Hacker News Reader

Comment by salimmadjd | Hacker News Reader

This is sad and tragic.

I remember going with a friend on a Tesla test drive. He was eager to try the Autopilot system. And within a few minutes we got freaked out with what seem to us as unpredictable action and we turned it off.

I think the engineer should have never relied on it. Especially if he had concerns with the system.

At the same time, "Autopilot" is misleading and Tesla does own responsibility here. Even if it's purely a marketing term it's designed to up sell you on a half-baked technology.

They should call it "Driver assist" or something like "cruise control plus".

It's a heartbreaking story.

View on HN

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.

Call it speculation, but most people don’t really want all the consequences that come with the distinct lack of it in the US. Especially when you consider the economic effects of walkable areas.

Most people don't really want all the consequences that come with the distinct lack of rural areas either..

But if you have denser cities, that leaves more room for rural areas. I don’t think anyone is arguing for less rural land, but as it stands, it is legally impossible to build another city like San Francisco or Manhattan in the US, and that’s a problem for sustainability. People want to live in those places, and not just because they are old.

It's legally impossible to even build San Francisco into a place like Manhattan. Most the city is only zoned for two stories.

LA is zoned for a lower population today than it was half a century ago, and in that time the population has increased by 50%. Enter housing crisis.

The problem is not that you can't legally build a city like SF in the middle of nowhere. There are hundreds of smaller cities that could grow into bigger ones, but their population growth just isn't there. People wouldn't come to your "New SF" either.

The real problem is:

- Many people choose to live in big hubs like SF or Manhattan because that's where the high paying jobs are

- You can't legally turn SF into Manhattan (i.e. making it denser)

While it is true that Americans rely a bit too much on their cars, and have built infrastructure based on it instead of public transportation, it's hard to envision a world where individual transportation is no longer needed.

Look at Japan, they have a great train network but still use cars a lot. Because while train works well in cities, you still need people living in the country side to grow your food.

There a massive gap between “all cars should be banned”, which you seem to be replying to, and “some urban areas should be allowed to densify and operate without cars if their citizens want that”.

That's true but it seems like the people advocating for policy on both ends don't understand the difference in density and therefore transportation needs between Boston and Boise. Everyone wants policy to be a state or federal level cudgel these days and quite frankly that is stupid.

I live in rural Japan and own a car. However, I lived here without one for over 5 years. My apartment building is surrounded on 3 sides by rice fields, so it gives you an idea of just how rural it is. I can walk from one end of the town where I live to the other in 20 minutes. In that space there are 3 grocery stores, 2 hardware stores, a butcher, 2 fish mongers, 1 tofu shop, 2 flower shops at least 5 barbers, 3 doctors, 2 optometrists, many bars, restaurants, cafes, etc, etc, etc. I moved from suburban Canada. In 20 minutes I wouldn't be able to walk out of my neighbourhood of cul-de-sacs lined with identical houses. Not even a single convenience store. It's a commercial waste land.

In rural Japan, cars are used. In suburban Canada, cars are necessary. There are definitely more rural areas in Japan than where I live. I live in an actual town. If you are up in the mountains, or live far away from a town, a car is probably necessary. However, for the vast majority of people who live in this country, it is not.

This is true. But what I wish for is better public transport in countryside. Automated train that can run also late at night and more bus would have relieved most of my woes when I was living in the countryside.

You could even make it so that you drive it yourself and keep it outside your home to use when you need.

What an innovation that would be

Would be great if you could afford the $40k and insurance to house that 4000lb hunk of precious metal you spend energy moving around everywhere you go. A bus pass is a $1.75 on the other hand...

263 million cars in an America.

Don't be silly, not everyone is buying $40k cars. And the bus is a great solution if it exists, but in most places on planet earth it does not.

The bus pass wouldn't be $1.75 if it serviced the rural areas. The economics just don't work with low population density.

Look at NJ Transit, in rural areas if you have a train station you get 2-3 commuter trains in the morning on week days and no service on the weekend or during the day/late at night. And thats for NJ, which has high population density even in rural areas, is physically small and is part of the north east corridor.

Yeah, very clever. There is many reason why one would like to avoid having to take a care all the time:

1. I cannot drive if I went to a friend a drunk alcohol or consume weed

2. Why would I drive if somebody/something could do it for me, and i get to do whatever i want during this time

3. Gazoline, insurance, maintenance, repair, ... a car can be very costly especially if use intensively

4. Cars are polluting a lot, maybe I don't want to contribute to the issue of global warming

And yet there are almost as many cars as adults in the USA.

