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!

 

[spacecoin]

I would not have taken that bet...and I roll my eyes as much as most (said so in this post) when it comes to WholeMarsCatalog pimping of FSD.

Curious though...what was your side of the wager?

He could change mine.

I may be interested, but would have to flesh out exact terms.
1) Any tesla vehicle (or just the next gen robotaxi vehicle which has not yet been announced?)
2) Must be Tesla owned? Or can it be any company utilizing Tesla vehicles / software?
3) Must actually be charging for rides? (Not just free?)

Any Tesla vehicle with an autonomy stack from Tesla operated autonomously by any owner in any US city on public roads.
Autonomous operation. No driver in the driver's seat, for any reason. It's fine if the rides are free at long they are open to the public and you can pick off and dropoff wherever you like in an ODD size similar to Waymo's in one city as of today, 2024-04-17.

 

[Joey D]

He could change mine.

Any Tesla vehicle with an autonomy stack from Tesla operated autonomously by any owner in any US city on public roads.
Autonomous operation. No driver in the driver's seat, for any reason.

Make it by end of 2027 and we have a bet.

 

[spacecoin]

Make it by end of 2027 and we have a bet.

Sure, if we say vision only and HW4 as well.

 

[Joey D]

Sure, if we say vision only and HW4 as well.

No. I don't know what Tesla may iterate on.

 

[spacecoin]

No. I don't know what Tesla may iterate on.

Then end of 2026. For any amount on any hardware.

 

[Joey D]

No...you claimed "decades away"

Then end of 2026. For any amount on any hardware.

Sorry, no. I would think for someone who claimed it was "decades away" (as in 20+ years), you would be jumping at 2027?
But if not, then we have no bet.

 

[spacecoin]

No...you claimed "decades away"

Sorry, no. I would think for someone who claimed it was "decades away" (as in 20+ years), you would be jumping at 2027?
But if not, then we have no bet.

OK! Level 5. Then 2027 is fine. All public US roads, 365x7x24. No geofence. Any hardware, any vendor in the world, not limited to Tesla.

We already have geo limited lidar based robotaxis TODAY, and since 2017 for that matter. What's Tesla's edge except being 8 years late?

 

[Joey D]

OK! Level 5. Then 2027 is fine. All public US roads, 365x7x24. No geofence. Any hardware, any vendor in the world, not limited to Tesla.

We already have geo limited lidar based robotaxis TODAY, and since 2017 for that matter. What's Tesla's edge except being 8 years late?

You keep changing the bet. These are the terms.

1) Level 4 (but driving without a driver present in vehicle)
2) ANY tesla vehichle, owned by ANY user
3) Operated on ANY public US Road. (Not required to operate on ALL public US Roads)
4) Operated at ANY time (Not necessarily 24x7)
5) This occurs by end of 2027

We have already discussed Tesla's edge: That is, should tesla be successful with their approach, they will have a much easier time scaling across country and ultimately globally, and will be able to incorporate the technology / software into a much larger fleet of vehicles,

If you don't understand what I'm saying, perhaps another bet?

I say that Waymo will not have this by 2030:
1) Sell any car to the public with it's technology.
2) Have any "unsupervised car" (Level 3+) that I can hail (or own) take me from my house in the suburbs, and request that it take me anywhere I want to go (say within 100 miles)
3) By 2030

Not sure if Tesla will be able to do that either (I would not bet on that)...but I would give Tesla a much better chance at achieving that than Waymo. That is "Tesla's Edge"

 

[spacecoin]

You keep changing the bet. These are the terms.

1) Level 4 (but driving without a driver present in vehicle)
2) ANY tesla vehichle, owned by ANY user
3) Operated on ANY public US Road. (Not required to operate on ALL public US Roads)
4) Operated at ANY time (Not necessarily 24x7)
5) This occurs by end of 2027

We have already discussed Tesla's edge: That is, should tesla be successful with their approach, they will have a much easier time scaling across country and ultimately globally, and will be able to incorporate the technology / software into a much larger fleet of vehicles,

If you don't understand what I'm saying, perhaps another bet?

I say that Waymo will not have this by 2030:
1) Sell any car to the public with it's technology.
2) Have any driverless car (Level 3+) that I can hail (or own) take me from my house in the suburbs, and request that it take me anywhere I want to go (say within 100 miles)
3) By 2030

Not sure if Tesla will be able to do that either (I would not bet on that)...but I would give Tesla a much better chance at achieving that than Waymo. That is "Tesla's Edge"

You need to understand when that I said "decades" I also said "general autonomy". That means Level 5. So you changed the discussion to L5 quoting me. I specifically said "ODD size like Waymo 2024". That's approx 225 sq miles.

