Tesla robotaxis launch in Austin with $4.20 invite-only service and human "safety monitors"

midian182

Posts: 10,780   +142
Staff member
What just happened? Tesla's long-awaited robotaxi service finally launched yesterday (June 23) in Austin, Texas. So far, it has been limited to invited participants and not the general public, and passengers aren't alone in the vehicles; Tesla's "safety monitors" sit in the front passenger seat ready to take over if something goes wrong – despite Elon Musk promising the service would launch unsupervised.

Musk's "tentatively, June 22" response early this month to an X user's question about when Tesla Robotaxi service would begin proved accurate.

The rides started yesterday, limited to a geofenced section of Austin, avoiding the likes of highways and complex intersections. The fleet of ten to twenty modified Model Y vehicles currently operate between 6 am to 12 am in the area, and there's a warning that they may not be available in inclement weather.

The service is limited to invitees made up of Tesla influencers and fans; they can bring a guest who is over 18. The videos they've been posting show the safety monitors in the passenger seat, a contrast to Musk's previous promise that the taxis would launch "unsupervised, no one in the car." The CEO has admitted that Tesla is being "super paranoid about safety," in Texas, where a new law that requires state permits for self-driving vehicles comes into effect on September 1. It's unclear if or when Tesla will stop using the monitors.

Musk said that initial Robotaxi customers would pay a flat fee of $4.20, a weed culture reference that he's fond of using.

The taxis use Tesla's Full Self-Driving (FSD) system and eight cameras. The company claims the technology allows self-driving to be deployed anywhere it's approved, and does not require expensive, specialized equipment or extensive mapping of service areas.

Reviews so far have been positive – that's to be expected, given the allegiances of those invited . The biggest complaint seems to be trying to find the pickup locations of the waiting Robotaxis. However, a 12-second clip of a journey by podcaster Rob Maurer seems to show the taxi repeatedly trying to swerve into the wrong lane.

The safety monitors ask riders to show their robotaxi apps, which are said to be "basically Uber," to prove their identities once they get in the vehicles. The monitors are then silent for the rest of the journey.

Rides are initiated by customers via the rear screen or the app. These Tesla users can see their preferred music apps on the rear screen, complete with playlists and saved tracks. There are also remote operators who can be contacted about any issues via a support button.

Tesla is entering a competitive market. Waymo has over 1,500 driverless vehicles in several states and plans to increase both its vehicle count and number of locations in the future. Musk has claimed that Tesla will have over a thousand driverless vehicles on the road "within a few months."

Permalink to story:

 
All the best everyone. I don't agree to the moderation of my comments at the level which Techspot demands. Cheers.
 
Would be good investment to buy such a self driven vehicle (of any brand) and just put it on the street to create cash income for you 24/7

Of course, at Shanghai auto show over 100 brands of Chinese electric vehicles were shown. They would literally conquer the world with it's biggest revolution if they also had self driving. Of course in 20 years would be the norm.
 
However, a 12-second clip of a journey by podcaster Rob Maurer seems to show the taxi repeatedly trying to swerve into the wrong lane.
I saw that, there’s definitely some work to be done. To me it looks like something a human driver would do if they weren’t familiar with roads (and a human did the same thing later in the same video).

The Robotaxi got in a left turning lane too early and realized it needed to turn left at the next intersection instead. It eventually made a lane change in the intersection so it could go straight when it was safe to do so. What it handled wrong was jerking the steering wheel left and right as it was indecisive whether it could switch lanes to go straight or it should continue with the early left turn. Instead, it should’ve slowed down, put on the tight blinker, and if it couldn’t switch lanes in time continue to make the left turn and reroute.
 
I saw that, there’s definitely some work to be done. To me it looks like something a human driver would do if they weren’t familiar with roads (and a human did the same thing later in the same video).

The Robotaxi got in a left turning lane too early and realized it needed to turn left at the next intersection instead. It eventually made a lane change in the intersection so it could go straight when it was safe to do so. What it handled wrong was jerking the steering wheel left and right as it was indecisive whether it could switch lanes to go straight or it should continue with the early left turn. Instead, it should’ve slowed down, put on the tight blinker, and if it couldn’t switch lanes in time continue to make the left turn and reroute.
No surprises that there is work to be done, but I don't agree with the justification of the action through saying "looks like something a human driver would do if they weren’t familiar with roads".

Good luck to those who use this. Given its base software is likely the "autopilot" software, I would not put much faith in a robotaxi. At least Texas is making laws to regulate such vehicles.
 
