Hacker News Reader

Hacker News

Comment by palmotea | Hacker News Reader

Parent Full thread

palmotea

> It's an absurd judgement.

> Consider also how many lives have been saved by the autopilot.

> Be careful what you wish for.

How many? Tell me.

View on HN

It saved my life. I was standing on the Golden gate bridge looking over the edge. A Tesla model 3 pulled over and started playing baby I need your lovin at full volume. I cried and climbed into the car and it turned on the seat heater.

So the opposite of that Herbie scene https://youtu.be/QpgrKDBHlgo?si=hA8iDnIZSVvvs_d2&t=77

There's legit dashcam video showing Autopilot preventing severe crashes. Go on YouTube instead of balking. Here's a few to get your algorithm working:

https://www.youtube.com/watch?v=A3K410O_9Nc

https://www.youtube.com/watch?v=Qy6SplEn4hQ

https://www.youtube.com/watch?v=GcfgIltPyOA

https://www.youtube.com/watch?v=Tu2N8f3nEYc

> There's legit dashcam video showing Autopilot preventing severe crashes. Go on YouTube instead of balking. Here's a few to get your algorithm working:

So? I just watched those. They don't prove anything about "lives [that] have been saved by the autopilot." They all look like scenarios a human driver could handle (and *I, personally have handled situations similar to some of those). If autopilot is saving lives, you have to show, statistically, it's better than human drivers in comparable conditions.

Also the last one appears to be of a Tesla fanboy who had just left a Tesla shareholder meeting, and seems pretty biased. I'd say his Cybertruck actually reacted pretty late to the danger. It was pretty obvious from the dashcam that something was wrong several seconds before the car reacted to it at the last second.

The last link is literally a man stating he could not handle the situation without FSD driving him. You're experiencing cognitive dissonance with the evidence.

> The last link is literally a man stating he could not handle the situation without FSD driving him. You're experiencing cognitive dissonance with the evidence.

And I have doubts about that man's reliability.

I can't speak for Tesla's FSD specifically, but Waymo did a study on the collision rate of their autonomous cars compared to human drivers: https://waymo.com/safety/impact/. They found that Waymos get into about 81% fewer crashes per mile. Compared to a statistical human driver, Waymo prevented around 411 collisions that would have resulted in any injury, and 27 collisions that would have resulted in serious injury or death. It seems like for Waymo specifically, self-driving cars are demonstrably safer than human drivers. Not sure if that generalizes to Tesla FSD, though.

> I can't speak for Tesla's FSD specifically, but Waymo did a study on the collision rate of their autonomous cars compared to human drivers: https://waymo.com/safety/impact/. They found that Waymos get into about 81% fewer crashes per mile.

I think that's true. Though I recall that Waymo limits their cars to safer and more easily handled conditions, which is totally the right thing to do, but it probably means that statistic needs an asterisk.

> Not sure if that generalizes to Tesla FSD, though.

I don't think it does. They're two totally different systems.

Statistics aren't collected on that. But I've read anecdotes where individuals recounted the autopilot saving them from a severe accident.

You can also google "how many lives has tesla autopilot saved?" and the results suggest that the autopilot is safer than human pilots.

That doesn't make any sense. If I, a human, hit the brakes in time and avoid an accident today, then hit someone tomorrow, should I not be held responsible for the second incident?

The point is to compare the consequences of more deaths by not using autopilot with deaths by the autopilot.

If you accidentally kill someone with your car, do you think you should have to pay $243m?

> If you accidentally kill someone with your car, do you think you should have to pay $243m?

If you have billions of dollars, and somehow can't go to prison, yes you should. If not in compensation to the victim, then in some kind of fine scaled to wealth. The amount paid needs to be high enough to be a deterrent to the individual who did wrong.

If this extreme judgement causes Tesla to abandon autodrive, there will be more deaths on the road. It just isn't rational.

Besides, nobody makes you turn on the autodrive.

You haven't established that it has saved any lives beyond vague, hand waving anecdotes.

What is autodrive? Are you talking about basic autopilot, enhanced autopilot, or full self-driving? They are separate modes:

https://en.wikipedia.org/wiki/Tesla_Autopilot#Driving_featur...

Which revision of the hardware and software is the "good one"? Remember that Tesla claimed in 2016 that all Teslas in production "have the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver". But that was, of course, a lie:

https://web.archive.org/web/20240730071548/https://tesla.com...

https://electrek.co/2025/10/22/tesla-changes-all-cars-have-s...

What Tesla used to claim was "full autonomy" is now called "Full Self-Driving (supervised)", whatever that's supposed to mean. How many times has "Full Self-Driving (supervised)" gone dangerously wrong but was stopped? How many times was supervision not enough:

https://electrek.co/2025/05/23/tesla-full-self-driving-veers...

Show me some concrete numbers to back your claims. If you can't do that then I think you've fallen victim to Tesla's dishonest marketing.

googling "how many lives has tesla autopilot saved?" produces:

It is impossible to determine the exact number of lives saved by Tesla Autopilot, as "avoided accidents" are difficult to quantify, but Tesla reports that its vehicles with Autopilot engaged are significantly safer than the U.S. average, with one crash recorded for every 6.69 million miles driven in Q2 2025. Conversely, independent trackers have identified over 65 deaths involving Autopilot or FSD as of October 2025.

Key Safety Data and Context:

Safety Claim: Tesla's Q3 2025 Safety Report indicates that Autopilot technology results in an accident frequency nine times lower than the US national average.

Fatalities: Data compiled by TeslaDeaths.com reported 65 fatalities linked to Autopilot through October 2025, with federal investigators having verified dozens of these cases.

Comparison: As of Q2 2025, Tesla recorded one crash for every 6.69 million miles with Autopilot, compared to 1.26 million miles without it, and a U.S. national average of one crash every 702,000 miles.

> It is impossible to determine the exact number of lives saved by Tesla Autopilot, as "avoided accidents" are difficult to quantify

Then you agree your claims have no merit. Don't be suckered by advertising.

I mentioned that upthread.

But there's strong related evidence: "Tesla recorded one crash for every 6.69 million miles with Autopilot, compared to 1.26 million miles without it"

Dude. That's not evidence, that's advertising. You're taking at face value a claim made by a company that has lied continuously for over a decade about the self-driving capabilities of their cars:

https://motherfrunker.ca/fsd/

https://en.wikipedia.org/wiki/List_of_predictions_for_autono...

Remember how there were supposed to be 1 million robotaxis on the road by 2020:

https://www.thedrive.com/news/38129/elon-musk-promised-1-mil...

Lying is part of the company culture at Tesla. They even lie about dumb things there's simply no need to lie about, but they go ahead and lie anyway:

https://www.motortrend.com/reviews/tesla-cybertruck-beast-vs...

https://www.youtube.com/watch?v=K0AJmLvKjxw

https://www.youtube.com/watch?v=5J3H8--CQRE

It's not rational to swallow Tesla's lies hook, line, and sinker. If you can't produce independent verification of Tesla's claims then there's nothing to talk about.

The judgement is only so high because of punitive damages for misleading marketing. Their actual liability for the collision itself is relatively low (and indeed the jury found them only 33% at fault).

> If you accidentally kill someone with your car, do you think you should have to pay $243m?

It would be a large portion of my net worth for sure. Maybe also prison time. Can we put Autopilot or Tesla in a prison?

Reply on news.ycombinator.com