Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Apple engineer killed in Tesla crash had previously complained about autopilot

kqed.org

Post reply on HN
Print view View original

331–340 of 919 posts

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#331

Post by toast0 »

curiousgal wrote: post #145

This happened 2 years ago. It was a left-hand exit (why are those even a thing?). The seperator did not have contracting protectors because a car had previously crashed into them, I wonder why. No one ever dies in car accidents. Hence, Tesla is super bad for marketing its autopilot according to HN.

There's a left-hand exit there because there's an interchange between freeways, and the HOV lane exits on the left, to enable free flowing HOV traffic to avoid merging through general traffic to get to a right exit.

At this interchange, I've seen lots of human drivers change between the exiting HOV lane and the continuing HOV lane (or vice versa) much later than is safe; presumably because they were surprised that their lane was exiting, or they noticed the exit too late, but didn't want to miss it. I haven't seen anyone drive in between the two lanes as if it was a lane, though; and usually the late lane changers are braking, not accelerating.

It is unfortunate that the crash attenuator wasn't reset. Looking at the docs for a similar attenuator[1], it seems like resetting is somewhat involved, and you'd need a trained and properly equipped repair crew to do it, even if the time required is not that much. Scheduling is probably an issue.

[1] https://www.dmtraffic.com/assets/sci_smart_cushion_design_an... page 9-13

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#332

Post by keanzu »

kgabis wrote: post #318
keanzu wrote: post #291

Earlier quoted context omitted.

> dead Uber pedestrian Elaine Herzberg was in dark clothing crossing a dark street well away from any crosswalk or street lighting. Would a human driver have performed better? From the footage I saw she was nearly invisible, I would have hit her too. https://www.youtube.com/watch?v=ufNNuafuU7M This was not a hard fail for the AI.

The footage shows how car's cameras saw the accident. Human eyes have much greater dynamic range than that, so I wouldn't assume that a human driver would not perform better. Something appearing pitch black on this footage could be well recognisable to you. Also, this car's lidar also failed to recognise the pedestrian so if this isn't an AI fail then I don't know what is.

This is absolutely a fail, a woman died. The question is whether or not this incident is an example of "fails harder than most average drivers" or a hard fail.

With the sensor array available to it the car should have done better, no question.

But to make the claim "fails harder" I would be looking for a clear cut situation where a human would almost definitely have outperformed the AI.

Human eyes do have miraculous dynamic range so we would likely see more. Can we say with 90% certainty that a human would have saved the situation?

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#333

Post by simonebrunozzi »

fxtentacle wrote: post #187

I avoid new technology, exactly because I'm an engineer. I wonder if that is just me, but when new technology is introduced and hyped, I usually take a quick look at implementations, research, and talks just to get an idea of what the state of the art really is like. As a result, I have become the late adopter among my group of friends because the first iteration of any new technology usually just isn't worth the iss…

I would almost agree with you, but please indulge with me in the following thought.

Let's say that you can go to work by feet or by bike.

Let's also say that by feet your probability of getting killed are 1 in 1,000,000, and the commute takes 20 minutes.

On a bike, commuting takes only 10 minutes, but the probability is up to 1 in 200,000. Five times more.

You end up deciding to use the bike every day to go to work.

In this example, you decided to trade comfort (faster commute) with a slightly higher probability to end up dead.

Imagine now you need to decide whether to commute in your Tesla with or without autopilot.

Let's assume (I might be wrong) that Tesla's autopilot increases your chances of getting killed. (for simplicity, let's ignore the consequences for other people on the road).

Would you still trade comfort (not needing to drive) with a slightly higher probability to die?

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#334

Post by ascar »

stephenr wrote: post #237
allovernow wrote: post #217

Earlier quoted context omitted.

I don't know where you got these papers but you clearly got an awful sampling and I don't think you're giving computer vision a fair evaluation. Granted, we're not quite ready for self driving, but there's no question that the neural network subfield of ML has absolutely exploded in the last 5-10 years and is bursting with productionizable, application ready architectures that are already solving real world problems.

> Granted, we're not quite ready for self driving And yet it's literally in cars on the road. I'm not saying you're wrong because of that. I just wonder how far from "ready" we are, and how much of a gamble manufacturers are taking, and how much risk that presents for not just their customers, but everyone else their customers may drive near.

Based on tesla's safety report [1] it's already less dangerous than letting humans drive (alone). The error rate of human drivers tends to be downplayed, while the perceived risks of automated driving is exaggerated, distorting the picture.

Yes, it's a hard problem, yes we are not nearly there and there is a lot of development/research to do. Yes, accidents will happen during the process. But humans suck at driving and kill themselves and other people daily. It's the least safe form of transportation we have.

