Take a look at this. A Tesla Robotaxi in Austin pauses, seems to reconsider a right-hand turn through some bollards, and then decides: “I’m just gonna send it.”
Tesla is giving robo-taxis a bad rep. Why is Musk so insistent on optical-only? Sunk cost fallacy? Even the increased processing power required alone could probably already offset the more expensive LIDAR, not to mention the research and loss in revenue.
because it’s cheaper, and the ONLY thing important is making more money. People be damned, as long as he makes another billion.
If it’ll mow down bollards, it’ll mow down toddlers. It can’t distinguish between the two.
Would be the ultimate irony if a robotaxi mows down Elon Muskuito.
Question: what value do self driving taxis add to regular taxis? I can’t think of any. All that money poured into something that solves no problem at all and just creates more, killing human jobs in the process. It’s so stupid.
You don’t have a creepy driver transport you/know your address.
The theory is that you can own a fleet of robotaxis while you have a job without the hassle of maintaining vehicles or employing drivers.
The fact that all self drive taxis are company owned and no plate owners will touch them tells the real story.
The claim that they can undercut human driven taxis I guess, but then as all enshittified services it won’t stay true. They would only do that until people have become used to them then jack up the rates far beyond what cabbies ever were
You only say that because that is what they have done for every other sector they weaseled their way into!
I can see the value of data networks. If every car on the road is self driving and interconnected, they know where every other car is and can cooperatively adjust their speed and routes to basically eliminate traffic. Even if it’s just every car in a taxi fleet, they can still share their sensor data to build a better operating map from multiple angles. The privacy side of that is rough, but it could drastically reduce accidents compared to individual human drivers.
Regular taxis kill a lot more people
Robotaxis have three times the accident rate of human drivers.
For comparison, human drivers in the United States average approximately one police-reported crash every 500,000 miles, according to NHTSA data.
Police reported
In training data, way higher. Also, notice I said they kill fewer people. So far that’s true
Waymos got 220 million miles and 0 deaths so far. They’d be close to 3 by now going by human averages. I dont know far they’ll make it, but you know whenever it does happen, there will be a huge uproar.
I’d be curious to see this quantified, and whether “more” takes into consideration that regular taxis far outnumber Robotaxis
Its been done a couple dozen times, always compared on a per-mile basis.
Decreased costs and increased safety.
The claim “increased safety” is rather ironic in light of the post you’re commenting on.
“Decreased cost” will only be true as long as they still have to compete with human drivers. Once they manage to drive (hah) those out of the market, they’ll start enshittifying and surcharging wherever they can.
“You’re absolutely right to be upset. I should have notified you ahead of time that this section of the highway is closed and I cannot complete the trip you paid for.” “Then exit the highway, dammit!” “Very well. The upgrade has been charged to your card.” “The fuck you mean, upgrade?” “I’m sorry, I should have been clearer on this. You just purchased the Road Map modules for municipalities within 30 miles of our current location in order for me to consider alternate routes not involving the highway. Now please enjoy this entertainment while I select the optimal route.” 90 seconds of ads pass “New route selected. Estimated arrival: three hours and fifty-three minutes. Overtime extensions have been charged preemptively.”
The claim “increased safety” is rather ironic in light of the post you’re commenting on.
Not really?
Are we really so blinded by (understandable) Tesla hate that we’re reducing ourselves to anecdotes to counter data and studies, the thing we’ve been complaining about from conservatives for decades now?
That wasn’t a serious comment on that general topic. I had a paragraph typed up on that, then realised I’m not versed enough in the state of that field to speak on it. It’s to be understood as a tongue-in-cheek allusion to the fact that “they’re safer” is a funny juxtaposition against a post about a safety failure.
If you do want me to contribute my poorly qualified opinion, I do believe that a thoroughly interconnected network of automated mobility can be safer than human control. I suspect that the technical bar for that would be fairly high, but humans are notoriously prone to error, road rage, distraction etc.
The question I’m more concerned about isn’t whether we could, but whether we should. Individual transport being so thoroughly surveilled is even worse than the surveillance of public transport, and in our current world, I don’t think I can safely condone that trend.
The claim “increased safety” is rather ironic in light of the post you’re commenting on.
