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]

This RoboTaxi will be a Model Y with HD radar.

‘nuf said

I thought Elon made it clear last night that it was going to be vision only?

"it is obvious that our solution with a relatively low-cost inference computer and standard cameras can achieve self-driving. No LiDARs, no radars, ultrasonic."

"It's become very clear that the vision-based approach with end-to-end neural networks is the right solution for scalable autonomy.

And it's really how humans drive. Our entire road network is designed for biological neural nets and eyes. So, naturally, cameras and digital neural nets are the solution to our current road system"

Tesla (TSLA) Q1 2024 Earnings Call Transcript | The Motley Fool

TSLA earnings call for the period ending March 31, 2024.

www.fool.com

 

[OxBrew]

Large numbers of self driven robot vehicles will not coexist with humans driving in the minority for very long at all before the only car you can buy has no steering wheel. Steer that.

Once there the insurance will take care of the rest. Insure robocar $$. Insure oldsmobile $$$$. Make insurance compulsory.

End of self drive except for a few rich lunatics. After a few fatal crashes caused by said lunatics they are banished to private raceways.

I may live to see it but doubtful.

Still, we are only one tech breakthrough from kicking it off and then it would only be 5 years I reckon to no steering wheels in anything new.

Humans driving in the MAJORITY will exist for decades past any fantasy self driving cars.

Do the math on turnover in new car sales. It will take 40 years to get there, minimum, if you're really optimistic about increasing turnover in the future.

Earth just does not have the capacity to make a switch in anything under 40 years, assuming you can actually solve Self Driving, make it affordable for everyone, and work everywhere. (you can't actually solve it.)

Try to get to my cabin down a 2 mile dirt road on July 4th weekend, where I have to get out several times and track down drunk people to move their cars out of the way.

lol times infinity, FSD exists to pump stock prices, nothing else.

 

[JulienW]

This RoboTaxi will be a Model Y with HD radar.

‘nuf said

Elon called it a Cyber taxi on the call so that is a BIG hint it will be all but certain it will be stainless steel. Also on the 1st qt statement Tesla says "Our purpose-built robotaxi product will continue to pursue a revolutionary “unboxed” manufacturing strategy." So an ALL new vehicle.

 

[stopcrazypp]

The perception has already been there for a while. There's a huge tremendous road from there to L4 autonomy with adequate performance and safety

Yeah, even for the previous versions, the biggest problems haven't been perception, it has been the planning.

 

[ngng]

Self driving is legal here in the birthplace of Tesla.

Only if you've applied for the permits, which Tesla has not.

 

[Daniel in SD]

Only if you've applied for the permits, which Tesla has not.

Tesla has a permit.

Autonomous Vehicle Testing Permit Holders - California DMV

Permit holders who are allowed to test autonomous vehicles

www.dmv.ca.gov

 

[diplomat33]

Tesla has a permit.

Autonomous Vehicle Testing Permit Holders - California DMV

Permit holders who are allowed to test autonomous vehicles

www.dmv.ca.gov

Scroll down to driverless permit and deployment permit. Tesla only has the permit to test with a safety driver. They do not have the permit to test driverless. They do not have the permit to deploy AVs. Tesla also does not have the CPUC permit to charge money for a robotaxi ride. So that is 3 additional permits that Tesla still needs to get before they could actually deploy a commercial robotaxi service in CA.

 

[QtownTaxiGreg]

Humans driving in the MAJORITY will exist for decades past any fantasy self driving cars.

That be like, your opinion, man. You're entitled to it. I can imagine an ongoing single decade only before the rout.

The more robocars on the road the more accidents will happen from humans cutting them up knowing they can't get banged into but the poor sod behind the robocar wears it.

Traffic will get slower and slower and slower and slower from all the bad driving by humans. There won't be any point in driving yourself any longer for most people.

The nanny state will mandate autodrive for all new vehicles and then take away the steering wheels. You will still be able to whip out your neural link to drive at walking pace to park your boat.

