345
Tesla Robotaxi Mows Down Bollards & Keeps Driving, After “Impeccable” Record Touted on Stock Call (fuelarc.com)
submitted 1 month ago by to c/nottheonion@lemmy.world
60 comments fedilink hide all child comments
 

I mean, I guess "impeccable safety record" is in the eye of the beholder?

you are viewing a single comment's thread
view the rest of the comments
[–] 31 points 1 month ago* (22 children)

Just the fact that it crosses that double wide white line, that means under no circumstance do not cross this line! Only to end behind a perfectly visible veritable wall of bollards! All this in high daylight too!
The strongest danger of Tesla FSD I think, is that in most situations it drives very confidently, and very well. So people begin to trust it.
But then suddenly it does something insanely stupid like this, or cross a red light without hesitation, or completely ignore a stop sign.
This is when you die. Not the 90% of the time it drives fine.

  • permalink
  • fedilink
  • source
  • hideshow 22 child comments
    •
  • [–] 11 points 1 month ago (17 children)

    Remember that the bar for self driving cars isn't perfect driving, it's better-than-human driving, and humans are generally pretty bad at driving, so that bar isn't as high as you probably hope it is.

    I'm not suggesting they've reached that bar, only that "sometimes a mistake happens that results in death" is also a problem with human driving, but we know that self driving cars don't get drunk or distracted by their phones, so they've already got a leg up on the competition.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 17 child comments
    •
  • [–] [S] 11 points 1 month ago (2 children)

    The human safety supervisor of a Tesla Robotaxi who crashed was apparently sleep deprived and having a mental health episode, so, if you think self-driving cars will insulate you from human foibles, just you wait. The futures going to get real weird!

    Broadly I agree, though. The goal is better-than-human. It really shouldn't ever cross a solid line, though.

    That was this case, it's all self-reported so take it with a grain of salt, but it was interesting food for thought: https://electrek.co/2026/07/28/tesla-self-driving-manager-rolling-hazards-lawsuit/

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 7 points 1 month ago (1 child)

    The goal is better-than-human. It really shouldn’t ever cross a solid line, though.

    The goal should be far fewer cars of any sort on the street. Investing in self driving cars instead of better infrastructure is stupid. Especially when the cars are privately owned

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 3 points 1 month ago (6 children)

    But take that to the next step. Let’s say it reaches the point where it drives better than humans. The problem is it’s not perfect, and has different weaknesses than humans. Even after self-driving is statistically safer, there will be accidents where a human could have done better. Would we accept it under those circumstances? Because those are the circumstances

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] 1 point 1 month ago (5 children)

    Even after self-driving is statistically safer, there will be accidents where a human could have done better.

    I think "could have" is doing a lot of heavy lifting here and it's making the question seem more complicated than it is. You can't make decisions based on hindsight from specific examples. If, hypothetically, self driving cars are statistically safer, then they're safer, regardless of any specific anecdote that suggests otherwise.

    Maybe I'm not grasping what you're saying?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] 2 points 1 month ago (4 children)

    By “statistically safer” I mean as an overall metric. However there will still be specific scenarios where self-driving is less safe, or even just people believe they could have done better. People will naturally focus on those specific scenarios and reject self-driving even when it is overall safer

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 0 points 1 month ago (3 children)

    Can you give me a hypothetical example? I'm not understanding the scenario you're referring to.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 3 points 1 month ago (2 children)

    As a fake example, let’s say Tesla full self-driving is overall safer than humans, however it doesn’t recognize traffic poles and will try to drive through them. People would probably reject FSD as not yet ready despite the evidence that it was safer than humans

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 1 month ago (1 child)

    I appreciate the clarification, but I doubt there will be a way for both scenarios to exist simultaneously; that is to say, I think any sufficiently common "exception" would result in self driving cars not being statistically safer than human drivers.

    It would have to be something sufficiently obscure, like, I dunno, going down a eastern-facing switchback road in the fog at sunrise. (Maybe too obscure.)

    That being said, I remember reading an article or maybe watching a video about the challenges for making a self driving car feel safer for people. Stuff like turn speed on main roads vs in residential neighborhoods. The forces being felt are exactly the same, and the safety is exactly the same, but (apparently) people feel like one is acceptable but the other is dangerous.

    If your ultimate question was along these lines, then yes, I agree there is more than just statistics that will be necessary before people are comfortable with mass adoption of self driving cars. And not for nothing but that tool Musk isn't helping anything by putting out "self driving" cars before they're actually ready.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 1 month ago

    making a self driving car feel safer

    Without judging the effectiveness, Tesla has made great progress on this. If you read through recent FSD reviews you’ll see many claim things like “the car drives more confidently”

  • permalink
  • fedilink
  • source
  • parent
  • [–] 3 points 1 month ago (6 children)

    A human can drive decades without accidents. I've done that myself. Day and night, rain, snow and fog.
    Most autonomous systems are not allowed to be used in any of those conditions!
    So there's a long way for FSD to match a competent driver.