The post I commented on is a single anecdote where no one was even remotely hurt. These things collectively have millions of miles on them. If you want counter evidence I can show you one of a billion anecdotes of terrible drivers.
“Decreased cost” will only be true as long as they still have to compete with human drivers.
I meant to the provider, not the end user.
The post I commented on is a single anecdote where no one was even remotely hurt.
And that line was just a comment on a juxtaposition I found funny. Don’t take it too seriously.
I can show you one of a billion anecdotes of terrible drivers.
I just replied to another comment to the same effect; I’m aware of terrible drivers being an issue. My concern is less whether it increases the road safety and more whether it helps societal safety.
I meant to the provider, not the end user.
Yeah, I don’t think that should ever be an endorsement criterion for anything at all. Any savings that don’t also save money for the user are just straight up greed, funelling money into the providers’ pockets (and if we’re honest, mostly their executives and shareholders, not their workers).
I don’t think that should ever be an endorsement
It wasn’t. It was an answer to the question that was asked.
I interpreted the original question “what value do self driving taxis add to regular taxis?” as referring to the value for society at large. I concede the potential safety improvement as a counterpoint to “solves no problem at all”, I just don’t really believe that the money savings will manifest for society at large (unless, of course, we manage to get rid of the whole “disrupt, monopolise, exploit, enshittify” mechanism).
Jobs are always going to be replaced by new technology and falling back to omg the jobs, as a means to prevent technology from advancing isnt the right way to think about things.
Maybe you should move to NJ so someone can pump your gas for you, or only buy from farms that hand pick everything.
Ive been in so many taxis driven by humans and feared for my saftey. People have been assaulted by drivers. The AVs can see all around them and will ultimately be definitively safer than a human can ever be. People are terrible drivers as seen by the immense death (1.16 million people in 2025) and destruction we cause behind the wheel.
Does anyone actually have the perfect AV right now? No. But they will figure it out.
Edit: clarity, but also dont forget taxi drivers who scam people by taking longer routes, ive also had my share of having to call them out at times when they think im a tourist.
So-called “ride-share” companies scam people on a bigger scale than individual cabbies ever could. Uber had a whole team devoted to avoiding legal scrutiny.
Robotaxi companies will most likely be just as shady, since they’re backed by the same investors.
As for NJ gas, that’s a false equivalence. A better analogy would be that NJ gas stations are like having a chauffeur, everywhere else is like doing your own driving.
Uber had a whole team devoted to avoiding legal scrutiny.
Pretty sure every company has a legal team…
Definitely no part about protecting jobs, definitely none.
State Sen. Joe Vitale, D-Middlesex, said he’s against this bill because the industry would “eliminate jobs for no other reason than to save a few bucks” and full-service gas would become more expensive.
As a coalition of organized labor, immigrant rights, and disability community advocates, we are coming together to oppose these proposed changes to New Jersey’s unique fueling rules because it will eliminate gas station attendant jobs that working families of diverse backgrounds depend on. While few to none of these workers are unionized, the New Jersey State AFL-CIO seeks to protect all jobs throughout New Jersey. Additionally, eliminating these jobs would be painful for vulnerable populations such as people with disabilities, the elderly and people with children, who rely on attendants to help them gas up. This bill would hurt consumers and provide little in return.
Companies will always try to get the most money out of you sure, but if I hop in a Uber and it says it’s $20.00, it’s going to be $20.00. Is that $20 a rip off, quite possibly, but I know that upfront, unlike a random driver trying to scam me and I don’t know until the end.
As a Keynesian, I’m in favour of gas pump attendants, while gas stations exist.
As an environmentalist, I’m in favour of phasing out fossil fuel.
As a human, I’m opposed to companies that engage in union-busting, labour casualisation, and surge pricing. AFAICT, all the robotaxi companies engage in those practices.
Can we train it to do this with ALPR cameras?
Lol.YOLO! - this tesla
A full minute of processing power and the best it could come up with was crashing through a divider and ending up in the far lane
In fairness, I feel that sometimes.
