There's no clear evidence Autopilot saves lives | Hacker News Reader

There's no clear evidence Autopilot saves lives

arstechnica.com

67 pointsJudgmentality68 comments

Something that works really well but is pretty uncommon:

My 2016 Honda Accord coupe has a wide-angle camera on right mirror whose feed displays in a center display on when I turn the right blinker on (or manually turn the camera on).

Rather than actively do anything (like stear for me) or abstract anything (like warn me when there is a vehicle in my blindspot), it just gives me a plain view of the road to the right of me extending back hundreds of feet, with only the small augmentation of 2 added virtual lines on the road, one at a single car length back and the other at two car lengths back, to help me guage distance.

Simple features like this, which increases situational awareness without abstracting important dicisions away from the driver, are terrific and, despite only having anecdotal evidence of their effectiveness, I would bet my last dollar that improvements like this reduce accidents drastically. I don’t cut anybody off anymore ever, not that I ever did it intentionally in the past, but I am always extremely confident now in knowing whether or not I’m going to be pulling into a lane that has a speeding car approaching from behind or whether a car is in my blind spot, or somebody is about to change into the lane next to me from two lanes over, etc.

Unfortunately, the market seems to be rewarding features that attempt to replace, rather than augment, situational awareness.

> Unfortunately, the market seems to be rewarding features that attempt to replace, rather than augment, situational awareness.

So true. Shortly after back-up cameras came out, there was a terribly sad story in my hometown about an older driver leaving a church service that had gotten used to just looking at the camera instead of physically turning their head before backing up and wound up hitting and killing a child that was behind the car but not in clear view of the camera.

I guess that's good, but I've always found properly positioned mirrors, and turning my head to give me a good ability to assess if there is anything that I need to be aware of when changing lanes. It's a little unclear what value add the camera gives you. Blind spots can be very small if the mirrors are well adjusted, and then turning your head allows you to fill in the blind spot.

You answer your own question here, but since you might not realize you did... You said:

"Blind spots can be very small if the mirrors are well adjusted, and then turning your head allows you to fill in the blind spot."

Nevermind what a driver should do, features like Honda's system can save the lives of people who do not use best-practice driving techniques. This describes most people.

Honda LaneWatch (blind spot camera) is easily my favorite car safety feature.

Unfortunately, it's more expensive to manufacture, compared to standard blind spot warning systems, and the average customer does not appreciate it (by sales). Instead, Honda has received many angry complaints that the radio/media console is unavailable when changing lanes or when the user has left the blinker on.

I really like that feature on my wife's Accord.

Watch the last two minutes of this test from Edmunds and tell me why Tesla is even allowed to sell this junk.

https://m.youtube.com/watch?v=RxeK0F-D3gg

(If you're too busy to watch: it repeatedly drives the car across the double yellow line dividing a 2-lane freeway, towards oncoming traffic, once right into the path of an oncoming semi).

Geez! That was scary.

I've heard that lane assist and blind spot sensor features could make things worse because people will start to pay less attention to the road.

This video is pretty frightening and seems like it could be closely related to the most recent Tesla crash into the exit barrier.

The biggest problem is Tesla marketing this shit as "auto-pilot" where as Honda and others tell you what it is: lane assistance. It's obviously far from perfect and I really don't think anyone should rely on it at all.

The article separates out auto-breaking and adaptive cruise control and I'm glad it does. Those are present on a lot of other vehicles and it makes sense that those have prevented accidents. Side/blind-spot cameras connected to turn signals are also incredibly helpful.

Going back to actual self-driving tech, I think this helps prove it's a long way away. I'd say at least 10+ more years of prototypes and I'm honestly sick of tax credits going into it. States should work on fixing our god damn public transport infrastructure first. It will cost less, transport orders of magnitudes of more people and self driving trains are real tech that exists now, verses some sci-fi fiction that may or may not be viable in a decade.

I don't think this proves anything about the state of self-driving tech generally. It just proves that Tesla doesn't have any, and that camera vision alone can't do much. Waymo's car doesn't drive like this.

As a very rational person (no doubt like many of you) I am cautious to explore my 'gut' feelings with a critical eye. Often 'gut' feelings can be misleading or at least alleviated by further info.

In autopilots case, and more generally with autonomous driving, I just can't shake the gut feeling that we're much much further away from achieving true autonomous safety than many people are saying.

I think I might be predisposed to ultra-conservatism (small c) with automobile safety, so that might be playing it's part I guess.

Good lord I've seen drunk drivers manage better. Is something completely wrong with their cameras engineering wise?