Between human behaviour, the insurance industry, and the nanny state, humans driving is doomed.

 

[QtownTaxiGreg]

The perception has already been there for a while. There's a huge tremendous road from there to L4 autonomy with adequate performance and safety

I don't think the perception is anywhere near there or the thing would be working like a charm already without eyes on. Beyond mistaking a semi for a billboard but still a long way from accurate.

 

[OxBrew]

That be like, your opinion, man. You're entitled to it. I can imagine an ongoing single decade only before the rout.

The more robocars on the road the more accidents will happen from humans cutting them up knowing they can't get banged into but the poor sod behind the robocar wears it.

Traffic will get slower and slower and slower and slower from all the bad driving by humans. There won't be any point in driving yourself any longer for most people.

The nanny state will mandate autodrive for all new vehicles and then take away the steering wheels. You will still be able to whip out your neural link to drive at walking pace to park your boat.

Between human behaviour, the insurance industry, and the nanny state, humans driving is doomed.

Wow, just totally divorced from reality. I'm sorry we can't have a reasonable discussion.

Global auto production capacity literally, factually, cannot come close to the fantasies you're spouting. On top of that, people don't have the actual economic capacity to buy all those magic new cars that can't actually be produced. We don't have enough income to buy your fantasy cars in 10 or 20 years.

40-60 years starts to be a plausible time frame, with HUGE assumptions about everything.

I refuse to google global auto production per year, vs global fleet size for you. I did it before, it takes 3 minutes, and it obliterates any fever dreams about both EV and self driving cars adoption.

The only way we convert personal transportation to EV by 2030 (the only time frame that matters, beyond that we don't have a habitable planet) is put 99% of the world's battery production into bikes, NEVs and busses.

People, this whole debate is moot. We are doomed by 2040, if not sooner. This summer is going to be -fun- (not actually fun).

I think I'm done with TMC, just a bunch of irrelevant drivel.

 

[HitchHiker71]

Although it makes sense to co-opt as much of an existing design as possible, I would think that the weight and passenger capacity of the robotaxi would dictate which existing Tesla model would serve as the template for the robotaxi. If the robotaxi is a smaller wheelbase vehicle, then it would draw from the Model 3. If it is larger, than possibly the Cybertruck would be the basis.

I haven't read anywhere as to definitely whether stainless steel exterior panels are less expensive to manufacture than painted metal, but if they are, the Cybertaxi could look like the CT.

The Limiting Factor posted a lengthy video about this exact topic for anyone interested in consuming the content:

Long story short, it won't be stainless steel. It'll be a painted surface and/or plastic exterior best estimate.

 

[QtownTaxiGreg]

Wow, just totally divorced from reality. I'm sorry we can't have a reasonable discussion.

Global auto production capacity literally, factually, cannot come close to the fantasies you're spouting. On top of that, people don't have the actual economic capacity to buy all those magic new cars that can't actually be produced. We don't have enough income to buy your fantasy cars in 10 or 20 years.

40-60 years starts to be a plausible time frame, with HUGE assumptions about everything.

I refuse to google global auto production per year, vs global fleet size for you. I did it before, it takes 3 minutes, and it obliterates any fever dreams about both EV and self driving cars adoption.

The only way we convert personal transportation to EV by 2030 (the only time frame that matters, beyond that we don't have a habitable planet) is put 99% of the world's battery production into bikes, NEVs and busses.

People, this whole debate is moot. We are doomed by 2040, if not sooner. This summer is going to be -fun- (not actually fun).

I think I'm done with TMC, just a bunch of irrelevant drivel.

You misconstrue what I am saying. I don't know when robocar proper will arrive. But when it does it's the end of humans driving for the reasons given.

If we are all in communal buses before it gets figured out that wouldn't be a bad thing.

 

[HitchHiker71]

I thought Elon made it clear last night that it was going to be vision only?

"it is obvious that our solution with a relatively low-cost inference computer and standard cameras can achieve self-driving. No LiDARs, no radars, ultrasonic."