    “sometimes a mistake happens that results in death” is also a problem with human driving

    The shit show Tesla is doing is nowhere near an average human driver, and the FSD has been shown to be lethal multiple times even under good circumstances that shouldn't pose any problems. All Tesla has is as good as a human driver 90% of the time. Then it fucks up. Tesla FSD would for sure not be able to get a drivers license here in Denmark.
    IDK the rules in USA. But if a reasonable drivers test is included, I think Tesla FSD would be very likely to fail it.

    So Tesla is definitely not anywhere near "better than a human". Not even under optimal conditions like daylight and no bad weather.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] 3 points 1 month ago (1 child)

    Aside from the interactive voice response, Tesla self-driving would easily pass a road test here. Although I never tried to get it to parallel park and certainly not “parallel park between those two cars”.

    The driving part is easy though. What I remember of the driving test was basic skills and confidence, and self-driving has those. Real life driving is a plethora of unexpected stuff and edge cases: that’s where self-driving still needs work but is also not covered in a road test

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 3 points 1 month ago

    OK, to be fair the roads in USA are 100 times easier to navigate than in Europe. At least when I compare Californian cities to Copenhagen and other European cities.
    But when the Tesla AI can be confused by the above, I'm not so sure it will easily pass a drivers test. Something that is completely obvious to a human, obviously is not to the AI of a Tesla Robo Taxi, that even has extra sensors compared to a normal Tesla.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 1 point 1 month ago (3 children)

    So there's a long way for FSD to match a competent driver.

    "Competent driver" is a higher bar than is strictly necessary to make the roads safer. That was my point. It's easy to read a story about a self driving car getting confused at some traffic pattern, but that happens all the time for humans, too.

    So Tesla is definitely not anywhere near "better than a human".

    I don't know how to quantify "not anywhere near" but I agree the industry isn't there yet, and there are some pretty significant challenges to figure out, like inclement weather.

    Tesla isn't the only company trying to crack this; don't let Musk being a terrible person blind you to the fact that very smart people are working on this.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 2 points 1 month ago (2 children)

    don’t let Musk being a terrible person blind you

    This was about a robotaxi, this is Tesla specifically.
    There are no other vendors that are equally irresponsible.
    I'm not arguing whether Tesla is ahead or behind, I'm only arguing that Tesla is acting with a level of irresponsibility that should have had them fined and controlled on these issues long ago.

    I have no doubt that eventually the technology will be good enough, but it's 10 years ago that Musk claimed FSD would be ready next year, and be safer than human drivers, and be able to pick you up and drive you across country, completely without human intervention.

    Musk is a liar, I never made that claim about the other companies working on it. But the "good people" at Tesla are still acting irresponsibly, and putting people's lives in danger for profit.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 1 month ago (1 child)

    You'll get no disagreement from me that Musk/Tesla is being irresponsible by over stating the capabilities of their vehicles.

    But I don't find it interesting to discus Musk and his many character flaws; I was always speaking more generally about the technology itself.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 2 points 1 month ago (1 child)

    I was a driver for Waymo for several years, working overnights. When I first started, you'd be lucky if you could go more than 2 or 3 minutes without having to comment/document something, and normally have a few dozen disengages in a shift, so it kept you on your toes. By the time I left about 4 years later, I was really glad that I'd moved up and wasn't driving every night because I honestly didn't know if I could do it. Drivers would tell me they'd have nights with only one or two disengages the entire night. That job 1000% got harder as the software got better.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 3 points 1 month ago

    Exactly it's way harder to stay vigilant when nothing happens for hours. When that becomes days and weeks, it becomes near impossible.
    I have a relatively "stupid" car by today's standards. It can only keep lanes, and adjust speed according to traffic and speed limits. Nothing more than a fancy cruise control really.
    And even that makes mistakes, but it's simple enough that it doesn't create "situations" because you still have to actually drive yourself.

    It's kind of funny when approaching a round about, it slows down automatically, and I enter the round about at a comfortable speed, but then suddenly half way through, it accelerates back to highway speed if you don't prevent it. This is extremely weird IMO, since you are still in the middle of turning to follow the round about?!
    It took me a bit by surprise the first time that it suddenly accelerated in a situation where it absolutely shouldn't. But not much. because I was absolutely paying attention, as it can't navigate by itself.

  • permalink
  • fedilink
  • source
  • parent
  • [–] -1 points 1 month ago (1 child)

    Yet a normal road designer might put cross hatches or something in there to indicate do not go.

    This is both my support of and criticism of self-driving cars.

    • This is easily the road designers fault and may also result in bad human decisions
    • On the other hand every drive seems to be a series of edge cases that self-driving isn’t ready for
  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 4 points 1 month ago

    This is easily the road designers fault and may also result in bad human decisions

    Did you actually take a look at the picture further down in the article?
    IDK what the purpose of that arrangement is, but it is extremely clear that you are not supposed to cross that line.

  • permalink
  • fedilink
  • source
  • parent