I had similar experience when I was test driving Tesla, Audi, and BMW recently. Tesla performed the worst of the 3.

If Tesla's autopilot is clearly 3 times better, put out the data. I know me and many other data scientists would be happy to analyze the data and see how Tesla autopilot is doing. Surely if Tesla's data proves out their point, why not release it?

In general, I think autopilot is a controversy magnet. From its name to its seemingly speedy initial rollout, to the highly publicized Harry Potter death, to the recent crash. Not to mention, anything to do with Musk and his futuristic ideas gets huge play in the press, though the actual reality is the features are quite slowly rolling out and are fairly underwhelming given the hype.

Also, the hype about this being self-driving: this is basically only a highway assist system, there are numerous challenges around the ~30 MPH, suburban/urban driving challenge to overcome. At that speed range in particular, I think lidar could be useful to Tesla, since it'd provide even further data in a variety of conditions, thus redundancy and more potential for completely safe driving in this most tricky and unsafe speed range where one must interact with pedestrians, bikes, and various things entering the roadway.

>to the highly publicized Harry Potter death

I must be living under a rock. What happened? Did Dumbledore die in a Tesla?

> [2016] Tesla owner killed in crash was watching 'Harry Potter' while using car's autopilot

> "They should be writing a story about how autonomous cars are really safe," Musk said in his Wednesday earnings call. "But that's not a story that people want to click on. They write inflammatory headlines that are fundamentally misleading to readers. It's really outrageous."

I'm so very tired of the attack-the-messenger (i.e., attack-the-media) strategy. Musk provides no support for his claim that autopilots are safer, but he does try to change the subject from autopilots to the reporters. What a cheap strategy, even if trendy, for such a smart person.

IME, usually the reporters turn out to have raised important questions, and perhaps the more the subjects of the report complain, the closer to home the news report has struck. (I'm excepting editorials, columns, blogs, etc. - I mean real journalism in serious publications.)

I don't understand why Tesla bet on Autopilot. They're a small car company that's just getting started, why not concentrate on selling electric cars? That in itself is unique enough, the Autopilot feature is mostly a gimmick which adds complexity and liability. Even if they think self driving cars are the future they could wait until it can be bought from suppliers, or at least develop it in house without releasing it to the public until it's done.

To a lesser extent I think the same about Falcon wing doors on the model X, although that isn't a safety issue and they seem to work ok now. But the car would have been done sooner and cheaper if they had released it with regular doors, and I doubt the customers would have minded.

I agree that the doors are idiotic, but they serve to make normal people who don’t care about cars want to spend >$100k on a vehicle. “It opens like a lambo!” Silicon Valley had a bit about this too: https://youtu.be/_ih1ptOguaM (warning: they swear during this)

The “autopilot” is obviously oversold and mislabeled, but again, it’s a way to get non-car people with too much disposable income to want one. “Can your car do that???!”

Free charging some places, HOV stickers, tax credits on >$100k vehicles (the Porsche 918 was also eligible)...and smug owners believing that they’re really helping to save the planet vs riding a bicycle / walking / using public transit / carpooling / not living an hour away from work.

> the car would have been done sooner and cheaper if they had released it with regular doors, and I doubt the customers would have minded

IME, and speaking very generally, when businesses take the approach that 'customers won't mind', their products revert to the mean, to commodities.

If you want to have a special brand, to charge a premium, you need to thrill customers, not do things in a way that the customers 'won't mind'. Consider Apple and their 'obsessive' focus on small details. Consider the Beatles - they didn't say, 'lets just do it the easy way; our listeners won't mind'.

Also, I think customers do notice these things, even if unconsciously. It applies to any craftsmanship. Most people can distinguish a cheap men's suit from an expensive one, even if they can't consciously name many reasons. The cheap suit is made to the standard of customers 'not minding'; it's a commodity and nobody cares about the brand.

I get that special features can be important to stand out from the crowd, but my point is that Tesla is the only available well designed electric car at the moment, have been for years, and will be for a while longer. That's why people buy them, not because of new kinds of doors or overhyped Autopilot. You could call the interior screen a bit of a gimmick too, but I think it's mainly a way to save on interior design cost, and then made out to be a futuristic feature.

I think it's more of an ego thing, they want to show that they're more high tech than the old "dinosaurs" that still use fossil fuel and requires you to drive your own car. Focusing on the car part and less on gizmos in the start might have been better.

I think that Tesla (Musk) are putting a lot of marketing into Autopilot. Other automakers have comparable technologies but they are not marketing them as aggressively and definitely not position them as an autonomous driving platform as Tesla does.

