Self-driving cars are often marketed as safer than human drivers, but new data suggests that may not always be the case.

Citing data from the National Highway Traffic Safety Administration (NHTSA), Electrek reports that Tesla disclosed five new crashes involving its robotaxi fleet in Austin. The new data raises concerns about how safe Tesla’s systems really are compared to the average driver.

The incidents included a collision with a fixed object at 17 miles per hour, a crash with a bus while the Tesla vehicle was stopped, a crash with a truck at four miles per hour, and two cases where Tesla vehicles backed into fixed objects at low speeds.

top 97 comments

sorted by: hot top controversial new old
[–] 152 points 9 months ago (40 children)

Optical recognition is inferior and this is not surprising.

  • source
  • hideshow 40 child comments
  • [–] 75 points 9 months ago (38 children)

    Yeah that's well known by now. However, safety through additional radar sensors costs money and they can't have that.

  • source
  • parent
  • hideshow 38 child comments
  • [–] 76 points 9 months ago (20 children)

    Nah, that one's on Elon just being a stubborn bitch and thinking he knows better than everybody else (as usual).

  • source
  • parent
  • hideshow 20 child comments
  • [+] 23 points 9 months ago* (18 children)
  • [–] 53 points 9 months ago (6 children)

    Even if they were, would it not be better to give the car better senses?

    Humans don't have LIDAR because we can't just hook something into a human's brain and have it work. If you can do that with a self-driving car, why cut it down to human senses?

  • source
  • parent
  • hideshow 6 child comments
  • [+] 5 points 9 months ago (3 children)
  • [–] 10 points 9 months ago (3 children)

    Cameras are inferior to human vision in many ways. Especially the ones used on Teslas.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 6 points 9 months ago (4 children)

    I am a Human and there were occasions where I couldn't tell if it's an obstacle on the road or a weird shadow...

  • source
  • parent
  • hideshow 4 child comments
  • [–] 6 points 9 months ago

    And, we humans have built-in binocular vision that we've been training for at least 1.5 decades by the time we're allowed to drive.

    Also, think about what you do in that situation where there's a weird shadow. Slow down, sure. But, also move our heads up and down, side to side, trying to use that powerful binocular vision to get different angles on that strange shadow. How many front-facing cameras does Tesla have. Maybe 3, and one of those is mounted on the bumper? In theory, 3 cameras could give it 3 different "viewpoints" for binocular vision. But, that's not as good as a human driver who can shift their eyes around to multiple points to examine a situation. And, if one of those 3 cameras is obscured (say the one on the bumper) you're down to basic binocular vision without even the ability to take a look from a different angle.

    Plus, we have evidence that Tesla isn't even able to use its cameras to achieve binocular vision. If it worked, it shouldn't have fallen for the Wile E. Coyote trick.

  • source
  • parent
  • [+] 6 points 9 months ago (2 children)
  • [–] 1 point 9 months ago*

    Well I mean if you believe that it is possible in a safe way it's the one thing that Tesla's got going for it compared to Waymo which is way ahead of them. Personally I don't but I can see the sunk cost.

  • source
  • parent
  • [–] 7 points 9 months ago (13 children)

    I don't think it's necessarily about cost. They were removing sensors both before costs rose and supply became more limited with things like the tariffs.

    Too many sensors also causes issues, adding more is not an easy fix. Sensor Fusion is a notoriously difficult part of robotics. It can help with edge cases and verification, but it can also exacerbate issues. Sensors will report different things at some point. Which one gets priority? Is a sensor failing or reporting inaccurate data? How do you determine what is inaccurate if the data is still within normal tolerances?

    More on topic though... My question is why is the robotaxi accident rate different from the regular FSD rate? Ostensibly they should be nearly identical.

  • source
  • parent
  • hideshow 13 child comments
  • [–] 5 points 9 months ago* (last edited 9 months ago) (4 children)

    Regular FSD rate has the driver (you) monitoring the car so there will be less accidents IF you properly stay attentive as you're supposed to be.

    The FSD rides with a saftey monitor (passenger seat) had a button to stop the ride.

    The driverless and no monitor cars have nothing.

