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!

 

[RDoc]

If you are expecting radar/lidar redundancy on the Tesla robotaxi, I think you will be disappointed. I think it is a pretty safe bet that the robotaxi will be vision-only. My guess is that the robotaxi will use "HW5". It will have higher resolution cameras than our cars and the more powerful FSD computer, that's it. They might add some extra cameras but that is about it IMO. I do think 8/8 will focus on the vehicle itself (and the ride-hailing app), not on any new autonomous driving tech. Remember, Elon thinks autonomous driving is basically solved already. So from his point of view, Tesla does not need any new tech. He will show off the cool looking robotaxi vehicle and claim that pending more E2E training, everything will be solved "soon".

My guess would be a major breakthrough in vapor emitters.

 

[aronth5]

If you are expecting radar/lidar redundancy on the Tesla robotaxi, I think you will be disappointed. I think it is a pretty safe bet that the robotaxi will be vision-only. My guess is that the robotaxi will use "HW5". It will have higher resolution cameras than our cars and the more powerful FSD computer, that's it. They might add some extra cameras but that is about it IMO. I do think 8/8 will focus on the vehicle itself (and the ride-hailing app), not on any new autonomous driving tech. Remember, Elon thinks autonomous driving is basically solved already. So from his point of view, Tesla does not need any new tech. He will show off the cool looking robotaxi vehicle and claim that pending more E2E training, everything will be solved "soon".

At a minimum I just want Tesla to solve the B-pillar problem. Until my ride this evening I had gone 9 weeks in between critical safety disengagements. The problem was pulling out from a low visibility intersection blocked by foliage. I could see by leaning forward a car was coming. The B-pillar couldn't. FSD started to pull out from the intersection with a car 95 feet away going about 40mph. I take this intersection out of my neighborhood at least once a day so am very aware of the B-pillar problem. Soon as the foliage drops FSD does a safe turn 100% of the time.

 

[aataskin]

Found the entire slide deck. Would be a lot of work to fake this.

Tesla Autonomy Investor Day Slides — The Tesla Show

Slides from Tesla’s Autonomy Investor Day.

theteslashow.com

Take a look at the Sensor Redundancy slide.

 

[aataskin]

If you are expecting radar/lidar redundancy on the Tesla robotaxi, I think you will be disappointed. I think it is a pretty safe bet that the robotaxi will be vision-only. My guess is that the robotaxi will use "HW5". It will have higher resolution cameras than our cars and the more powerful FSD computer, that's it. They might add some extra cameras but that is about it IMO. I do think 8/8 will focus on the vehicle itself (and the ride-hailing app), not on any new autonomous driving tech. Remember, Elon thinks autonomous driving is basically solved already. So from his point of view, Tesla does not need any new tech. He will show off the cool looking robotaxi vehicle and claim that pending more E2E training, everything will be solved "soon".

I think part of the definition of a robotaxi that everybody agrees on is it's where the manufacturer is liable. To me, the most important part of the presentation is what tech Tesla believe is enough to get them to that point, and whether they currently have anything even close to a working prototype. A ride hailing app is something anyone can do, same as a two-seater car or a bus, all of this is just so not exciting, really.

 

[diplomat33]

To me, the most important part of the presentation is what tech Tesla believe is enough to get them to that point, and whether they currently have anything even close to a working prototype.

I predict Elon will do more of the same. He will say vision-only is sufficient. He will say end-to-end is the way. He will say that Tesla FSD has solved most of autonomy now and the only thing left are some edge cases which he will admit "is the hardest part" but he will express confidence that Tesla can do it with more E2E training. He will repeat his promises about how V13 will be able to drive over a year without a safety critical intervention. He will make vague claims that HW4 will be 2x better, HW5 will be 5x better.

I hope they do show a working prototype. They could show a demo of the robotaxi prototype doing a short autonomous trip with no human in the car. I am sure Tesla could do some E2E training for HW5, enough to do a 5 mile driverless ride on an easy road, by 8/8. A 5 mile driverless demo with a robotaxi prototype would not prove that they have solved autonomy but it would lend credence to their robotaxi efforts.