If we just look at the set of technologies (ACC, line assist, emergency braking, etc.), it is actually quite impressive how far we jumped in only a few years. However, it is also clear that true autonomous driving is definitely not available in consumer cars today (I don't count Waymo and a couple other companies as consumer car company). We are closer to it today than we were even 5 years ago. But not there yet.

Disclaimer: I mentioned it on HN before that I had an overnight test-drive of Tesla X and I didn't like the autopilot at all (among other things). I do like the set of similar technologies (adaptive cruise control, keep lane assist, etc.) in my Q7.

They should renamed it to "Super Cruise"

That's what Cadillac calls it

http://www.cadillac.com/world-of-cadillac/innovation/super-c...

Huh. I still don't like that it doesn't require hands on the wheel (I fell like a creative person could trick the eye tracking). But unlike the Tesla, it does have LIDAR.

> a creative person could trick the eye tracking

People have already started to bypass hand requirements by placing a weight on the steering wheel . At a certain point, you can't really blame the manufacturer when the driver is actively trying to defeat safety features put in place for their own good.

Sure you can, because those features aren't just there for the driver's own good: they protect all the other people on the road.

Don't forget, a big part is absolving the car manufacturer of liability.

Well, we're going to need something as the Boomers continue to age, eventually aging out of being able to drive themselves en masse but being... let's say disinclined to give up their car keys and rely on someone else to drive them around.

Meeting them halfway with self-driving cars is going to save a lot of lives, compared to having octogenarians with horrible reaction times and the idea that having their own car makes them an adult filling the roads.

This is a non sequitur. There's no contradiction between "There will eventually be autonomous programs that are safer than human drivers" and "Tesla's current autonomous program is not safer than human drivers". There's no principle requiring Tesla's current program to be good enough, or even that it will ever be good enough– the future might come from someone else.

I never said any of that. You're responding to a post I never made.

We could do what every other country does and build a really good public transport infrastructure so they can take trains and buses everywhere. That will also greatly reduce drunk driving and help people who are very poor whose lives can literally fall apart if something on their car breaks and they can't afford to fix it right away (and then can't get to work, and then lose their job -- or they fix it and forego insurance, and then get in a wreck and then own more money and legal fees).

Cars should be an optional luxury, unless you intentionally chose to live in a suburban area.

You're ignoring the massive rural areas of the country, which are neither urban nor suburban and will never be served by a mass-transit system.

Actually in the US, rural areas are served pretty well by AmTrak as a principal way of getting to big city airports (although more expensive than a drive, AmTrak is way cheaper than using a rural airport to connect to a flight).

But if you live in rural areas, you're choosing to require a vehicle. That's fine. But in cities, are transport is shit in the US.

Even with rural areas, Northern Indiana has the same population density of Scottland. It could easily utilize a decent intercity rail system. The US use to have more passenger track at one time than Western Europe has today!

We had rail once. It went away and we spread out, but if we rebuild it, people will start living closer to it and everyone can benefit.

I think once the US hits a really hard economic collapse (not that piddly one in 2008 the banks just used to buy other banks) is the only time when US politicians will be forced to fix infrastructure (because the poor literally won't be able to go to work, and politicians really only care about the poor once they're no longer generating taxable revenue they can feed to their supporters). Once that happens you will see things change very fast, but things do have to get considerably worse before politicians invest in people.

> Actually in the US, rural areas are served pretty well by AmTrak as a principal way of getting to big city airports

This alone tells me you've never lived in a rural area in the US or relied on Amtrak.

> But if you live in rural areas, you're choosing to require a vehicle. That's fine.

It proves my point, to the extent that people who live in rural areas (by choice or not) need individual vehicles, and that given that the number of people who are too old to drive will be increasing in the medium term, we'll need something to provide people who are too old to drive with individual vehicles.

> Even with rural areas, Northern Indiana has the same population density of Scottland.

I'm sure it does. Now shift your gaze westwards, to North Dakota, Wyoming, eastern Montana, and, really, most of the rest of what used to be called the Great American Desert and is now referred to as the high plains or, more simply, flyover country.

You're doing something I see software developers do for their programs: Take a hard problem, solve the easy part of it, and consider their job done because they don't/won't/flat-out refuse to see the really hard bits their "solution" leaves unsolved. It's like their vision... slides over the hard parts, their minds automatically move away, as if you were trying to make two bar magnets touch at the wrong ends.