How do you think all those issues are solved presently?

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

If you visit Europe you can see this system working with a much less dense population and connections for small towns.

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.

I sometimes wonder, from a science fiction point of view, how cities would be designed if we somehow had foresight of the consequences (social, environmental, etc.) of building infrastructure centered around the usage of cars for daily transportation. We don't really hold much resentment for how things started since people just didn't know about the problems that over-reliance on cars would cause decades down the line. Would people still consider the convenience/economic opportunity the automobile afforded to outweigh the problems?

(Also, there is a body of fiction set in a future Earth where, since people several decades ago did have knowledge of the problem of climate change, they end up being collectively resented by their descendants for their inaction/ignorance in addressing it.)

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.

Those pesky local electorates. Always stopping progress.

Maybe, just maybe, the majority of people who live in those cities like the way they're arranged?

A small minority biased towards homeowners who stand to benefit economically from constrained supply and rising costs are the ones voting, not the majority of people who live in these cities.

it's becoming more and more clear. self driving technology will only put more 2000 lb machines on highways clogging up space. USA NEEDS TRAINS!

Not necessarily: a self driving car takes you to a self driving coach for the long distance, then back to sdc for the last miles. All electric. Very efficient.

Not as efficient as a freaking train with hundreds of people transporting basically only what is needed.

Trails will help highway traffic between major cities, but it doesn't address how spread out suburbs are from people's jobs and support infra (grocery stores, doctors, restaurants, etc). It's a chicken-and-egg problem, now:

- people are spread out because they've owned cars, so the distance doesn't bother them

- If you take their cars away, they're too spread out to support themselves, because of how the towns were built.

Totally. American cities we're built with cars in mind, and it's really crippled their potential. Densely packed cities with subways and a lot of vertical freedom are the best way to build. They prevent deforestation because you don't need to expand outward, and they help with pollution because you can have maybe a couple hundred metro cars instead of millions of individual cars. I'm not sure how this would be implemented now though, because like I said, American cities have been crippled by their car infrastructure.

The 'problem' doesn't even exist; non-autonomous cars work fine in America. Autonomy is a luxury product.

Aside from those massive numbers of dead pedestrians. (Edit: not that I think autonomous cars is the answer - but it's clear that regular cars are a problem)

I hear you about pedestrian deaths, but the numbers aren't actually that bad (accepting some level of accidental deaths will occur no matter what we do).

Cyclists coming in at 857 in 2019 and people on foot coming in at 6,283[1].

That's not too trivialize those deaths... but from what we've seen so far, I'm not convinced self-driving cars are going to do any better.

[1] https://www.usnews.com/news/us/articles/2019-10-22/pedestria...

lol, there is a serious car accident in America every 3 seconds. People die every day in car accidents. Is that what you consider "working fine"?

> 'Is that what you consider "working fine"?'

First off, yes. It's a big country so figures like that don't mean a whole lot. I wager I'm more likely to be killed by cheeseburgers than by a car. 50,000 cheeseburgers are eaten every three seconds in America.

But another thing: if Tesla were really trying to save lives, they'd have this software turned on for all their cars. That's not what they're doing. Rather they're selling activation of this software for thousands of dollars. They're selling a luxury product.

Looks like about 4755 every 3 seconds

https://southfloridareporter.com/the-average-american-consum...

Owning your own car seems lime the much bigger waste in the long run. Imagine only needing a fraction of vehicles in total.

Owning your own car will become a luxury for many people. Kinda like owning a horse.

I‘d still rent one every once in a while for road trips and camping though.

I never owned a car, but going anywhere, anytime, at will is a very real thing and cars are the obvious solution. Humans are terrible at safe driving and self driving will eventually solve this. I don't see how this is not the future of transportation.

The obvious solution often isn't the right solution.

We're going to lose the global warming battle if we don't reduce personal cars in transportation. (And no, EVs don't solve it, they still have way too big CO2 footprints)

> personal cars in transportation.

the EPA[1] says 4.6 metric tons of co2 from an average car. There were[2] 1bn cars in service globally, as of 2010.

“It has been estimated that just one of these container ships, the length of around six football pitches, can produce the same amount of pollution as 50 million cars."[3]

1bn/50m = 20 cargo ships equal the estimated world carbon footprint of automobiles.

