Make your robotaxi predictions for the 8/8 reveal

So Elon says that Tesla will reveal a dedicated robotaxi vehicle on 8/8. What do you think we will see? Will it look like this concept art or something else?

I will say that while this concept drawing looks super cool, I am a bit skeptical if it is practical as a robotaxi. It looks to only have 2 seats which would be fine for 1-2 people who need a ride but would not work for more than 2 people. I feel like that would limit the robotaxis value for a lot of people. Also, it would likely need a steering wheel and pedals for regulatory reasons even if Tesla did achieve eyes-off capability.

So I think this is concept art for a hypothetical 2 seater, cheap Tesla, not a robotaxi.

Could the robotaxi look more like this concept art but smaller? It could look a bit more like say the Zoox vehicle or the Cruise Origin, more futuristic box like shape IMO and seat 5-6 people.

Or maybe the robotaxi will look more like the "model 2" concept:

Other questions:
- Will the robotaxis be available to own by individuals as a personal car or will it strictly be owned by Tesla and only used in a ride-hailing network?
- What will cost be?
- Will it have upgraded hardware? Radar? Lidar? additional compute?
- Will Elon reveal any details on how the ride-hailing network will work?

Thoughts? Let the fun speculation begin!

 

[petercc9]

It will be manned robotaxies for quite a while

I agree. Tesla isn't going to encourage people to get out of the driver's seat until they are sure there is not going to be any screwups.

and initially only in geofenced heavy urban centers.

Not sure why that would be needed

It will likely be years before any actual true robotaxi comes to market.

They will likely take the Waymo route of having safety drivers for a while.

Tesla has been taking M3 leasebacks and holding them back from resale for this purpose.

I would expect Tesla to build mainly a fleet of refurbished cars in the early days when drivers are required.
They may even announce a sweet trade deal to early adopter Tesla owners with cars too old to self-drive.

The 8/8 unveil will be more about the Tesla robotaxi network - how to use the Tesla app in a manner similar to Uber/Lyft, possibly drafting manned personally owned Tesla vehicles into the fleet with the explicit agreement to use FSD for the vast majority of fleet ride shares to collect data, etc. 8/8 is more about process and software than hardware IMHO. The hardware will come later (the actual robotaxi itself).

I'm sure they can ask but I doubt they expect many people to quit their jobs and become drivers.

We might see an actual concept of the robotaxi, just like we did with the CT in 2019, but it’ll take 18-24 months to build the first one on the new unboxed manufacturing line that doesn’t exist either, at the very least.

They certainly are not known for getting things to market speedily.

 

[spacecoin]

Waymo does not have the "brain" that Tesla does, which is arguably significantly more important. Just saying.

That's hillarious. Waymo is L4 and driverless. Tesla has a wide-ODD L2. They use the same tech, only Waymo does it better and was where Tesla is today 5-8 years ago.

 

[aataskin]

That's hillarious. Waymo is L4 and driverless. Tesla has a wide-ODD L2. They use the same tech, only Waymo does it better and was where Tesla is today 5-8 years ago.

Waymo uses radar and lidars, Tesla only vision. Maybe this is the answer.

 

[spacecoin]

Waymo uses radar and lidars, Tesla only vision. Maybe this is the answer.

It's a lot of small things in all aspects of the stack. Removing the human requires rigorous processes and advanced engineering (not just machine learning).

 

[HitchHiker71]

Not sure why that would be needed

It's a matter of ROI. The heavy urban centers are where rideshare penetration is the highest. That's where the initial focus will be - as is clearly evidenced by the job postings that are all in large urban areas. I'm not saying this will be 100% focused on large urban areas - but the lion's share of focus will follow the population density - robotaxies will start in the areas with the highest population densities and gradually move out to areas will less population densities over a period of many years.

They will likely take the Waymo route of having safety drivers for a while.

Exactly.

I would expect Tesla to build mainly a fleet of refurbished cars in the early days when drivers are required.
They may even announce a sweet trade deal to early adopter Tesla owners with cars too old to self-drive.

Potentially - hard to say. 2024.3.15 just released yesterday and now supports legacy MX/MS. This is a good thing because the take rates for FSD were highest back in the "good ole days" when FSD hype was over the top and prior to everyone realizing what Musk was saying about "next year" was far from accurate/true. Take rates are at all time lows right now (12.5% claims Tesla - used to be as high as 50% back in the early days). Getting data from the legacy models for FSD12 will help without a doubt - at least those that support it as you said.

I'm sure they can ask but I doubt they expect many people to quit their jobs and become drivers.

The job postings are for full time people - two different things. For the vast majority of Tesla owners that rideshare - they don't do it full time - they do it part time when they have availability. This is the co-hort that Tesla may target apart from the full time job postings - to boost up FSD adoption, take rates, and data collection. They will likely provide some kind of FSD subscription discount or price discount for folks who are willing to be manned robotaxi early adopters using personally owned vehicles. Again, we'll have to wait and see - just my two cents here.

They certainly are not known for getting things to market speedily.

Exactly. If we actually look at Musk's track record of bringing products to market on time based upon his claims vs reality - then we're talking years out. This is a fact that no one can dispute. This isn't to say that eventually it doesn't happen - eventually Tesla delivers - but it's typically years out with big misses on initial delivery timing expectations for new products. FSD robotaxi is an entirely new product.