    So you get more accidents as you remove that supervision.

    Edit: this would be on the same software versions... it will obviously get better to some extent, so comparing old versions to new versions really only tells us its getting better or worse in relation to the past rates, but in all 3 scenarios there should still be different rates of accidents on the same software.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 3 points 9 months ago (3 children)

    The unsupervised cars are very unlikely to be involved in these crashes yet because according to Robotaxi tracker there was only a single one of those operational and only for the final week of January.

    As you suggest there's a difference in how much the monitor can really do about FSD misbehaving compared to a driver in the driver's seat though. On the other hand they're still forced to have the monitor behind the wheel in California so you wouldn't expect a difference in accident rate based on that there, would be interesting to compare.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 2 points 9 months ago* (last edited 9 months ago) (2 children)

    There are multiple unsupervised cars around now, it was only the 1 before earnings call (that went away), then a few days after earnings they came back and weren't followed by chase cars. There's a handful of videos over many days out there now if you want to watch any. The latest gaffe video I've seen is from last week where it drove into (edit: road closed) construction zone that wasn't blocked off.

    I would still expect a difference between California and people like you and me using it.

    My understanding is that in California, they've been told not to intervene unless necessary, but when someone like us is behind the steering wheel what we consider necessary is going to be different than what they've been told to consider necessary.

    So we would likely intervene much sooner than the saftey driver in California, which would mean we were letting the car get into less situations we perceive to be dicey.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 9 months ago (1 child)

    Yeah I seen that video and another where they went back and forth for an hour in a single unsupervised Tesla. One thing to note is that they are all geofenced to a single extremely limited route that spans about a 20 minute drive along Riverside Dr and S Lamar Blvd with the ability to drive on short sections of some of the crossing streets there, that's it.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 9 months ago (7 children)

    Which one gets priority?

    The one that says there's a danger.

  • source
  • parent
  • hideshow 7 child comments
  • [–] 1 point 9 months ago (6 children)

    Alright, so the radar is detecting a large object in front of the vehicle while travelling at highway speeds. The vision system can see the road is clear.

    So with your assumption of listening to whatever says there's an issue, it slams on the brakes to stop the car. But it's actually an overpass, or overhead sign that the radar is reflecting back from while the road is clear. Now you have phantom braking.

    Now extend that to a sensor or connection failure. The radar or a wiring harness is failing and sporadically reporting back close contacts that don't exist. More phantom braking, and this time with no obvious cause.

  • source
  • parent
  • hideshow 6 child comments
  • [–] 1 point 9 months ago (5 children)

    Now you have phantom braking.

    Phantom braking is better than Wyle E. Coyoteing a wall.

    and this time with no obvious cause.

    Again, better than not braking because another sensor says there's nothing ahead. I would hope that flaky sensors is something that would cause the vehicle to show a "needs service" light or something. But, even without that, if your car is doing phantom braking, I'd hope you'd take it in.

    But, consider your scenario without radar and with only a camera sensor. The vision system "can see the road is clear", and there's no radar sensor to tell it otherwise. Turns out the vision system is buggy, or the lens is broken, or the camera got knocked out of alignment, or whatever. Now it's claiming the road ahead is clear when in fact there's a train currently in the train crossing directly ahead. Boom, now you hit the train. I'd much prefer phantom breaking and having multiple sensors each trying to detect dangers ahead.

  • source
  • parent
  • hideshow 5 child comments
  • [–] 1 point 9 months ago* (4 children)

    FYI, the fake wall was not reproducible on the latest hardware, that test was done on an older HW3 car, not the cars operating as robotaxi which are HW4.

    The new hardware existed at the time, but he chose to use outdated software and hardware for the test.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 1 point 9 months ago (3 children)

    Hardware that was still on the road, or something that had been recalled?

  • source
  • parent
  • hideshow 3 child comments
  • [–] 1 point 9 months ago* (last edited 9 months ago) (2 children)

    As a consumer product, you are responsible and supposed to be paying attention at all times and be ready to take over.

    It is completely acceptable that it does not function perfectly in every scenario and something like a fake wall put on the road causes issues, that is why you need to pay attention.