[1] https://www.tesla.com/VehicleSafetyReport?redirect=no

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#335

Post by asdff »

ironmagma wrote: post #180
PopeRigby wrote: post #146

Earlier quoted context omitted.

I think self driving cars are a ridiculous solution to a problem we created and doesn't need to exist. The US building out the country for cars wasn't a good way to do transportation. If we just had trains going from city to city and subways in the cities, we wouldn't even need self driving cars.

This is true, but there’s no rolling back history. We’re quite locked in to our current situation and the best we can do is work within the constraints that have been established.

Or set the clock right and invest in road diets and protected bus lanes. These are solutions that any city can implement but doesn't, because they aren't politically agreeable to local electorates. Nearly all american metros were laid out around the street car, anyway.

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#336

Post by keanzu »

michaelt wrote: post #327
keanzu wrote: post #291

Earlier quoted context omitted.

> dead Uber pedestrian Elaine Herzberg was in dark clothing crossing a dark street well away from any crosswalk or street lighting. Would a human driver have performed better? From the footage I saw she was nearly invisible, I would have hit her too. https://www.youtube.com/watch?v=ufNNuafuU7M This was not a hard fail for the AI.

LIDAR would have picked that up dead easily. Just like LIDAR would have picked up https://www.extremetech.com/extreme/297901-ntsb-autopilot-de... And just like LIDAR would have picked up https://youtu.be/-2ml6sjk_8c?t=17 And just like LIDAR would have picked up https://youtu.be/fKyUqZDYwrU And just like LIDAR would have picked up https://www.bbc.co.uk/news/technology-50713716 These accidents are 100% due to the decis…

"Brad Templeton, who provided consulting for autonomous driving competitor Waymo, noted the car was equipped with advanced sensors, including radar and LiDAR, which would not have been affected by the darkness."

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

The car had LIDAR.

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#337

Post by Piskvorrr »

keanzu wrote: post #291
fxtentacle wrote: post #249

Earlier quoted context omitted.

The dead Uber pedestrian where the AI dismissed the data as faulty and did not slow down at all. I think every human would slow down if they see things that they cannot explain. An AI will not. It's basically the same problem as when an image recognition AI is 99% sure that the photo of your cat shows guacamole. Current AIs do not have a concept of absolute confidence, they only produce an estimate relative to the ot…

> dead Uber pedestrian Elaine Herzberg was in dark clothing crossing a dark street well away from any crosswalk or street lighting. Would a human driver have performed better? From the footage I saw she was nearly invisible, I would have hit her too. https://www.youtube.com/watch?v=ufNNuafuU7M This was not a hard fail for the AI.

Nice story...if only it would match the evidence. The algorithm has detected the person 6 seconds before the crash, but as it didn't match any of the object classes conclusively, it took no action. You read that right: "there's something at 12 o'clock. Bike? Person? Unknown? Who cares, proceed through it!" If that's not a hard fail, IDK what would.

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#338

Post by mistahenry »

pferde wrote: post #298
perl4ever wrote: post #64

Earlier quoted context omitted.

"you have to always be prepared for it to do something completely unexpected on a dime" This is inherently absurd, even if you remove the last three words. You could write an essay, or a book, on why this is an example of a class of impossible problems. Or, inverting it, it's practically the generic framework of all catastrophes.

You are being downvoted because your reply is basically an unnecessarily long "nuh-uh!" - a dismissal without any counterargument. Wrapping that in a string of fancy words does not change the fact that you are not providing any argument at all.

As far as I can tell, the previous commenter was in agreement with the sentence he quoted, going so far as to say that the “on a dime” stipulation isn’t even necessary to consider the idea that we “must always be prepared for it to do something unexpected” an absurd reality

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#339

Post by keanzu »

leereeves wrote: post #321
sebcat wrote: post #314

Earlier quoted context omitted.

The exposure time and the dynamic range of the sensor affects the visibility of the person in that video - it is very likely that a non-distracted human would have performed better. The vehicle was equipped with both radar and lidar. The victim was detected as an unknown object six seconds prior to impact, and at 1.3 seconds prior to impact when the victim entered the road the system determined that emergency braking…

> the system determined that emergency braking by the driver was required, however the (distracted) driver was not alerted of this. Why would the system notify the driver that emergency braking was required instead of simply braking?

"the nascent computerized braking technology was disabled the day of the crash"

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

Re: Apple engineer killed in Tesla crash had previously complained about autopilot

#340

Post by asdff »

mrits wrote: post #167
PopeRigby wrote: post #146

Earlier quoted context omitted.

I think self driving cars are a ridiculous solution to a problem we created and doesn't need to exist. The US building out the country for cars wasn't a good way to do transportation. If we just had trains going from city to city and subways in the cities, we wouldn't even need self driving cars.

Not everyone wants to live in a dense population.

True, but why not build out transit for existing dense populations which are congested?
Post reply on HN