"It's become very clear that the vision-based approach with end-to-end neural networks is the right solution for scalable autonomy.

And it's really how humans drive. Our entire road network is designed for biological neural nets and eyes. So, naturally, cameras and digital neural nets are the solution to our current road system"

Tesla (TSLA) Q1 2024 Earnings Call Transcript | The Motley Fool

TSLA earnings call for the period ending March 31, 2024.

www.fool.com

I think it remains to be seen whether Vision Only will provide enough "visibility" for autonomous vehicles - especially with edge cases in moderate to heavy inclement weather are taken into account. Even FSD v12.3.4 still complains a LOT during moderate to heavy rain for example - and on a few occasions it has dis-engaged. Will the system be able to train given enough data to overcome camera distortions during inclement weather? Only time will tell. Radar may required. HW4 has a Tesla designed radar unit IIRC.

 

[spacecoin]

I think it remains to be seen whether Vision Only will provide enough "visibility" for autonomous vehicles -especially with edge cases in moderate to heavy inclement weather are taken into account. Even FSD v12.3.4 still complains a LOT during moderate to heavy rain for example - and on a few occasions it has dis-engaged. Will the system be able to train given enough data to overcome camera distortions during inclement weather? Only time will tell. Radar may required. HW4 has a Tesla designed radar unit IIRC.

I can tell you right now that there will be no autonomy in a wide ODD from Tesla on current HW.

I’m just reporting what Elon said on the call.

 

[brkaus]

I think it remains to be seen whether Vision Only will provide enough "visibility" for autonomous vehicles -especially with edge cases in moderate to heavy inclement weather are taken into account. Even FSD v12.3.4 still complains a LOT during moderate to heavy rain for example - and on a few occasions it has dis-engaged. Will the system be able to train given enough data to overcome camera distortions during inclement weather? Only time will tell. Radar may required. HW4 has a Tesla designed radar unit IIRC.

If it's that bad, people should not be driving. Even those people that put on their hazard flashers and lean over the steering wheel and squint like it helps.

 

[HitchHiker71]

Humans driving in the MAJORITY will exist for decades past any fantasy self driving cars.

Do the math on turnover in new car sales. It will take 40 years to get there, minimum, if you're really optimistic about increasing turnover in the future.

Earth just does not have the capacity to make a switch in anything under 40 years, assuming you can actually solve Self Driving, make it affordable for everyone, and work everywhere. (you can't actually solve it.)

Try to get to my cabin down a 2 mile dirt road on July 4th weekend, where I have to get out several times and track down drunk people to move their cars out of the way.

lol times infinity, FSD exists to pump stock prices, nothing else.

It's really hard to predict the future. Tony Seba studies market disruption for a living and, in general, once a new set of integrative technologies matures enough to disrupt an existing market - the transition generally occurs within 10-20 years. If we look back at the iPhone introduction in 2007 - that's only 16 years ago - and smartphones have certainly disrupted or outright displaced many markets since their introduction. The majority of people use their smartphones for photos, music, and videos just to name a few. While digital cameras and video cameras and music storage devices (iPods/CDs/records/tapes) still exist of course, only hobbyists or professionals tend to still buy dedicated devices for the most part. The vast majority of people use the integrated functions on their smartphone for these functions now.

What Musk/Tesla really seems to be focused on - is what Tony Seba talks about with regard to TaaS (Transportation as a Service), which only becomes possible on a mass scale when true autonomous driving is introduced at scale. This is why Musk is pushing so hard on FSD. Think of Uber/Lyft on steroids with driverless fleets operating 24/7 and offering up transportation on demand whenever you need it - without all of the added costs of electric/gas fuel, insurance, vehicle purchase, maintenance, etc. Given vehicles are the second most costly budget item for most families, if TaaS offers up the ability to save 75-80% back into their monthly budgets - people will start choosing those options. If I'm spending $1000-2000/month on two vehicles all-in, and I could use TaaS for $250-500/month for the vast majority of daily transportation needs - I'd at least consider it. If this type of model actually becomes reality, where I don't agree with others that are touting that human drivers will no longer be on the road, I think pleasure driving will become the norm. People may continue to own weekend cars purely for fun - much like many folks own classic cars today. I don't think this element will go away, if at all, for a really long time.