*flex posts. Comments on reddit locate it here: https://maps.app.goo.gl/VZRjSc325Kg7Vpq17
Yeah it shouldn’t have gone in there because of the white line. But shouldn’t there also be at least one other flex post there blocking it?
Either way, at least it managed to get out appropriately.
I’m honestly surprised that a Tesla Model Y fit in this gap. Maybe it ran over the first bollard a bit?
Edit: And, IDK, “managed to get out appropriately” in this case would unfortunately mean throwing it in reverse and doubling back through the cross-walk, I think. It’s never a good idea to run stuff over to get where you’re going.
I was just looking into this a bit more, “managed to get out appropriately” is potentially actually going through the bollard after it fucked up.
It broke the law going in there, but reversing through the intersection (crosswalk) to get out is also against the law in Austin and many many places.
I know a lot of people do it (and I think I’ve seen a Tesla do it before too) but if you’re stopped at an intersection and you fuck up and are slightly in the crosswalk, you’re in most places, not allowed to back up. If you back up, you could actually be cited for 2 infractions, one blocking the intersection, and then another for the backing up. (edit: and in this case, it’s just a matter of property damage if they go over it, and they’re plastic because they know people will hit them, which is also sad)
I really doubt the car was smart of enough to realize I’m not allowed to reverse, and i could safely drive over this though. I think something else in its decision making process happened with failures all over the place.
I can’t wait to see what the NHSTA report on this one says, hopefully there’s some meaningful info in it.
Maybe we’ll even hear it called home for help and a person drove it over? There was one report where a support person remotely drove it into a chain at slow speed.
Yeah it ran over the first flex post.
You don’t want to back over flex posts, because if there’s one still under the vehicle, it’s going to end up like a spear in the belly of an animal. I once saw someone parked like that and was morbidly curious to see what happened when they tried to leave.
To be fair that looks like a shitty implementation of the bollards.
I wonder if it might also include outdated maps and this is newer infrastructure?
Still not an excuse, it should have seen it and not crossed the white line, but maybe it had conflicting info and chose wrong?
An automatic car should take into account any obstacle in the path whether or not it matches a map.
It did take the obstacles into account though at first. It was going very slowly and managed to squeeze itself into that bad spot, before ultimately realizing it was stuck and deciding to go over them once it had gone too far (I would love to know how the fuck it decided to do that last part though)
What it didn’t take into account from the start was the white line or bike lane sign, but those first bollards on their own it made it past.
If it had bad navigation data, that could have conceivably played a part, but bad map data seems less likely given there’s an image of that intersection from 2024.
Everything about this was still very bad obviously, but even if it wasn’t the case here, bad map data could get you into a bad situation like that if you can drive between the initial bollards.
Jesus Christ, it shouldn’t be squeezing through such a small gap in bollards on a fucking road. That should be a massive red flag that it is going the wrong way on top of the well marked white lines. If some random person did it we would be mocking them for being stupid no matter how well they kept from scratching the paint in the process.
It did a shit job the whole video. Stop making excuses for it.
Im not saying it should, and in this case and have clearly said what it did was wrong.
Im just trying to reason out how or why something like that may have happened, and at first bad map data seemed like it could be part of it wtf went wrong. (edit: even OP said they looked into it, implying they thought that might be a factor too!)
Road says go right. Oh shit its tight but I can make it. Im fucked. YOLO!!!
And yes, it shouldn’t have ignored the bicycle signage or the white line.
Have you never driven around a construction local traffic
roadonly sign that gets placed in the middle of the road or blocking a lane? AVs need to be able to do weird things like this, but in the correct situations, which this was not.Jesus Christ, it shouldn’t be squeezing through such a small gap in bollards on a fucking road.
…
Still not an excuse, it should have seen it
Everything about this was still very bad obviously
???
I actually looked into that…!
The Google Maps interface lets you see when the street was imaged. It’s been that way since 2024, at least. A helpful Redditor shared the link to it.
https://maps.app.goo.gl/VZRjSc325Kg7Vpq17
I’m bad as an automotive journalist. What I lack in skill, I make up for with effort and comedic timing
Oh wow, that’s cool. I didn’t know they had date stamps on it like that. Definitely more than enough time to have updated the maps, especially since I believe Tesla uses data from google for navigation.
