing to take risk? Also I'm dubious as to Waymo's claim of public and fully driverless. Suspect this is going to remain small scale and public will get 95% of rides with a safety driver. Hope I'm wrong and Waymo is widely successful.
Waymo has many advantages that are non Lidar related. Such as: Experience since 2009 1500 employees, much more than Tesla autopilot team Humongous datacenters at their disposal TPUs while Tesla is dreaming of dojo World class scientist team at Google Spending and losing billions each year Custom hardware designed at Google: edge TPUs Waymo has remote monitoring
I don't believe Waymo is driving solid state LIDAR yet. An indicator of the intention to scale may be seeing Waymo I Pace with more integrated sensor systems. I don't think that Jaguar made a deal with Waymo to have stupid looking cars driving around.
In a desert, in tiny subsection of the city, and for now only to people that were already approved for the closed beta. Yes, I know.
That is quite impressive. I haven't watched all of it, but the first thing that stuck out to me was at the very beginning where the car did a three-point turn in a driveway to turn around. That was a maneuver I wasn't expecting. It certainly wasn't perfect, but it navigated the city pretty well, and the visualization seemed to pick up on all the important things. As I said, I only watched the first ride, and the photographer mentioned that his camera work improved in later rides, but I was a little disappointed not to see what the visualization showed when they passed the work truck on the side of the road with the workers picking up cones, etc. Would have been interesting to see what the car picked up. I'm ready for my Tesla to do those things!
There's a 10 minute Jump Cut version that skips the boring parts for those who don't want to watch the full hour and 21 minutes. The Reddit thread has some discussion and one commenter who highlighted most of the interesting points. Ha, they only perform miracles in a 5 mile by 10 mile area. They're DOOMED! They have test vehicles running in 25 or so cities and could start services in most of them tomorrow if they were anxious to lose even more money. They need to get their business model and logistics fixed before scaling. As I've said for years, Tesla has the vastly superior business model, regardless of whether their technology works or not. Waymo addressed this as an engineering and science problem. Elon looks at it from a marketing angle. The better marketeer usually wins in consumer markets.
If Musk could go back in time he would be crazy to change anything. Even if Tesla has to eventually pay back buyers of FSD the write off and embarrassment will be well worth it. Likely Tesla plans to increasingly add useful "FSD" only features. It will be interesting to see if Tesla continues to neglect increasingly standard features such as good blind spot warning and rear cross traffic. I assume Tesla starting to use the in car camera means that they planning to be stuck at level 2/3. Tesla can't stand still with technology unless the cars truly have the hardware capable of FSD. The new iPhone 12 is state of the art and is an interesting comparison to what Tesla is delivering.