 

[DanCar]

That's hillarious. Waymo is L4 and driverless. Tesla has a wide-ODD L2. They use the same tech, only Waymo does it better and was where Tesla is today 5-8 years ago.

Waymo has tons of compute on board, much more than Tesla. Source: I talked to a waymo engineer some years ago, and they said it was an ongoing goal to get as much compute possible on board. At that time Google's compute was 10x+ of Tesla. Suspect that is still true today. Waymo has their edge inference device on board, edge tpu. Would be interesting to know how many. In addition they have had the most powerful xeon computers on board. I wonder if Google's vast data-centers can help with real time navigation and control.
Summary: I don't consider them in the same league tech wise.

 

[jabloomf1230]

That's hillarious. Waymo is L4 and driverless. Tesla has a wide-ODD L2. They use the same tech, only Waymo does it better and was where Tesla is today 5-8 years ago.

Stop stating this as a universal fact. Waymo is Level 4 in extremely limited geographic areas and does not use highways. When Waymo covers an entire political jurisdiction like the US, then it can be considered as SAE Level 4.

 

[DanCar]

... When Waymo covers an entire political jurisdiction like the US, then it can be considered as SAE Level 4.

What makes you think that level 4 is equated with an entire political jurisdiction? That is not in the standard.

 

[diplomat33]

Waymo is Level 4 in extremely limited geographic areas and does not use highways.

Waymo does use highways. They started employee driverless rides on highways in Phoenix.

From surface streets to freeways, safely expanding our rider-only testing

Waymo will begin testing its fully autonomous passenger cars without a human driver on freeways in Phoenix to soon help Waymo One riders get where they’re going safely and efficiently. Just as we've taken a phased approach to rolling out our technology and operations in the past, we will...

waymo.com

When Waymo covers an entire political jurisdiction like the US, then it can be considered as SAE Level 4.

Covering the entire US is not L4. L4 means full autonomy in a limited ODD. So Waymo is L4 now since they are geofenced. If Waymo did cover the entire US, it would basically make them L5.

 

[aataskin]

It's a lot of small things in all aspects of the stack. Removing the human requires rigorous processes and advanced engineering (not just machine learning).

Agree, and I think some of those other bits are important for the reason that doesn't get mentioned enough. Humans are vastly superior at processing video information than the best computers, for now. Both better and requiring much less energy. Which is why if you can offload some of the video processing tasks to crude but reliable technology like radars, USS, etc., then you are making the remaining tasks more achievable.

 

[Dracaris]

On 8/8, if he says anything, he'll probably just announce that he meant 8/8/2088.

 

[ItsNotAboutTheMoney]

(1) Model 2 cum Robotaxi design with some hint at unboxing manufacturing approach.
(2) A Robovan design for Loop.

 

[Joey D]

That's hillarious. Waymo is L4 and driverless. Tesla has a wide-ODD L2. They use the same tech, only Waymo does it better and was where Tesla is today 5-8 years ago.

Saying that Waymo and Tesla uses "the same tech" just underscores how much you don't know about the tech in each.

 

[spacecoin]

Saying that Waymo and Tesla uses "the same tech" just underscores how much you don't know about the tech in each.

Enlighten me!

 

[ThomasD]

How far can Waymo operate from the Mothership?

 

[spacecoin]

How far can Waymo operate from the Mothership?

How do you validate the safety of a new release if you need to take on civil and legal liability for driving EVERYWHERE? A regression in driverless autonomy may kill someone.

 

[diplomat33]

How far can Waymo operate from the Mothership?

What "mothership"? What are you talking about? Are you asking how big the geofences are? We know the sizes of the geofences in LA, SF and Phoenix. Those sizes are not a mystery.

 

[spacecoin]

What "mothership"? What are you talking about?

I just assumed this was a "lol geofence".

 

[diplomat33]

That's hillarious. Waymo is L4 and driverless. Tesla has a wide-ODD L2. They use the same tech, only Waymo does it better and was where Tesla is today 5-8 years ago.

Waymo and Tesla do not use the same tech:

• Tesla uses off-the-shelf cameras. Waymo builds their own in-house customized HD cameras, lidar and radar.
• Waymo builds their own HD maps with their own vehicles. Tesla uses regular maps.
• Tesla trains a deep neural network, end-to-end, from camera input to control. Waymo trains individual neural networks to handle perception, prediction and planning separately.

The tech is quite different.

 

[spacecoin]

Waymo and Tesla do not use the same tech:

• Tesla uses off-the-shelf cameras. Waymo builds their own in-house customized HD cameras, lidar and radar.
• Waymo builds their own HD maps with their own vehicles. Tesla uses regular maps.
• Tesla trains a deep neural network, end-to-end, from camera input to control. Waymo trains individual neural networks to handle perception, prediction and planning separately.

The tech is quite different.

That's obviously not what I meant. Waymo is more advanced than Tesla in all aspects that I can think of. It's not like Waymo uses a rules based approach and no NN:s.

Let me put it like this: Are you aware of any technology that Tesla uses that Waymo doesn't? Are you aware of any novel approaches that Tesla invented or uses that Waymo doesn't?

It's my understanding that Waymo uses several end to end networks by the way, for smaller stuff like parking.