Again we already have geo-limited robotaxis using expensive hardware, and Tesla bulls much like yourself has explained to me that it doesn't scale.

Now you're suggesting I should take on a bet that when Tesla realises that's the only viable way to autonomy, that they might go driverless in almost three years from now and deploy in a single city in a small scale? Ten years after Waymo? What a joke.

My conclusions based on this discussion is:
1) Neither of us think HW4 isn't enough for autonomy
2) Neither of us think Autonomy from Tesla before 2026 is happening.
3) You seem to think that it's a feat if Tesla deploy by 2027 in one city in a limited geofence without a safety driver, ten years after Waymo.

 

[spacecoin]

You keep changing the bet. These are the terms.

1) Level 4 (but driving without a driver present in vehicle)
2) ANY tesla vehichle, owned by ANY user
3) Operated on ANY public US Road. (Not required to operate on ALL public US Roads)
4) Operated at ANY time (Not necessarily 24x7)
5) This occurs by end of 2027

6) Minimum ODD size 225 sq miles. Pickup and dropoff anywhere in that geo.
7) Available to anyone in the public. I want to try it myself.
8) Operates in fog and rain (like Waymo today).

F!ck it. Sure, by eoy 2027.

$10k USD? We can escrow it.

 

[Joey D]

You need to understand when that I said "decades" I also said "general autonomy". That means Level 5. So you changed the discussion to L5 quoting me. I specifically said "ODD size like Waymo 2024". That's approx 225 sq miles.

Again we already have geo-limited robotaxis using expensive hardware, and Tesla bulls much like yourself has explained to me that it doesn't scale.

Now you're suggesting I should take on a bet that when Tesla realises that's the only viable way to autonomy, that they might go driverless in almost three years from now and deploy in a single city in a small scale? Ten years after Waymo? What a joke.

My conclusions based on this discussion is:
1) Neither of us think HW4 isn't enough for autonomy
2) Neither of us think Autonomy from Tesla before 2026 is happening.
3) You seem to think that it's a feat if Tesla deploy by 2027 in one city in a limited geofence without a safety driver, ten years after Waymo.

This is exactly why I insisted on clarifying the "Terms" of the bet. It helps flesh out exactly what our viewpoints are....even though you still don't seem to grasp mine.

Your conclusions are wrong, BTW.

1) I never said I don't think HW4 isn't enough. I'm saying for betting purposes, it is too restrictive to insist on HW4. Tesla may or may not iterate. Would I be shocked if HW4 is enough? No. You would be. That is the difference.
2) Similarly, I would not be surprised at all if Autonomy from Tesla happens by end of 2026. There is a difference between "what is your best guess" and "would you be surprised if...." CLEARLY, you would be shocked if Tesla has some form of autonomy by end of 2026. I would not. This does not mean my "best guess" (bet) would be 2026
3) It is a feat if Tesla deploy by 2027...and is able to then leverage their technology to relatively rapidly scale afterwards. I do not see Waymo in any position to do this now, or 4 years from now, even though it will have been "10 years" of autonomous driving for them (I assume you are correct).

 

[Joey D]

6) Minimum ODD size 225 sq miles. Pickup and dropoff anywhere in that geo.
7) Available to anyone in the public. I want to try it myself.
8) Operates in fog and rain (like Waymo today).

F!ck it. Sure, by eoy 2027.

$10k USD? We can escrow it.

Nope.. But keep trying to change the terms. And no, I won't be wagering $10K on anything. (I have a spouse to answer to). Friendly bet (if we can agree on terms), sure.

 

[spacecoin]

Nope.. But keep trying to change the terms. And no, I won't be wagering $10K on anything. (I have a spouse to answer to). Friendly bet (if we can agree on terms), sure.

6) Minimum ODD size 225 sq miles. Pickup and dropoff anywhere in that geo.
7) Available to anyone in the public. I want to try it myself.
8) Operates in fog and rain (like Waymo today).

1kUSD?

Which of the above points above do you have a problem with?

 

[Joey D]

6) Minimum ODD size 225 sq miles. Pickup and dropoff anywhere in that geo.
7) Available to anyone in the public. I want to try it myself.
8) Operates in fog and rain (like Waymo today).

