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!

 

[mborkow]

I am just curios with respect to timing. I personally have found the length of time between announcements and action to be too long. It gives way to much power to the media to keep talking about how this or that would never happen. I do not see FSD in the next year happening (could be wrong) and would like to see maybe a 6 month or max one year gap between announcing products and getting them out there. Model x was what 5 years? Cybertruck , same?

This announcement is nothing more than an attempt to boost flagging stock price.

 

[stopcrazypp]

They must be either vested share holders or never tested FSD!

Or they could have tested in California in areas the system is "overfit" for and been thoroughly impressed by the progress. It seems the experience of people are fairly different depending on where they are. There are those that saw night and day differences between v10/v11 and V12, and others where V12 still performed poorly.

 

[DanCar]

Or they could have tested in California in areas the system is "overfit" for and been thoroughly impressed by the progress. It seems the experience of people are fairly different depending on where they are. There are those that saw night and day differences between v10/v11 and V12, and others where V12 still performed poorly.

I live and drive in the over fitted area. It is no where near fully self driving. Just a bunch of shills / overly enthusiastic peeps. I might change my mind some day if it handled potholes.

 

[aataskin]

It will be a Robotaxi (Supervised) car rental service. Just another brand name, just like the FSD (Supervised) autopilot.

 

[JulienW]

It will be a Robotaxi (Supervised) car rental service. Just another brand name, just like the FSD (Supervised) autopilot.

So kinda like the Vagas' Loop full autonomy with a Tesla driver in every car.

 

[Joey D]

By when do you expect cancer diagnosis in radiology to be unsupervised autonomous? Why is this relevant? It's a form a simpler computer vision application that still isn't autonomous. It's safety critical, but not time critical. It uses still images instead of video and real time decision making.

I believe you could argue that right now, AI does a better job at diagnosing based on images alone. I would also argue that diagnosis done by professional humans today is not solely based on imagery

Waymo has hardware that sees about three football fields in all directions and in most weather, and they aren't still able to remove the human in the loop at highway speeds. Just saying.

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

 

[azred]

Based on some of Elon’s past, rock-solid predictions like the one in early 2017 that there would be a fully autonomous FSD trip CA to NY by Dec 2017, I would guess he will predict Robotaxis will flood the country in no more than two years.

Funny guy that Elon.

As for the shape of the vehicle, I predict butt-ugly on a par with the hideous pickup truck.

 

[azred]

So kinda like the Vagas' Loop full autonomy with a Tesla driver in every car.

Hmm, Robotaxi Driver might be a good part-time gig. And based on the speed of completing FSD, it would probably be a many years long gig.

 

[stopcrazypp]

By when do you expect cancer diagnosis in radiology to be unsupervised autonomous? Why is this relevant? It's a form a simpler computer vision application that still isn't autonomous. It's safety critical, but not time critical. It uses still images instead of video and real time decision making.

On that subject, even though it's still in infant stages, there are already studies that have AI perform just as good as a radiologist.

Advancing Breast Cancer Detection with AI - Google Health

From the government article, there are multiple examples given where AI performs better than human radiologists at not only detecting cancer or pre-cancer signs, but also at eliminating false positives.
Can Artificial Intelligence Help See Cancer in New Ways?

But as like most AI work, the capabilities can vary a lot depending on what kind of investments are made in a particular field and by a particular company/team, just like how the resources for AVs are not evenly distributed and not shared.

 

[spacecoin]

On that subject, even though it's still in infant stages, there are already studies that have AI perform just as good as a radiologist.

Advancing Breast Cancer Detection with AI - Google Health

From the government article, there are multiple examples given where AI performs better than human radiologists at not only detecting cancer or pre-cancer signs, but also at eliminating false positives.
Can Artificial Intelligence Help See Cancer in New Ways?

But as like most AI work, the capabilities can vary a lot depending on what kind of investments are made in a particular field and by a particular company/team, just like how the resources for AVs are not evenly distributed and not shared.

The point was that there’s still a human in the loop. It’s not unsupervised.

 

[petercc9]

Or they could have tested in California in areas the system is "overfit" for and been thoroughly impressed by the progress. It seems the experience of people are fairly different depending on where they are. There are those that saw night and day differences between v10/v11 and V12, and others where V12 still performed poorly.