 

[kpanda17]

Why not on 8/8:

Level set on the design and operation, vehicle (this one and our vehicles), app UX (Tesla owned vehicles controlled and/or ours), infrastructure charging/maintenance, park/wait, operation how it will operate and disengage remote takeovers (like waymo does), etc
Talk about vehicle code and app statuses and roadmap
where they are with one or more target city trials for later this year, Austin and Fremont
Staged/live or recorded demo
With Robo/cyber cab next steps
Our vehicles as robos, next steps
Cybervan review, status, next steps
Other Robo options
When we meet again to review all of the above progress? Q3, Q4, 2025
Next Robotaxi Day

What am I missing guys? No snarks please

 

[DanCar]

When robotaxi is viable in the 4+ year time frame, I think it might shift people to smaller vehicles. I like driving a larger vehicle for the few times that I need it, but if I can rent larger vehicle easily, comes to my door, then yes, I would shift to a smaller car. Although if the robotaxi is super cheap and convenient, I might just get rid of my vehicle.
48 days until 8/8.

 

[kpanda17]

When robotaxi is viable in the 4+ year time frame, I think it might shift people to smaller vehicles. I like driving a larger vehicle for the few times that I need it, but if I can rent larger vehicle easily, comes to my door, then yes, I would shift to a smaller car. Although if the robotaxi is super cheap and convenient, I might just get rid of my vehicle.
48 days until 8/8.

and thats the point, robotaxi actually means less vehicles, higher utilization, as stated in the master plan part 3
a large group of the population will just hail a robo and not buy/lease vehicles

 

[ThomasD]

Musk will come out with something. Does not matter what it is people will Ooh and ahh and the Stock price will go up.

 

[Just a Reader]

I never said L3 was good enough. Just pointing out that for those FSD owners who have no interest in Robotaxi, which is the vast majority, L3 is definitely an option for Tesla if they chose to implement it. You need to think out of the box on L3. Tesla could first implement L3 on controlled access highways which in the US is a significant percentage of the roadway. Owners could then read, text and watch videos. That would be helpful and drive the take rate. Tesla could also expand that to include city/streets. Perhaps limited but the beauty of L3 is there is no definition as to what is included and what isn't. Just look at Mercedes dumb implementation.

Mercedes' "dumb implementation" follows regulatory limits and they are currently increasing the available top speed. While its usage may be limited, it's quite useful for those who often have to commute in slow-moving highway traffic. I would have been quite grateful for such a system back when I had to commute into Munich by car.

 

[Just a Reader]

Why would any city be keen on having these robotaxis? They are destroying local jobs and suck money out of the local economy for the benefit of a US corporation. It's not usable for someone who is using a wheelchair and its low seating position makes it difficult to use for older people (provided it's designed as that two seater that is shown here).
A robotaxi would be far more useful and socially accepted in a rural environment to cster to elderly people and where taxi services are difficult to maintain economically. Even then its design wouldn't be particularly useful, though.

 

[aronth5]

Mercedes' "dumb implementation" follows regulatory limits and they are currently increasing the available top speed. While its usage may be limited, it's quite useful for those who often have to commute in slow-moving highway traffic. I would have been quite grateful for such a system back when I had to commute into Munich by car.

If you would have found Mercedes helpful that would have been great. But at $2,400 a year for a L3 solution that requires a lead car, restricts speeds to 40/mph, doesn't work in the rain, at night on sharp highway turns, only works in defined areas, and cannot change lanes, doesn't sound like a practical solution for most people. Tesla, if they wanted too could include for half the price a L3 for the highways that has none of those restrictions. Hopefully Mercedes can improve upon Drive Pilot in the future.

 

[AlanSubie4Life]

Tesla, if they wanted too could include for half the price a L3 for the highways that has none of those restrictions.

You keep saying this! Evidence needed. Remember FSD has only had an extremely small number of miles traveled - only 1 billion or so. And there have been some accidents in that interval, per Tesla's data.

Having unrestricted highway L3 in rain at night on any highway (or even call it only freeways) as you propose seems like it would be awesome but seems extremely difficult and I see no evidence that FSD is anywhere near that. I can't see Tesla wanting to pay for lots of busted battery packs.

