FRESH

Hacker News

Home

Tesla’s autonomous vehicles are crashing at a rate much higher tha human drivers

477 points by breve

by z7

17 subcomments

The comparison isn't really like-for-like. NHTSA SGO AV reports can include very minor, low-speed contact events that would often never show up as police-reported crashes for human drivers, meaning the Tesla crash count may be drawing from a broader category than the human baseline it's being compared to.
There's also a denominator problem. The mileage figure appears to be cumulative miles "as of November," while the crashes are drawn from a specific July-November window in Austin. It's not clear that those miles line up with the same geography and time period.
The sample size is tiny (nine crashes), uncertainty is huge, and the analysis doesn't distinguish between at-fault and not-at-fault incidents, or between preventable and non-preventable ones.
Also, the comparison to Waymo is stated without harmonizing crash definitions and reporting practices.

by SilverBirch

7 subcomments

To be honest I think the true story here is:
> the fleet has traveled approximately 500,000 miles
Let's say they average 10mph, and say they operate 10 hours a day, that's 5,000 car-days of travel, or to put it another way about 30 cars over 6 months.
That's tiny! That's a robotaxi company that is literally smaller than a lot of taxi companies.
One crash in this context is going to just completely blow out their statistics. So it's kind of dumb to even talk about the statistics today. The real take away is that the Robotaxis don't really exist, they're in an experimental phase and we're not going to get real statistics until they're doing 1,000x that mileage, and that won't happen until they've built something that actually works and that may never happen.

by sammyjoe72

3 subcomments

Elon promised self driving cars in 12 months back in 2017? He’s also promising Optimus robots doing surgery on humans in 3 years? Extrapolating…………… Optimus is going to kill some humans and it will all be worth it!

by culi

0 subcomment

This comes after a recent iSeeCars study that found that Tesla as a brand had the highest fatal crash rate in the US (with Kia being a very close second)
https://www.iseecars.com/most-dangerous-cars-study#:~:text=T...

by ryandvm

3 subcomments

Tesla has completely fumbled a spectacular lead in EVs and managed to snatch defeat from the jaws of victory. And instead of turning it around, we're supposed to believe they are going to completely pivot and then take over a market with far more developed competitors (e.g. Boston Dynamics).
That Elon is riding this wave amidst the transparency of the whole thing is the funniest part. It's like watching people lose money at the "three cup" game but the cups are clear.

by mejthemage

2 subcomments

Humans average one police reported accident per 500,000 miles?!
TIL I'm incredibly unlucky.

by _ph_

2 subcomments

As far as I understand, those Robotaxis are only available within Austin so far. That is slow city traffic, the number of miles per ride is very small. However the number for human drivers seem to take all kind of roads into respect. Of course, highways are the roads where you drive most of the distance at the least risk for an accident. Has this been taken into account for the evaluation?
It would be ironic that people are claiming the Tesla numbers for Autopilot are to optimistic, as it is used on highways only and at the same time don't notice that city-only numbers for the FSD would be pessimistic statistics-wise.

by giorgioz

3 subcomments

Yes on one side Tesla is not transparent but on the other side the author of the article is an hypocrite given they went with the click-bait title "Tesla’s own Robotaxi data confirms crash rate 3x worse than humans even with monitor" Tesla secrecy is likely due to avoid journalists taking any chance they can to sell more news by writing an autonomous vehicles horror story. Given the secrecy we don't know what happened, yet the journalist did choose to go with the worse scenario title.

by artembugara

0 subcomment

By the law of large numbers, it's not a significant distance.

by 4fterd4rk

0 subcomment

The human accident count per mile is brought down by a lot of highway miles. The Robotaxi is, at present, geofenced. It's not going to be getting a lot of highway miles. Most crashes happen on city streets.

by stevage

1 subcomments

Hard to believe that in 2017, I was utterly convinced that self-driving cars would be the majority of all cars on the road within 5 years.

by rich_sasha

2 subcomments

As much as I'd love to pile in on Tesla, it's unclear to me the severity of the incidents (I know they are listed) and if human drivers would report such things.
"Rear collision while backing" could mean they tapped a bollard. Doesn't sound like a crash. A human driver might never even report this. What does "Incident at 18 mph" even mean?
By my own subjective count, only three descriptions sound unambiguously bad, and only one mentions a "minor injury".
I'm not saying it's great, and I can imagine Tesla being selective in publishing, but based on this I wouldn't say it seems dire.
For example, roundabouts in cities (in Europe anyway) tend to increase the number of crashes, but they are overall of lower severity, leading to an overall improvement of safety. Judging by TFA alone I can't tell this isn't the case here. I can imagine a robotaxi having a different distribution of frequency and severity of accidents than a human driver.

by fabian2k

3 subcomments

As long as there are still safety drivers, the data doesn't really tell you if the AI is any good. Unless you had reliable data about the number of interventions by the driver, which I assume Tesla doesn't provide.
Still damning that the data is so bad even then. Good data wouldn't tell us anything, the bad data likely means the AI is bad unless they were spectacularly unlucky. But since Tesla redacts all information, I'm not inclined to give them any benefit of the doubt here.

by OfflineSergio

1 subcomments

this is not good but the point is this can be improved much easier than improving human accidents rate. Both are very difficult problems, but one is certainly harder.

by oleumbudget

0 subcomment

Wow thats interesting to know and is quite concerning

by spprashant

1 subcomments

Most of us are very well aware of Tesla's shortcomings with FSD and inflated valuations.
But electrek's reporting is biased and in bad faith when it comes to Tesla/Musk.