    There is nothing to recall about this situation.

    If the car is failing on things it shouldn't be, like both Tesla and Waymo failing to properly stop for school busses while in autonomous mode, that does require an update. Alhough ive seen 0 reports of an autonomous Tesla doing this yet only supervised ones.

    A Tesla not stopping for a school bus in supervised mode is acceptable though because the driver is responsible to stop.

    Edit: and note, a problem like the school busses is a visual processing understanding problem. Lidar won't help with that kind or problem.

    Edit: and sorry to be clear, it is hardware still on the road, but I'm saying its acceptable that hardware does it because its not autonomous. If the newer hardware running without supervisors was doing it, that's another story.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 64 points 9 months ago (5 children)

    Wow, thank goodness nobody gutted the authority in charge of making sure that wouldn't happen...

    https://www.theverge.com/news/646797/nhtsa-staffers-office-vehicle-automation-safety-firing-doge-tesla

  • source
  • hideshow 5 child comments
  • [–] 16 points 9 months ago (4 children)

    The AI companies put out a presser a few years back that said "Um, aktuly, its the humans who are bad drivers" and everyone ate that shit up with a spoon.

    So now you've got Waymos blowing through red lights and getting stuck on train tracks, because "fuck you fuck you stop fighting the innovation we're creatively disruptive we do what we want".

  • source
  • parent
  • hideshow 4 child comments
  • [–] -2 points 9 months ago* (last edited 9 months ago) (3 children)

    That doesn't mean that waymo is more error prone than human drivers.

    Humans are awful at driving and do stuff like stop on train tracks and blow through red lights all the time.

    Edit: I'm still waiting on someone to prove me wrong with actual data, because this article is about Tesla, not Waymo.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 4 points 9 months ago (2 children)

    The article is about Robotaxis crashing 4x as much as human driven cars.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 9 months ago

    Meanwhile, Waymo has logged over 127 million fully driverless miles, with no safety driver, no monitor, no chase car, and independent research shows Waymo reduces injury-causing crashes by 80% and serious-injury crashes by 91% compared to human drivers. Waymo reports 51 incidents in Austin alone in this same NHTSA database, but its fleet has driven orders of magnitude more miles in the city than Tesla’s supervised “robotaxis.”

    Point me to where Waymo taxis are just as bad as Tesla.

  • source
  • parent
  • [–] 60 points 9 months ago
    [–] 40 points 9 months ago (2 children)

    They'll work perfectly as soon as AI space data center robots go to Mars. I'd say a Robovan will be able to tow a roadster from New York to Hong Kong by... probably July. July or November at the latest.

  • source
  • hideshow 2 child comments
  • [–] 9 points 9 months ago (1 child)

    I really fucking hate how his fans can just listen to him lie like this over and over and it doesn't affect their opinion of him. I remember falling for it a couple times before I started asking "Is this like the last time you promised dates?"

    By that time it was a moot point, however, because that "pedo guy" comment was just around the corner. Now anyone who likes him after that needs to go to therapy to figure out a few things.

    I won't comment on people who support him after the other things.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 38 points 9 months ago (6 children)

    Use lidar you ketamine saturated motherfucker

  • source
  • hideshow 6 child comments
  • [–] 11 points 9 months ago (3 children)

    Can't do that. Then he would have to upgrade all legacy cars. And he is missing the lidar dataset.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 10 points 9 months ago (2 children)

    The best time to add lidar would have been years ago, the second best time is right now. I don't think he would have to update the old cars, it could just be part of the hardware V5 package. He's obviously comfortable with having customers beta testing production vehicles so he can start creating a lidar set now or he can continue failing to make reliable self-driving cars.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 7 points 9 months ago

    Agree, but since he stated multiple time that all cars since xxx years were hardware capable of L5 self-driving next year (no need to precise the year, the statement is repeated every year), adding LIDAR now would be opening the way to a major class action. So he painted himself in a corner, and like all gigantic-ego idiots, he doubles down every time he's asked.

  • source
  • parent
  • [–] 24 points 9 months ago (7 children)
  • [–] 18 points 9 months ago (6 children)
  • [–] 5 points 9 months ago (5 children)

    Are they even insured like typical insurance?

