-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
What should Waymo do next?
- Thread starter diplomat33
- Start date
I agree that licensing Waymo Driver to an automaker would be a big step forward. I am not sure if Lucid is the right partner though. I love the Lucid Air as a car. But Lucid is a new start-up that is struggling to ramp up production. It might not be smart for Waymo to partner with a company that may not survive.
There is also the issue that Waymo would take on all the risk and liability if there were an accident. So Waymo would need to work very carefully on how they want to implement the system, what ODD, what sensor package. I think Waymo could do L4 highway. Waymo would need to validate that L4 highway is safe enough. And L4 highway could probably work with less sensors than the I-Pace. Waymo would need to work that all out.
I get that Waymo has not "solved L4" yet. But I think some type of limited application of their L4 on a personal car would be a great way to get the tech out and generate revenue. Not to mention it would really help show that Waymo's approach is viable for wider use and not just geofenced robotaxis.
I'll be curious how Waymo (and any AV for that matter) will handle Manhattan traffic/pedestrians. AVs are, by design, safe and courteous drivers. They should never drive aggressively or force their right-of-way. Will they just end up stun-locked.
View attachment 844209
Mobileye says they plan to address that by designing their driving policy to be assertive when needed. I think most AVs will need to figure out that sweet spot where the AV is assertive enough without being unsafe. AVs will need to have some level of assertiveness for these types of scenarios otherwise they will be paralyzed and won't get anywhere.
FYI, here is Mobileye's unedited NYC drive on their L2 vision-only stack:
Waymo has done 500k driverless miles in metro Phoenix now. The fact that it is driverless is significant because driverless has to have very low safety critical rate. So 500k miles of driverless in Phoenix is a significant milestone for Waymo IMO. I think it shows that Waymo Driver is getting really good at city driving.
I see this in more positive light for Waymo:
- Small company → Lucid can quickly make such a risky decision as starting to resell Waymo Driver
- Small company → Waymo could cap the risk by limiting their own liability contractually. If *sugar* hits the fan, Ludic dies but Waymo and Alphabet could be relatively protected
- Small company → Operations are naturally limited to very small scale that 1) allows Waymo to treat the collaboration as research experiment and 2) limit the risk (recalling 1,000 cars after catastrophe would have different outcomes from recalling 1,000,000)
- Attractive cars with lidar on the road today → Waymo could move quick by retraining with data (and simulation env) transformed to emulate Lucid sensor suite
- Lucid Air would become most advanced car overnight with a wide margin and 0 competition. Tesla would not have anything to compete with them for years. People with unlimited resources would put Model S and RAV4 in the same category, while Lucid could in category of 0 for quite a while
- Shipping L4 would give valuable validation for Lucid that could attract investments and provide path to a mid-priced volume car
- If everything goes well, Alphabet might become a deep pocketed potential acquirer (especially if Apple ships a car and Google feels left behind)
- If tech works, but financials do not, Alphabet might do a small investment just to keep them afloat
Just to demonstrate this, here is a familiar vision that could become true quickly with Waymo - Lucid collaboration:
Waymo Driver on Lucid Air
All you will need to do is get in and tell your car where to go. Take a nap, read emails or watch a video while you relax on your way there. Your Lucid Air will figure out the optimal route, navigate urban streets (even without lane markings), manage complex intersections with traffic lights, stop signs and roundabouts, and handle densely packed freeways with cars moving at high speed. When you arrive at your destination, simply step out at the entrance and your car will enter park seek mode, automatically search for a spot and park itself. A tap on your phone summons it back to you. Or have your kids summon the car from across the town ask it to pick them up from school.
I would buy that. Now when I think about, I did
Waymo Driver on Lucid Air
All you will need to do is get in and tell your car where to go. Take a nap, read emails or watch a video while you relax on your way there. Your Lucid Air will figure out the optimal route, navigate urban streets (even without lane markings), manage complex intersections with traffic lights, stop signs and roundabouts, and handle densely packed freeways with cars moving at high speed. When you arrive at your destination, simply step out at the entrance and your car will enter park seek mode, automatically search for a spot and park itself. A tap on your phone summons it back to you. Or have your kids summon the car from across the town ask it to pick them up from school.