If TaaS were to take hold, then a driverless vehicle operating 16-20 hours a day can displace a lot of personally owned vehicles that only operate 1-3 hours a day on average. This results in less traffic, a much smaller need for vehicle parking space, and a laundry list of other significant changes when you play it all out. Is this just a dream? Again, only time will tell, but this is clearly what Musk is focused on with Tesla. He literally said on the investors call, if you don't believe in FSD/autonomous driving via AI, don't invest in Tesla stock. It's a sea change without a doubt. I have my doubts and suspicions on the timeline for all of this. Likely this will become an eventual reality. IME the younger generations already don't care near as much about driving themselves around, don't equate driving with individual freedom of choice, etc, like the older generations tend to do. This is therefore likely to be tied up in generational shifts/changes with different beliefs/priorities as much as technical innovation/disruption in other words. Just my two cents of course.

 

[aataskin]

If it's that bad, people should not be driving. Even those people that put on their hazard flashers and lean over the steering wheel and squint like it helps.

Human brain requires far fewer examples to learn and is better suited to processing video information than current generation of computers. And it does it at a fraction of the energy. It's been nearly two years since USS have been ditched and the car still cannot park itself, something that humans driving cars other than Tesla do with the help of USS day in and day out without spending a Watt of energy thinking about it.

 

[stopcrazypp]

I think it remains to be seen whether Vision Only will provide enough "visibility" for autonomous vehicles - especially with edge cases in moderate to heavy inclement weather are taken into account. Even FSD v12.3.4 still complains a LOT during moderate to heavy rain for example - and on a few occasions it has dis-engaged. Will the system be able to train given enough data to overcome camera distortions during inclement weather? Only time will tell. Radar may required. HW4 has a Tesla designed radar unit IIRC.

Note for L4, the car doesn't have to handle inclement weather at all. Most of the issues currently with V12 seems unrelated to perception if you go to the threads, it's stuff like setting the appropriate speed for the road, getting into the correct lane at the correct time for the route, unnatural stopping behavior (may be related to stop sign recall), etc. Basically throwing in more sensors will make no difference for those issues, there's still plenty of stuff to solve unrelated to perception.

 

[DrChaos]

Note for L4, the car doesn't have to handle inclement weather at all. Most of the issues currently with V12 seems unrelated to perception if you go to the threads, it's stuff like setting the appropriate speed for the road, getting into the correct lane at the correct time for the route, unnatural stopping behavior (may be related to stop sign recall), etc. Basically throwing in more sensors will make no difference for those issues, there's still plenty of stuff to solve unrelated to perception.

I see lots of mapping errors: too cheap to buy expensive maps. That would help the lane selection problems.

 

[MarkDave]

It's really hard to predict the future.

It's hard to predict the future but it's easy to look back on history.

I just want to know when in the history of humanity has any problem been solved by minimizing the givens?

I also want to know when in the history of humanity has a robotics problem been solved by minimizing the givens?

Isn't a central tenet to solving any problem, be it a high school math problem, an SAT problem, a robotics problem, an engineering problem, whatever: know your givens, and maximize the number of givens?

In the same vein, I think a "Cybercab" or "Robotaxi" and attempts to solve the Robotaxi issue by minimizing the givens and maximizing the number of unknowns is interesting.

Also, I think it's interesting, at least to me, that many of the great problems of humanity were solved nearly simultaneously by independent people and/or entities. The development of Calculus, for example: two different people, working independently, came up with it at roughly the same time. In this same vein, I see the robotaxi problem being solved simultaneously by various independent entities. Well, I suppose we could say that the robotaxi problem has already been "solved," as in, there exist proofs-of-concept from various companies already. No one has a massively deployed robotaxi network yet, but the proof-of-concept certainly exists.