    If Tesla owns it, don't they just pay out of pocket as needed, they don't actually have a monthly payment to themselves or anything?

  • source
  • parent
  • hideshow 5 child comments
  • [–] 3 points 9 months ago* (2 children)

    There's no way, then why can't I drive around "uninsured" with the promise I'll pay out of my pocket for any damage.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 6 points 9 months ago* (last edited 9 months ago) (1 child)

    I'm not sure if this is available everywhere, but in California you can put up a bond in lieu of insurance with the DMV, either with your own money or with a surety bond company.

    So you can do it, they just require proof in the form of the bond that the money is available when needed. They won't just take your word for it. They might take the word of a company with a $1.3 trillion market cap, which is probably a bad idea. You get hit with one of these fuckers and instead of the admittedly shitty but known process of dealing with an insurance company, now you have to deal with a huge company that doesn't want to admit their dumb camera only system is at fault.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 3 points 9 months ago

    This is what I was thinking was possible ya. If they have enough money, they could just cover it themselves.

    I really don't know if that can be done everywhere though.

    And ya, as an individual self insuring this way, it would be a disaster going against a behemoth like Tesla.

  • source
  • parent
  • [–] 23 points 9 months ago (5 children)
  • [–] 11 points 9 months ago (4 children)
  • [–] 2 points 9 months ago* (3 children)

    Mentally challenged POS Nazi Pedophile

  • source
  • parent
  • hideshow 3 child comments
  • [–] 1 point 9 months ago (2 children)

    Hey, let's not denigrate the mentally challenged now.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 0 points 9 months ago (1 child)

    I think it needs to be acknowledged, Musk is mentally challenged, has mental illness, call it what you will, he has a diagnosable condition, and as such has been tooled by Peter Thiel. Perhaps blackmailed as well, alongside the carrots of going along. His faults aren't solely because of the mental illness perhaps, but it's a factor in why he does what he does, and how acts, made worse by his drug use.

    If he was an ordinary person he would be recognized as such, if not committed involuntarily at some point, but being filthy rich he's just eccentric as far as the system is concerned. Just a fat out of shape slob throwing seig heils representing nazis that championed the idea of killing undesirables including fat out of shape slobs. But he gets a pass because he's rich. Nazis also killed mentally ill people, just as a matter of course.

    Point being Musk would be eliminated by the Nazis he wants to resurrect and put in absolute power, that he was compelled to represent and put in power. He obviously didn't appreciate whomever gave him that black eye, and I would guess it was thiel or yarvin or one of them, the actual brains behind doge and musk that are using him, that groomed those kids from sex and drug fueled parties, to secretly export all fed agency data to their secret data servers, under the guise of cutting costs.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 19 points 9 months ago

    Clearly, AI isn't just challenging human performance, it's exceeding it. Four times the crash rate is just the beginning. Just imagine the crash rate when super intelligence comes!

    🚘💥🚗

  • source
  • [–] 18 points 9 months ago (1 child)

    I do (sarcastically) love knowing Leave the World Behind is a documentary.

  • source
  • hideshow 1 child comment
  • [–] 13 points 9 months ago*

    Even for the first piss poor epigone of Neuromancer, the name "Robotaxi" would've been laughed at.

    Mulon Esk made the dumbest name happen for the xth time.

  • source
  • [–] 10 points 9 months ago (7 children)

    Got this saved next time someone tells me that a robot can drive better than a human. They almost had me there, but data doesn’t lie. 

  • source
  • hideshow 7 child comments
  • [–] 8 points 9 months ago

    A robot can theoretically drive better than a human because emotions and boredom don't have to be involved. But we aren't there yet and Teslas are trying to solve the hard mode of pure vision without range finding.

    Also, I suspect that the ones we have are set up purely as NNs where everything is determined by the training, which likely means there's some random-ass behaviour for rare edge cases where it "thinks" slamming on the accelerator is as good an option as anything else but since it's a black box no one really understands, there's no way to tell until someone ends up in that position.