"Waymo has over 1,500 driverless vehicles in several states and plans to increase both its vehicle count and number of locations in the future"

Tesla's can become Robotaxi capable with an upgrade. Waymo has to scale seriously to compete.
 
Vaporware

Time will tell if you are right?

"Vaporware refers to a product, typically software or hardware, that is announced or promoted but never actually released or completed. It’s often hyped up with marketing or promises but fails to materialize, either due to development issues, lack of resources, or intentional deception to gauge interest or mislead consumers"

 
"Waymo has over 1,500 driverless vehicles in several states and plans to increase both its vehicle count and number of locations in the future"

Tesla's can become Robotaxi capable with an upgrade. Waymo has to scale seriously to compete.
No to mention the cost of each Waymo... From Waymo's co-CEO: https://www.nytimes.com/2024/09/04/technology/waymo-expansion-alphabet.html
New York Times said:
The equipment on Waymo’s fifth-generation robot taxis — electric Jaguar I-Pace vehicles — costs as much as $100,000, Dmitri Dolgov, Waymo’s co-chief executive, said on a podcast in February. And while Waymo doesn’t have to pay for drivers, it employs technicians behind the scenes to monitor rides.
Meanwhile a Model Y (all hardware included) costs about $40k each after excluding profits, and Tesla is working on a two seater stripped down model that's expected to cost $30k or less. The lower the cost of the vehicle, the faster they break even/greater ROI of each one OR if they wanted to dominate the market/increase revenue they can charge less money for each mile driven.

Waymo does have the ability to switch vehicles and manufacturers, but they don't have the ability to improve vehicle longevity, cost, or integrate sensors into models themselves. They're more likely to sell the software to automakers, and that would also be an increase in base cost versus what Tesla can offer (and hardware for that matter if it relies on Lidar).
No surprises that there is work to be done, but I don't agree with the justification of the action through saying "looks like something a human driver would do if they weren’t familiar with roads".

Good luck to those who use this. Given its base software is likely the "autopilot" software, I would not put much faith in a robotaxi. At least Texas is making laws to regulate such vehicles.
It's different software from Autopilot (which was modular instead of end to end AI trained). They dropped all of the Autopilot software (even the visualizations are different now).

And I've done the same thing myself, when I tried to follow GPS and didn't realize the road I needed to turn on was not the one I thought it was. It involves coordinating what you see on a map and what is shown on the roads. And if a human driver wouldn't do it, then why did this one do the same thing in the same video/drive? See the black crossover in the left turning lane (it was worse because it came close to hitting an oncoming vehicle by braking first before changing his mind):
 
The most telling thing to come out of this experiment will be the liability and compensation from the first accident caused by one of the robotaxis and Tesla's exposure. If/when someone is injured or killed because of the actions of one of these cars, how much will Tesla settle for, and how often? Will an injured party go after the "deep pockets" and refuse to settle? In most cases, it's pretty easy to settle. Most humans don't have the resources to sue for millions of dollars. But Tesla? Let's see what happens.
 
And I've done the same thing myself, when I tried to follow GPS and didn't realize the road I needed to turn on was not the one I thought it was. It involves coordinating what you see on a map and what is shown on the roads. And if a human driver wouldn't do it, then why did this one do the same thing in the same video/drive? See the black crossover in the left turning lane (it was worse because it came close to hitting an oncoming vehicle by braking first before changing his mind):
For me, all it takes is to pay attention to where the vehicle marker is on the map. In my case, if it is at the road where it wants me to turn, then that is where I turn. The operation is not complex, IMO, but requires map data to be accurate.

;)
 
No surprises that there is work to be done, but I don't agree with the justification of the action through saying "looks like something a human driver would do if they weren’t familiar with roads".

Good luck to those who use this. Given its base software is likely the "autopilot" software, I would not put much faith in a robotaxi. At least Texas is making laws to regulate such vehicles.

This is based on a newer version (including newer hardware) of the FSD software. Not Autopilot, which is lane keeping/autosteer and adaptive cruise only... without navigation.
 
This is based on a newer version (including newer hardware) of the FSD software. Not Autopilot, which is lane keeping/autosteer and adaptive cruise only... without navigation.
I have lane keeping/autosteer and adaptive cruise only... without navigation on my 24' Prius Prime, although, someone tried to tell me it wasn't the equivalent. :rolleyes:
As a matter of fact, all Toyota's have this as part of Toyota's "Safety Sense" package. I can also switch on another function which will bring my Prius Prime safely to a stop in city driving without my having to do anything.
 
Back