As I said maybe most of the training videos have been done in dense narrow urban areas where things in more open areas are uncommon..like multiple left turn lanes or wide 4 lane streets.

We may need to do a poll to see where people who think v12 is awesome vs those who think it is nowhere close to working as it seem we have some extreme camps here on its current status and viability.

This is the typical type of experience I see.

My car drives around in my suburb fine...just like this. Even hands off the wheel for long stretches of time.

 

[JulienW]

.....We may need to do a poll to see where people who think v12 is awesome live vs those who think it is nowhere close to working as it seem we have some extreme camps here on its status.

Nowhere close to working?????......but it is and had been working for years. Originally we had v10.2 and it worked. One could argue that v10.3 didn't work but......other than that all versions have worked.

You could ask subjectively how well it works but asking if it is working is a false question since it 100% objective works.

 

[DanCar]

... asking if it is working is a false question since it 100% objective works.

Doesn't work as full self driving (FSD). Needs supervision and takeovers.
Lots of cars with adaptive cruise control and lane keep assist work well too. Minor supervision and occasional takeovers is all that is needed.

 

[JulienW]

Doesn't work as full self driving (FSD). Needs supervision and takeovers.
Lots of cars with adaptive cruise control and lane keep assist work well too. Minor supervision and occasional takeovers is all that is needed.

Go back and read what I quoted. It was about making a poll about what people think about V12. Is V12 working "awesome" or "nowhere close to working". V12 is FSDS L2 and it does work. V12 FSDS is not intended to be or ever will be a L4 system.

 

[petercc9]

Go back and read what I quoted. It was about making a poll about what people think about V12. Is V12 working "awesome" or "nowhere close to working". V12 is FSDS L2 and it does work. V12 FSDS is not intended to be or ever will be a L4 system.

ok, YEARS and YEARS away from being autonomous in typical driving scenarios.

 

[JulienW]

ok, YEARS and YEARS away from being autonomous in typical driving scenarios.

...but you did say "current status" and to me that means what it is currently. So that is what I went by.

 

[DanCar]

Go back and read what I quoted. It was about making a poll about what people think about V12. Is V12 working "awesome" or "nowhere close to working". V12 is FSDS L2 and it does work. V12 FSDS is not intended to be or ever will be a L4 system.

You were quoting Peter and that I doubt that is his context. What is your source for saying v12 is never intended to be an L4 system? If that is true elon should refund us our money since that is what we were promised.

 

[HitchHiker71]

It will be manned robotaxies for quite a while and initially only in geofenced heavy urban centers. It will likely be years before any actual true robotaxi comes to market.

Tesla has a bunch of these JDs posted this week:

These JDs are posted in major urban areas, along with ADAS test specialist job postings. This all but proves where the initial focus will be come 8/8. Tesla has been taking M3 leasebacks and holding them back from resale for this purpose. These prototype/adas jobs will use the M3 leasebacks as the beginnings of the manned robotaxi fleet. The robotaxi isn’t even built yet - that’s likely many months if not years out as well.

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).

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.

 

[QtownTaxiGreg]

As I said maybe most of the training videos have been done in dense narrow urban areas where things in more open areas are uncommon..like multiple left turn lanes or wide 4 lane streets.

We may need to do a poll to see where people who think v12 is awesome vs those who think it is nowhere close to working as it seem we have some extreme camps here on its current status and viability.

This is the typical type of experience I see.

My car drives around in my suburb fine...just like this. Even hands off the wheel for long stretches of time.

Towards the end it failed to anticipate the back end of the heavy vehicle moving the opposite direction due to the significany rear axle overhang. Have the guy a proper fright. You can't tell me it timed its overtake with swerve to perfection. Rookie driver stuff.

 

[kabin]

Towards the end it failed to anticipate the back end of the heavy vehicle moving the opposite direction due to the significany rear axle overhang. Have the guy a proper fright. You can't tell me it timed its overtake with swerve to perfection. Rookie driver stuff.

Yeppers. No way they have training data for that scenario. Like you said the aft end moves closer to the passing vehicle as it turns. It reminds of new RV owners crushing their rigs when they make sharp turns as gas stations.