0 subcomment

by 93po

0 subcomment

I really wish we'd ban electrek articles on HN, they're constantly false, misleading, or just grinding an axe.

by viraptor

2 subcomments

> showing cumulative robotaxi miles, the fleet has traveled approximately 500,000 miles as of November 2025.
Comparing stats from this many miles to just over 1 trillion miles driven collectively in the US in a similar time period is a bad idea. Any noise in Tesla's data will change the ratio a lot. You can already see it from the monthly numbers varying between 1 and 4.
This is a bad comparison with not enough data. Like my household average for the number of teeth per person is ~25% higher than world average! (Includes one baby)
Edit: feel free to actually respond to the claim rather than downvote

by lvl155

3 subcomments

I am so tired of people defending Tesla. I’ve wrote off Tesla long time ago but what gets me are the people defending their tech. We all can go see the products and experience them.
The tech needs to be at least 100x more error free vs humans. It cannot be on par with human error rate.

by ModernMech

1 subcomments

Of course, it could be no other way for a company that unleashed "FSD Beta" onto the streets and allowed all of us to be subjected to their bloody (literally) beta test. You don't get a safer future with "move fast and break things" mentality. Especially when the CEO is as illiterate as Musk about his own technology that he discount the results of actual experts in the field.
I mean, just look at the trail of headless corpses (there actually are multiple) left by Tesla during this beta test. Weren't we all here to witness a previous version of the thing running straight through a cartoon wall? Of course this thing was always going to end in disappointment -- it's sucked its whole existence. It's never been serious it's always been an 80/20 play hoping to get away with the con without delivering the rest of the 20% that makes it work.
Tesla's technology is bunk, their entire FSD thesis of "vision only" has been a dismal failure, and it's actually going to tank the entire Tesla car brand. I've been saying this for a while and it looks like it's finally starting to happen: Tesla is going to exit the car business never having delivered FSD in any viable capacity (although they'll claim total success), and Musk will retarget his empire to running the same FSD grift but with robots. Musk learned the bigger the promise, the more runway people give you to make it a reality. Spin a big enough yarn and Musk can live the rest of his life delivering nothing -- not Mars, not FSD, not AI, nada -- and people will still call him a genius.

by moorebob

0 subcomment

[dead]

by fatheranton

0 subcomment

[dead]

by Markaplan

0 subcomment

[dead]

by onetokeoverthe

0 subcomment

[dead]

by ProfessorZoom

0 subcomment

[flagged]

by smoovb

1 subcomments

electrek.co recent Tesla headline summary with sentiment:

  (negative) Tesla to stop selling Full Self-Driving package, moves to subscription-only

  (negative) Elon Musk says Tesla 'almost done' with AI5 design, 6 months after saying it was 'finished'

  (negative) Tesla's full 2025 data from Europe is in, and it is a total bloodbath

  (neutral)  Tesla updates 2026 Model Y with new features, launches tiny third row in the US

  (positive) Tesla launches US-made solar panel, a rare sign of life for its solar business

  (negative) Elon Musk moves goalpost again: admits Tesla needs 10 billion miles for 'safe unsupervised' FSD

  (negative) Are Tesla Gigafactory Berlin's days numbered?

  (negative) Elon Musk shows total ignorance of Tesla's current falling sales trajectory

  (negative) Tesla rolls out 0% financing to boost declining sales

  (negative) Tesla (TSLA) releases Q4 delivery results: confirms decline in sales is accelerating

  (negative) Tesla Cybercabs spotted testing, unsurprisingly with steering wheels

  (negative) Elon Musk's top 5 Tesla predictions for 2025 that didn't happen

  (negative) Tesla (TSLA) does something unusual ahead of Q4 delivery results

  (negative) Elon Musk drops 'sustainable' from Tesla's mission

  (negative) Tesla's Robotaxi project in Austin is much smaller than Musk claims

  (neutral)  Tesla Robotaxi spotted without a safety driver in Austin; Musk confirms testing begins

  (negative) Tesla US sales drop to under 40,000 units following tax credit expiration

  (neutral)  Tesla CEO Elon Musk claims driverless Robotaxis coming to Austin in 3 weeks

  (positive) Tesla announces 2025 holiday update with a few cool features

  (negative) Tesla (TSLA) sales keep crashing in Europe with a single market temporarily saving it

by jsight

2 subcomments

TBH, the comments here amaze me. The claim is that a human being paid to monitor a driver assistance feature is 3x more likely to crash than a human alone.
That needs extraordinary evidence. Instead the evidence is misleading guesses.

by timzaman

0 subcomment

Author (fred lambert) is a tesla shorter, and a total asshole. Didnt see that disclaimer on his website.

by heisenbit

0 subcomment

Elon know FSD still takes time and that is the reason he is now ramping up the robot production. Who else to turn to to steer is upcoming fleet of taxies?

by mikkupikku

4 subcomments

All these self driving and "drivers assistance" features like lane keeping exist to satisfy consumer demand for a way to multitask when driving. Tesla's is particularly cancerous, but all of them should be banned. I don't care how good you think your lane keeping in whatever car you have is, you won't need it if you keep your hands on the wheel, eyes on the road, and don't drive when drowsy. Turn it off and stop trying to delegate your responsibility for what your two ton speeding death machine does!