1kUSD?

Which of the above points above do you have a problem with?

These are my terms:
1) Level 4 (but driving without a driver present in vehicle)
This means that it may or may not operate "in rain and fog". It means the car must pull over (without driver) in conditions where it does not have confidence. "Driving in Rain and fog" is too ambiguous. I'm sure there will be levels of each and at certain speeds where at some point, it won't be able to safely drive at speed. BTW, Waymo does not drive in "Rain and fog" at high speeds on the highway.
2) ANY tesla vehicle, operated and owned by ANY user
3) Operated on ANY public US Road. (Not required to operate on ALL public US Roads)
4) Operated at ANY time (Not necessarily available 24x7)
5) This occurs by end of 2027
6) 225 sq. mile ODD minimum.
7) I do not agree with "pick up and drop off anywhere" in that geo. Too easy to abuse. Does this include inside gated communities? If the pickup is 50 ft from the requested area, does that disqualify? 100 ft? If you request to be picked on on an interstate and it denies...is that not OK? I would be OK with "pick and drop-off anywhere reasonable in the geo. Again...this is going to be an area of subjectivity.
8) Available to anyone in the public...as long as you are in the servicing area, sure.

$200. I am more interested in this being an exercise in vetting each other's actual viewpoints...than the bet itself.


EDIT: Having thought about it. the 225 sq. mile ODD may be more reasonable. So I agree with that and amended #6.

 

[diplomat33]

* An L2 car does not need to ask the user to intervene in low confidence situations. The user must proactively intervene: The user is responsible for accidents at all times.
* An L3 car will ask the user to take over in low confidence situations. The "car" is responsible for accidents while driving.
* An L4 car does not need a driver and must pull - over / disable driving in a low confidence situation. The "car" is always responsible.

SAE levels have nothing to do with confidence levels. They are about what the system is designed to do:

• L1 can do either steering or acceleration/braking but not both at the same time.
• L2 can do both steering and acceleration/braking at the same time but is not designed to do all driving tasks.
• L3 is designed to perform all driving tasks on its own in an artificially limited ODD but cannot perform the fallback.
• L4 is designed to perform all driving tasks and fallback on its own but only in an artificially limited ODD.
• L5 is designed to perform all driving tasks and the fallback on its own in a wide ODD of a typical human driver (ie everywhere, day, night, all drivable road and weather conditions).

Confidence determines when the system is ready for deployment based on what ODD or what feature the company is trying to deploy. But the level is not determined by the confidence. The level is based on the engineering design of the feature/system.

A typical L2 system would be lane keeping + cruise control at the same time. but that's it. So it can stay in the lane and maintain a safe distance from a moving lead car at the same time but nothing else. So it cannot move out of the lane to avoid an obstacle, it cannot respond to traffic lights or stop signs etc... Hence, a L2 system always requires a human driver since it needs a human to perform the driving tasks that it is not designed to do. So confidence has nothing to do with the level itself. You could have L2 that is very reliable or you could have L2 that is not very reliable. And it would also depend on the ODD. On interstate highways, with clear lane markings, no construction zones, on a clear day, with light traffic, L2 might go a long time with no interventions because lane keeping and cruise control are the only driving tasks that are needed for awhile. But put L2 in the city where it is not designed to stop for red lights or make protected turns, and it will likely require frequent disengagements. We saw that in the early days of FSD before FSD beta was released.

L3 requires a human to perform the fallback but otherwise can perform the driving on its own. Again, confidence has nothing to do with being L3. You could have a L3 system that almost never asks the human to take over or a L3 system that is designed to ask the human to take over quite frequently because of a very limited ODD.

You could have a L4 system that is not very reliable and as a result, needs a safety driver to intervene a lot. You could have a L4 system that is more reliable and the company decides to deploy it as a driverless robotaxi in a geofence. We see this with some of the L4 systems being tested. Some companies are testing L4 with safety drivers because their disengagement rate is still quite poor. Others like Waymo are deploying L4 without safety drivers in limited geofences because they have confidence the system is safe enough in that limited ODD.

Likewise you could have a L5 system that is not very reliable and requires human supervision or a L5 system that is safe and reliable enough that the company trusts it unsupervised.

On this last point, it should be noted why L5 is so much harder to deploy commercially than L4. With L4, you only need to achieve high safety in a limited ODD, like a small geofence, and you can deploy say a commercial robotaxi service. But to deploy L5 commercially, you need to achieve high safety basically everywhere, all the time.

cc: @jebinc

 

[spacecoin]

I am more interested in this being an exercise in vetting each other's actual viewpoints...than the bet itself.