    The tech still belongs in universities, not on public roads as a commercial product/service. Certainly not by the type of people who would at any point say, "fuck it, good enough, ship it like that", which seems to be most of the tech industry these days.

  • source
  • parent
  • [–] 5 points 9 months ago* (last edited 9 months ago) (2 children)

    This is more specific to Tesla than self driving in general, as Musk decided that additional sensors (like LiDAR and RADAR on other self driving vehicles) are a problem. Publicly he’s said that it’s because of sensor contention - that if the RADAR and cameras disagree, then the car gets confused.

    Of course that raises the problem that when the camera or image recognition is wrong, there’s nothing to tell the car otherwise, like the number of Tesla drivers decapitated by trailers that the car didn’t see. Additionally, I assume Teslas have accelerometers so either the self driving model is ignoring potential collisions or it’s still doing sensor fusion.

    Not to mention we humans have multiple senses that we use when driving; this is one reason why steering wheels still mostly use mechanical linkages - we can “feel” the road, we can detect when the wheels lose traction, we can feel inertia as we go around a corner too fast. On a related tangent, the Tesla Cybertruck uses steer-by-wire instead of a mechanical linkage.

    This is why many (including myself) believe Tesla has a much worse safety record than Waymo. I’ve seen enough drunk and distracted drivers to believe that humans will always drive better than a human robot. Don’t get me wrong, I still have concerns about the technology, but Musk and Tesla has a history of ignoring safety concerns - see the number of deaths related to his desire to have non-mechanical handles and hide the mechanical backup.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 9 points 9 months ago (4 children)

    They're 4 times as capable of crashing as a human driver. How efficient!

  • source
  • hideshow 4 child comments
  • [–] 3 points 9 months ago (3 children)

    Whaaa how do you do subscript (?) text! Aaaaah!

  • source
  • parent
  • hideshow 3 child comments
  • [–] 9 points 9 months ago
    [–] 7 points 9 months ago

    I thought ai was going to replace all jobs in a year and a half

  • source
  • [–] 7 points 9 months ago

    This is a really funny thing to see a few scrolls down from an article about Tesla's first drivingwheelless vehicle and finally "solving autonomous driving"

  • source
  • [–] 7 points 9 months ago (5 children)

    Hopefully, the French judicial system will throw his worthless pedo neo-Nazi ass into prison.

  • source
  • hideshow 5 child comments
  • [–] 6 points 9 months ago (4 children)

    Curious how the French have jurisdiction over Austin?

  • source
  • parent
  • hideshow 4 child comments
  • [–] 10 points 9 months ago* (last edited 9 months ago) (1 child)

    "In the recorded history of what is now the U.S. state of Texas, all or parts of Texas have been claimed by six countries: France, Spain, Mexico, the Republic of Texas, the Confederacy during the Civil War, and the United States of America." [ https://en.wikipedia.org/wiki/History_of_Texas ]

    Does THIS answer your question? That's right, Captain. Put'em up! Or shall I say... peut demains apres?

    EDIT: It's the Epstein files. The French are investigating. Musk is in them.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 6 points 9 months ago (1 child)

    Only 4x? Wao, they're way better than I expected then.

  • source
  • hideshow 1 child comment
  • [–] 6 points 9 months ago*

    It's Austin. The traffic is so shitty you can't go fast enough to get in a wreck most of the time.

    I live in the area, and can confirm anecdotally that the Teslas are bad drivers and the Waymos generally are excellent.

  • source
  • parent
  • [–] 5 points 9 months ago*

    Rogan: so eaaauhm, yeah that’s definitely a thing huh? But you know all progress must go uphill without breaking a few eggs…right?

    Musk: makes that stupid nazi face where he’s smoking weed So we’re going to make Grok a subscription model that watches you sleep in your car as we plug you into the bio battery of your Tesla. Then your mind gets used to train AI models as you’re driving. But you know, I’m expecting that to work last month, give or take a year or 10.

    Rogan: Pluggin in huh? How’s that work?

    Musk: Either a port in the back of your arm or an arm up your back, not sure yet.

    Rogan: Wow, … so anyway wanna do some dmt?

    We can plug it in if you want.

  • source
  • load more comments
    view more: next ›