Berichten zufolge stürzt Tesla Robotaxis mit einer Geschwindigkeit ab, die viermal höher ist als die von Menschen

https://gizmodo.com/tesla-robotaxis-reportedly-crashing-at-a-rate-thats-4x-higher-than-humans-2000722989

40 Kommentare

  1. TwoWeaselsInDisguise on

    Ban them?!

    If they keep running start trapping and impounding them, there’s no reason for this shit.

  2. Big-Chungus-12 on

    I thought they were a lot better than Waymo? Wonder if Elon’s lying again 😂

  3. DressedSpring1 on

    I think two things can be true. I think it’s possible and even likely that robot cars will drive to a significantly higher standard of safety than humans if not right now then within the next year or two.

    And I also think a company that exists to pump their stock on lies about technical capabilities they are nowhere near realizing should have no fucking business being allowed to have their cars drive themselves on public streets.

  4. Shiftymennoknight on

    so who is gonna be held accountable when these things mow people down?

  5. Tbf this is counting minor events that would never be reported for statistics on human drivers

  6. This is based on just 5 incidents in the city of Austin in a single month. They include things that a human might not even report, like touching a parking sign. So pretty small sample size to draw extrapolations like „Tesla Robotaxis Reportedly Crashing at a Rate That’s 4x Higher Than Humans Transportation“

    Edit: does every single piece of data need to filtered through some lense? Can’t we just be objective? This is nothing. I’m not praising Tesla. Just saying this isn’t enough data to base any kind of conclusions on. 

    Read the article. Then comment on the data. 

    *Reads headline* „I KNEW it!“ *posts personal opinion* -Everyone in the comments  

  7. intensive-porpoise on

    And those are the *reported* numbers! I’m sure there is some kind of emergency GTFO protocol for a *self driving vehicle* that gets in a fender bender. Like, there is no need to exchange info. Just redline it back to HQ, get a new panel… Never happened!

  8. BeckerHollow on

    Correct my logic if wrong, but if a computer is only as good as its programming, and its programming is made by humans, wouldn’t the error rate be normal human error multiplied by the number of devices using that same program?

    One human makes a mistake driving, that’s one accident. According to the article, 1 per 229,000 miles on average. 

    One human makes a mistake programming, that’s one accident multiplied by # of devices per mile. Article states it’s apparently 4x higher.

    I don’t know 

  9. This article states that low-speed collisions with objects are also included in this small dataset, but looks like those involving human drivers aren’t recorded.  🤔

  10. Overclocked11 on

    I remain bewildered how these are allowed on the road – yes, money will pay for a lot of people to turn and look the other direction, but holy shit. Its insane that these things were just thrown on the road so carelessly (along with Tesla self-driving).

    Totally insane to me.

  11. That’s what having a drug addicted egomaniac at the helm will do for you.

    But it’s a meme stock so results don’t actually matter evidently.

  12. Alan_Reddit_M on

    Man CGP-Grey couldn’t have been more wrong about self-driving cars. Turns out they’re even STUPIDER than humans

  13. notPabst404 on

    Yet people on the sub for my city are just parroting that these are somehow safer 🤮.

  14. Not sure what is more the number of tesla crashes or the number of ads in that gizmodo article.

  15. Bluefeelings on

    Self driving is high subscription cost due to insurance and human loss cushion for the company. This should not be allowed.

  16. Humans are literally the cheapest robots we have to operate. Why can’t we just stick to human powered taxis?

  17. Robotaxis and humaniod robots are all a distraction from Musks failure.

    **Facts – Tesla.**

    Financials:

    * First-ever annual revenue decline: $94.8B (2025) vs $97.7B (2024)​
    * Net income crashed 46% to $3.8B​
    * Q4 2025 profit plunged 61% to $840M​
    * Operating income fell 38%

    Product failures:

    * Cybertruck disaster: 20,237 units sold (down 48%), 92% below 250K target
    * Model S/X discontinued: Production ending June 2026 after 12-14 years​
    * No new competitive models: Aging lineup (Model 3 from 2017, Model Y from 2020) while competitors launched 150+ new EVs
    * $25K Model 2 cancelled: Mass-market vehicle abandoned

    Market share collapse:

    * Lost global EV leadership to BYD: 1.63M vs 2.26M deliveries​
    * Second consecutive year of declining sales (down 9%)​
    * European sales collapsed 28% while market grew 27%
    * Germany sales down 72% from 2022 peak

    Brand destruction:

    * Brand value plummeted 36% in 2025 alone ($43B → $27.6B), third consecutive annual decline​
    * Musk’s political role in Trump administration/DOGE alienated customer base
    * Federal EV tax credit eliminated September 2025, crushing demand​
    * California registrations down 21% in Q2 2025 alone

    Tesla is done.

    The robotaxis/ humanoid is all an elegant narrative escape-to save Musks ego and brand.

    All Musk had to do was concentrate and not open his stupid mouth. Give it 5 years or less, and Telsa and Musk will be a casestudy, on how to screw yourself.

  18. Five total incidents that were previously not disclosed were added to the total known count, which now sits at 14 total accidents involving Tesla Robotaxis since June 2025.

    >The [five new] 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.

    I’m really interested in the crash with the bus when the Tesla was stopped.

    >Tesla also appears less transparent than its competitors when it comes to crash reporting. Unlike Waymo and Zoox, Tesla has redacted the incident narratives for each crash in the NHTSA database, citing “confidential business information.”

    >Additionally, Electrek reports that Tesla updated a crash report from July that was originally filed as “property damage only” and now lists the incident as “Minor w/ Hospitalization,” indicating someone later required hospital treatment.

    This stood out to me as a bright red flag as well.

  19. When will the market wake up to the fact Musk is a BS merchant? It’s insane.

  20. Robotaxis seem like an easy way to target people. Calling these “accidents” is an assumption.

  21. Anyone who has ever used FSD knows how it performs

    It is great for highway driving. But not lane changing

    Street driving is so sus.

  22. Tesla’s so-called self-driving should be banned. Waymo has the only system that comes close, and even it needs work.

  23. ForsakenRacism on

    No cus it disconnected 4 milliseconds before impact so it’s your fault

  24. bandswithgoats on

    It’s just swell that this asshole is allowed to test his products on our streets and bodies. What’s some dead human beings here and there?

  25. love how we went from „full self driving by end of year“ to „crashes 4x more than humans“ in like 3 years. the roadmap was always vibes

  26. Traveler_90 on

    Cause tesla relies on camera heavily and the basic sensors it has. Whereas Waymo has cameras, sensors, radar, lidar.

  27. Tesla spaceX and his AI firm will not be around in ten years from now. Musk will be living either in south africa russia or a prison

Leave A Reply