I would buy that. Now when I think about, I did
Just to demonstrate this, here is a familiar vision that could become true quickly with Waymo - Lucid collaboration:
Waymo Driver on Lucid Air
All you will need to do is get in and tell your car where to go. Take a nap, read emails or watch a video while you relax on your way there. Your Lucid Air will figure out the optimal route, navigate urban streets (even without lane markings), manage complex intersections with traffic lights, stop signs and roundabouts, and handle densely packed freeways with cars moving at high speed. When you arrive at your destination, simply step out at the entrance and your car will enter park seek mode, automatically search for a spot and park itself. A tap on your phone summons it back to you. Or have your kids summon the car from across the town ask it to pick them up from school.
I would buy that. Now when I think about, I did
You own a Lucid?
yeah, I think Lucid with the Waymo Driver would be amazing. A beautiful car that was true L4!
Impressive - it drives more like a human than I thought it would. There were even moments where it came to a stop straddling a lane, with a truck pulling up next to it.
I think so too. Mobileye has really good vision-only. But I think it helps that they use maps and have really solid driving policy too. The maps and driving policy help make the vision-only smarter IMO. In fact, the Mobileye CTO just shared this clip to demonstrate how the REM maps help their vision-only drive in heavy rain when vision is reduced.
If they can take this vision-only with maps and driving policy that is already really good and add extra reliability with radar and lidar then they should be able to do L4. And Mobileye has a path to do L4 on personal cars which is very attractive to me. That is one reason why I am a big supporter of Mobileye.
I wish
Out of my price range."I would buy that. Now when I think about, I did" was a reference of buying Model S with almost the same "ad".
EVNow
Well-Known Member
On paper yes. Even REM looks interesting.
But no way to check it ourselves since no OEM in US trusts them enough to actually use them. Why only Zeekr ?
We don't know why Mobileye picked Zeekr but they must have seen each other as a good business match. I am sure when Mobileye is ready to release SuperVision in the US, they will. They have mapped most US roads and are actively testing SuperVision in the US. IMO, that's a good sign that they are planning to release SuperVision in the US.
EVNow
Well-Known Member
Release …. How ? Is there a secret OEM … because OEMs take years to do this kind of work. Multiple years.
It is possible that Mobileye could have a secret deal that they have not announced yet. Or maybe Zeekr still start selling their cars in the US. Also, Mobileye says that their goal is to make SuperVision and Chauffeur a "turn key" solution, meaning all the hardware and software would be done for the OEM. The OEM would just "plug and play" so to speak. Also, Mobileye could do geofenced robotaxis in the US first before making a deal with an OEM.
Obviously, I am speculating since Mobileye has not announced any concrete US plans yet. But considering all the mapping and testing, Mobileye has done in the US, it only stands to reason that Mobileye intends to release SuperVision and Chauffeur in the US at some point.
Last edited:
Daniel in SD
(supervised)
I think that no US OEM wants to do a supervised city streets system like FSD beta. This is of course contradicted by GM claiming that they're going to release UltraCruise and leverage Cruise's technology in future systems. I'll believe it when I see it.
I know Elon is arrogant and parted ways with Mobileye ages ago, but given what you've shown with Mobileye's vision-only today, I wonder if Telsa would be better off just buying Mobileye and bringing the tech inhouse. I think Mobileye's recent IPO valued at $50B, vs Tesla's nearly $1T. Perhaps my first sentence answered my own question.
I know Elon is arrogant and parted ways with Mobileye ages ago, but given what you've shown with Mobileye's vision-only today, I wonder if Telsa would be better off just buying Mobileye and bringing the tech inhouse. I think Mobileye's recent IPO valued at $50B, vs Tesla's nearly $1T. Perhaps my first sentence answered my own question.
Elon can't even go through with the Twitter purchase so....
Would Tesla be better off just using Mobileye's tech? IMO, yes. But I doubt Elon would buy Mobileye because it would be an admission that the Tesla approach failed. I also doubt Mobileye would do it since they want to be their own company that grows. But I know what you mean.
Similar threads
- Poll