I've never had any safety intervention on the freeway that I can recall on recent versions. My conclusion from that is that it is very unlikely that there would be no safety interventions in a reasonable sample size, which would mean L3 is inadvisable at this time.

 

[Just a Reader]

If you would have found Mercedes helpful that would have been great. But at $2,400 a year for a L3 solution that requires a lead car, restricts speeds to 40/mph, doesn't work in the rain, at night on sharp highway turns, only works in defined areas, and cannot change lanes, doesn't sound like a practical solution for most people. Tesla, if they wanted too could include for half the price a L3 for the highways that has none of those restrictions. Hopefully Mercedes can improve upon Drive Pilot in the future.

Well, if they could, then perhaps they should, rather than blaming their customers whenever they use it that way.
Apparently most Tesla customers don't consider FSD as that useful either, given that few people buy/rent it.

 

[stopcrazypp]

Well, if they could, then perhaps they should, rather than blaming their customers whenever they use it that way.
Apparently most Tesla customers don't consider FSD as that useful either, given that few people buy/rent it.

Literal multiple orders of magnitude more people rent or buy FSD than they did the Mercedes system (even the total volume of the vehicles that can potentially be optioned with it is way less than the total FSD users).

This is why I said elsewhere before Tesla shouldn't bother with limited L3. Far more people have interest in the door-to-door L2 that FSD is than those interested in a heavily limited L3, so it's not worth the liability risk.

 

[aataskin]

Why not on 8/8:

Level set on the design and operation, vehicle (this one and our vehicles), app UX (Tesla owned vehicles controlled and/or ours), infrastructure charging/maintenance, park/wait, operation how it will operate and disengage remote takeovers (like waymo does), etc
Talk about vehicle code and app statuses and roadmap
where they are with one or more target city trials for later this year, Austin and Fremont
Staged/live or recorded demo
With Robo/cyber cab next steps
Our vehicles as robos, next steps
Cybervan review, status, next steps
Other Robo options
When we meet again to review all of the above progress? Q3, Q4, 2025
Next Robotaxi Day

What am I missing guys? No snarks please

The only question is if it's going to be essentially another roadmap of great things to come, then why did he pre-announce it several months in advance? All of the stuff that you mention could be released several months ago. When people say there's a major announcement in five months then there's probably going to be some, well, major announcement, no?

I understand the stock pumping argument of @ThomasD , but I am not sure how the market will react to him re-using the same presentation deck every five years. My sense is that most people in this forum know that sooner or later he will have to bite the bullet and admit that Vision Only is not enough. Yes, this may open him up to lawsuits (and maybe it won't), but the company will be able to move on with its autonomy efforts much quicker.

 

[aronth5]

You keep saying this! Evidence needed. Remember FSD has only had an extremely small number of miles traveled - only 1 billion or so. And there have been some accidents in that interval, per Tesla's data.

Having unrestricted highway L3 in rain at night on any highway (or even call it only freeways) as you propose seems like it would be awesome but seems extremely difficult and I see no evidence that FSD is anywhere near that. I can't see Tesla wanting to pay for lots of busted battery packs.

I've never had any safety intervention on the freeway that I can recall on recent versions. My conclusion from that is that it is very unlikely that there would be no safety interventions in a reasonable sample size, which would mean L3 is inadvisable at this time.

L3 can be implemented in different ways so Tesla could determine their level of risk. For example I haven't had a critical safety disengagement on my highway drives for over 6 months and that was in a torrential downpour where FSD made me take over which is exactly what L3 is supposed to do. What they would need to do is have a more graceful handoff. Do I think Tesla will do L3, probably not but I can still hope they will. I believe the increased take rate for FSD might off set the financial risk.

 

[DanCar]

L3 can be implemented in different ways so Tesla could determine their level of risk. For example I haven't had a critical safety disengagement on my highway drives for over 6 months and that was in a torrential downpour where FSD made me take over which is exactly what L3 is supposed to do. What they would need to do is have a more graceful handoff. Do I think Tesla will do L3, probably not but I can still hope they will. I believe the increased take rate for FSD might off set the financial risk.

How many potholes in your area? A major pothole developed in a freeway near me and there were lots of Tesla's disabled on the side of the freeway because of this.