> This alone tells me you've never lived in a rural area in the US or relied on Amtrak.

ad hominem

You know absolutely nothing about me. I have lived in rural areas. I've never relied on Amtrak, but I've used it quite a bit. Most of what I was saying comes from facts out of people like Wenover Productions:

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

This video doesn't go into the rural -> plane connection; can't find that one at the moment. But he has another video on it. But yes, people do use Amtrak for connecting to major airports from rural areas.

> I'm sure it does. Now shift your gaze westwards, to North Dakota, Wyoming, eastern Montana ....

I was not addressing any of those areas. At all. You built some argument I wasn't making. I never said trains would be good for those regions. I think if we did build good rails, those areas would get some service eventually.

Australia has a similar situation (a region I have lived in btw). They only over 30M ~ 40M people. I have taken trains intercity and they are very touristy and not meant for general transport. But each capital city: Adelaide, Perth, Melbourne, Sydney ... they have train systems that would make any US capital be ashamed of itself. The only capitals without any rail are Hobart, ACT and Darwin.

Australia has the infrastructure to accommodate high-speed intercity if they ever got around to implementing it. The United States does not and it has a much higher population density and much larger cities.

American cities, CITIES (which I've been saying a lot) need real transportation. If we built that, the rest would start to naturally follow into place. We'd see rural rail as a byproduct. We use to have it in the US and it's gone now thanks to GM/Ford/big auto buying rail and bus lines and then killing them.

I'm looking back over my comment just to make sure I'm talking about cities. I really feel like your anti-rail stance goes to the hart of the problem. American hate rail and public transport for some fucked up nonsensical reason and it doesn't make any sense to me at all.

I lived without a car for five years in three different countries and I simply think American cities shouldn't require cars to be livable. I don't think that's unreasonable.

I'm talking about the whole country, not just the cities. Why are you only talking about the cities?

What I'm really talking about is all old Americans, all across the country. I'm not focusing on any subset.

> American hate rail and public transport for some fucked up nonsensical reason

Nonsense. This is simply wrong, as talking to an American would tell you.

This argument seems so pedantic to me.

Human driving has basically capped out at its potential to improve. It may even be starting to get worse...

Automated driving, has barely begun and even if it is less safe today, it will clearly be much safer than human driving soon. And will only continue to get safer and safer as the technology improves.

Do you we need to be so micro-focused as to have an exact day/model/release where autopilot is safer than human drivers?

I for one have been hit by a drunk driver. My wife was in an accident where the other driver was ACTIVELY playing candy crush. We can all see that automated driving will make the roads a safer, better place even if that takes 20 years.

The interesting question, which this article touches on is the following. Given that we don't have full autonomous driving--and won't depending on the environment for quite a while--what is the appropriate level of assistive driving systems? For example, as this article indicates, automated emergency braking seems to be a fairly obvious win. Adaptive cruise control probably is as well. But there may be issues with putting incrementally better systems into cars until they are really good enough to operate on their own.

It's hard to call it capped out when we could improve a lot by reducing drunk driving and discouraging phone use while driving. Don't you think so?

"Humans are stupid, tech will solve it" attitude is missing the point that humans need to build, maintain and take care of tech. If humans are stupid, tech built by their designs and under their control will be shit too.

Your first point is good. I should have stated that "Hypothetical Best Possible Human Driver" has been established. We can do a lot to improve safety in the average driver. But human reaction times, spacial awareness have hard limits. You can only react so fast, and you can only look at what your two eyes can show you at once. Once the average driver starts to approach the best possible human driver, improvements will be very hard to come by.

But your second point, if humans are dumb, their tech will be dumb is demonstrably false. Groups of idiots can collaborate to make 1 fantastic solution to a problem. Or a solution that improves over time. Software also tends to have fairly stable performance, except for specific outages. Where a human may have the capability to be a safe driver, but performance varies drastically based on mood, age, distraction.

I could definitely be wrong. But I would be comfortable betting every penny I have that automated driving will be safer even the best human drivers in 20 years. No one can know the future, but betting on human power instead of technology has been a losing strategy since the printing press was invented.

Humans can be pretty good if they're trained like jet fighter pilots. I'm sure automated cars will outperform them (or at least be as good as them) if they're built and maintained like jet fighters. If we can't make average human a jet fighter pilot, can we make them maintain their cars at that level? And expect to pay for them accordingly.

On the second point, I'm more talking about hacking scene and maintenance. I'm pretty sure today's jackasses will find a way to hack auto-piloted cars to act like jackasses. And drunktards will drive unmaintained cars with not working and counterfeit sensors too.