So am I. The money is irrelevant and the escrow costs isn't worth it for a few hundred bucks.

Thanks for the discussion though.

 

[Joey D]

SAE levels have nothing to do with confidence levels. They are about what the system is designed to do:

• L1 can do either steering or acceleration/braking but not both at the same time.
• L2 can do both steering and acceleration/braking at the same time but is not designed to do all driving tasks.
• L3 is designed to perform all driving tasks on its own in an artificially limited ODD but cannot perform the fallback.
• L4 is designed to perform all driving tasks and fallback on its own but only in an artificially limited ODD.
• L5 is designed to perform all driving tasks and the fallback on its own in a wide ODD of a typical human driver (ie everywhere, day, night, all drivable road and weather conditions).

Thank you for the detailed explanation! There are so many contradictions when searching it's laughable.
To be clear, you're saying that the difference between L4 and L5 is the breadth of the ODD. L5 is basically "anywhere, any time, any conditions that a human would reasonably drive", vs. L4 which introduces some other "artificial" constraint. Correct?

Likewise you could have a L5 system that is not very reliable and requires human supervision or a L5 system that is safe and reliable enough that the company trusts it unsupervised.

This I do not understand. I assumed that if active Human supervision is required, that it could not be L3, L4 or L5.

EDIT: Let me put it another way: If a company does not "trust" its L3, L4, or L5 to run without active supervision, then IMO it means to me that system is still in a "under development" or in "pre-release" type state. I don't see how a company can actually market, sell or deploy (to the public) an "official" L3, L4, or L5 system while requiring the user to actively supervise while the vehicle is operating in those modes.

 

[HitchHiker71]

Space travel is expensive too?

I'm assuming your comment was in jest - but just in case it's not - SpaceX is a completely separate entity/company - with it's own distinct balance sheet and finances/funding. Tesla monies have no bearing on SpaceX in any way.

 

[diplomat33]

To be clear, you're saying that the difference between L4 and L5 is the breadth of the ODD. L5 is basically "anywhere, any time, any conditions that a human would reasonably drive", vs. L4 which introduces some other "artificial" constraint. Correct?

That is 100% correct.

This I do not understand. I assumed that if active Human supervision is required, that it could not be L3, L4 or L5.

That is a common misconception. Human supervision just means that your L3, L4 or L5 is not ready for unsupervised deployment. If you think about it, all L3, L4 and L5 require active human supervision at first, when they are being developed and tested. Of course, L2 and L3 will require a human driver. L4 and L5 are designed to eventually not require a human driver. But L3, L4 and L5 will need active human supervision during the testing phase.

This might help:

If you are designing a driving system that will only perform certain steering and acceleration/braking actions at the same time (like lane keeping + cruise control) but is not designed to replace the human driver then you are designing a L2 system.

If you are designing a driving system to do all the driving on its own but not replace the human driver, then you are designing a L3 system.

If you are designing a system to eventually replace the human driver, then you are designing a L4 or L5 system.

If you intend the system to replace the human driver but only be deployed in an artificially limited ODD, then you are designing L4. During testing, when it is not reliable yet, you will need a safety driver. When it reaches a certain reliability where you are confident it would be safe enough unsupervised, then you can consider deploying it without a safety driver but constrained to a limited ODD, like a geofence.

If you intend the system to replace the human driver and be deployed "everywhere" then you are designing L5. During testing, when it is not reliable yet, you will need a safety driver. When it reaches a certain reliability where you are confident it would be safe enough unsupervised, then you can consider deploying it without a safety driver.

 

[Joey D]

That is a common misconception. Human supervision just means that your L3, L4 or L5 is not ready for unsupervised deployment. If you think about it, all L3, L4 and L5 require active human supervision at first, when they are being developed and tested. Of course, L2 and L3 will require a human driver. L4 and L5 are designed to eventually not require a human driver. But L3, L4 and L5 will need active human supervision during the testing phase.

Makes sense...key being "testing phase." In short, what I take from this is that a company cannot market, sell, etc. a system as L3, L4 or L5...if human supervision is required when operating in those regimes. Of course, your system can be developed with the goal to achieve L3/4/5... and part of that testing / development process will almost undoubtedly utilize human supervision.

I assume "liability" to the manufacturer also comes with an officially marketed / sold L3/L4/L5 system.