That's the point: I'm pretty sure that somebody will create automated driving technology that's better than humans within the next decade or so. But it won't necessarily be Tesla. If I had to guess I'd say Waymo is likely to get there first. And as I say in the piece, there's good reason to think that Tesla's incremental approach—start with ADAS, evolve into full self-driving—may be an evolutionary dead end because humans are really bad at supervising a driving technology that works 99.99 percent of the time but gets in a fatal crash the other 0.01 percent.

> And as I say in the piece

Are you the author of the article?

Twitter handle is the same as his hackernews handle, so I'd assume it's likely.

Yes.

"humans are really bad at supervising a driving technology that works 99.99 percent of the time but gets in a fatal crash the other 0.01 percent"

I've tried to express the same concept but you said it in far fewer words. Nice.

Good article.

What about Tesla's choice to design its technology around cameras rather than lidar? They seem to be the only ones going that route. I've seen quotes from a number of people who build this kind of technology who say that's a fundamentally less-safe way to build self-driving vehicles. Did you consider addressing that point in the article?

The argument is not pedantic.

>it will clearly be much safer than human driving soon

You cannot say really that. The technology at this current stage is much worst than human drivers when it comes to understanding its environment and detecting hazards.

So it is quite unreasonable to ask if this system is safer right now, because some future system almost certainly will be?

If Tesla had not prominently promoted the exact claim that this system is safer right now (and by a remarkable margin), that claim would not be under scrutiny. Tesla's statements are now being subjected to an objective, reasoned analysis, even though they were not made in that spirit.

But the counter-argument, to allow autopilot to catch on if it is NOT safer today is to cause deaths, even if over the long run (6 months? 6 years? 60 years?) it will be a net positive.

The stats need to compare a Tesla not to the "average" car, but the average car in that market segment driven by an owner of that socioeconomic class to tell us much of anything. And then we need to separate out when autopilot is active vs when it's merely functioning in a passive sense. Hell, my VW Golf has forward collision warning with automatic emergency braking. Even Corollas these days have laser cruise.

Further investigation required: are humans worse at driving or statistics when loss of life is a factor in each?

One of my early thoughts with some of the driver assist technologies was "oh cool, I can take off my jacket while driving, since the car can pay attention in that short time span." It was at that time that I realized that there was a problem with autonomous driving: there is a gap where it's short of allowing true inattention but is usually safe enough that it is going to train users to be even more inattentive. The natural consequence of this is that autonomous driving, at a certain point, will decrease safety, not increase it.

Waymo seems to realize that this is a problem, and is actively pursuing a development model that doesn't release self-driving cars except on the other side of the unsafe gap. Tesla seems to be trying for a constant incremental improvement model, which means they're going to be putting people into the unsafe gap... and their marketing department is going to be counterproductive.

>Human driving has basically capped out at its potential to improve. It may even be starting to get worse...

But we can significantly improve on that. Augmenting human capability is far easier than converting a pile of transistors to perform at the same level as humans. Improving road safety infrastructure is also possible - today, Improving technology in cars to detect when the driver is impaired is also easily possible - today. We don't need pie in the sky "autopilot" technology to improve on the current situation.

>Automated driving, has barely begun and even if it is less safe today, it will clearly be much safer than human driving soon. And will only continue to get safer and safer as the technology improves.

You can say that about anything. One day a bacterium will evolve into a human. Also, its more than reasonable to say, we don't want to get killed because some dude in the valley pushed a firmware update. If the tech sucks, don't endanger other humans. Go fix it, and come back and we'll gladly take a look at it again since it does have value. As of now, people are better off staying away from any Teslas on the roads.

No, automated driving won't be doing that. It will never make your car safer than human drivers. You think every human driver is out there playing Candy Crush?

And there are several car models that have ZERO fatalities. How is Tesla supposed to exceed that safety standard?

Right now Tesla is being criminal in comparing their safety rates to the average vehicle. That includes comparisons to motorcycles, which have death rates that are 50x-100x more than luxury passenger cars.

If Tesla is comparing death rates of their luxury vehicles to motorcycles, they have already failed.

And you know whats safer than cars? NOT cars. As in, mass transit.

What car models have zero fatalities?

At the bottom of this page: http://www.iihs.org/iihs/news/desktopnews/on-the-road-again-...

"Human driving has basically capped out at its potential to improve. It may even be starting to get worse..."

That's not true at all. If you've driven in other states and especially other countries, you know there is a lot of variations on average driving abilities. California drivers are among the worst and German drivers are among the best I've personally encountered. It's not that CA drivers are dipshits but if you look around enough as a pedestrian or while riding a bus passing other cars, you'll notice MANY of them are on their phones. This isn't humans reaching our cap but a recent degradation that can be corrected.

Reply on news.ycombinator.com