There are 50,000 cargo ships globally. There's your problem, go nuclear or go home.

[1] https://www.epa.gov/greenvehicles/greenhouse-gas-emissions-t... [2] https://www.worldometers.info/cars/ [3] https://inews.co.uk/news/long-reads/cargo-container-shipping...

Cargo ships looks like a diversion from lobbies. It should be relatively easy to solve (upgrading 50k motors is much easier than redesign transport on land to trains, subway and bus). The trick is to conflate nitrogen oxide emission with air pollution.

When I look at data I find that transport is "20% of CO2 emission" (https://ourworldindata.org/co2-and-other-greenhouse-gas-emis...) but shipping is "2.2% of the global human-made emissions in 2012" (form wikipedia http://www.imo.org/en/OurWork/Environment/PollutionPreventio...)

What does it means ? Well, shipping uses dirty fuel and is very polluting but that's also a scapegoat for individual consumerist culpability.

Have a look at figure 8.1 here:

https://www.ipcc.ch/site/assets/uploads/2018/02/ipcc_wg3_ar5...

Road transport is growinv catastrophically in addition to dominating other transport in absolute terms

I have difficulty understanding how you can present a figure like 4.6 tons per car and argue that it's not a lot, especially as developign countries are headed dangerously toward widespread car ownership.

The "50000x" number about cargo ships isn't CO2 emissions, it's about other pollution due to cargo ships using bunker fuel.

We can go anywhere, anytime, at will after the roads have been built and the pedestrians cleared away, no? The occasional neighborhood also cleared away for a freeway.

All that money and energy could have gone somewhere else.

> Humans are terrible at safe driving

I'd dispute this. I'd say humans are excellent at many aspects of it. The high mortality is simply because driving is inherently dangerous and we do a lot of it.

I imagine the solution is basically what has happened over the last 60 years: gradual changes to improve safe driving.

"Nearly 1.25 million people die in road crashes each year, on average 3,287 deaths a day.

An additional 20-50 million are injured or disabled."

https://www.asirt.org/safe-travel/road-safety-facts/

"Over 90% of all road fatalities occur in low and middle-income countries, which have less than half of the world’s vehicles." Not really relevant statistical baseline here. Self driving cars aren't going to help in areas where people can't afford them.

Not in the near future. Think 100+ years. I can totally imagine owning a human driven car to be the luxury choice.

Seems like we're proportionally better at the seemingly more complex task of driving, than we are at maintaining our health. Heart disease in the U.S alone apparently kills about as much as half the number of worldwide car related fatalities each year.

anywhere there's a road, and not forest, park, building, garden, trains, or people...

I think it's pretty absurd that we don't have automated trains and light rails but somehow we think self driving cars are in the near future. Compared to cars on roads it should be trivial to automate trains on tracks. Why don't we go for the lower hanging fruit first?

> I think it's pretty absurd that we don't have automated trains and light rails

We do have them: https://en.wikipedia.org/wiki/List_of_automated_urban_metro_...

"And within a few minutes we got freaked out with what seem to us as unpredictable action and we turned it off."

On the one hand, I'm affected by the FUD and will not buy a Tesla, and perhaps also will have schadenfreude to the extent their CEO has misfortunes.

On the other hand, to be fair, other driver assistance features are creating problems too. I think it was Subaru I was just reading about having serious problems (and a recall) where automatic braking triggers inappropriately and may cause an accident.

People have incredible faith in software and technology, and that's fine and all, but I wonder what people like that hang around HN for.

I have a Subaru with some sort of "Eye Sight" feature. It is quite handy in maintaining a constant distance from a car in front of you in moderate traffic - but you have to always be ready for someone to cut you off. It tracks the lane markers and rumbles the steering wheel. If a car stops short ahead of you, it beeps and flashes an indicator - I'm usually on the ball enough to hit the brakes myself though, so I don't know if its automatically braking or not.

That being said, there is a place on my commute where going around a curve in one particular spot (southbound highway 17 in Santa Cruz, before the Glenwood Cutoff if anyone is playing at home) where 9 times out of 10 the car thiunks thinks I'm about to drive into a wall, sounds the alarm, and then regains its senses.

So I haven't had it fail in a harmful way but I also don't think I'd ever give it enough rope to hang me with.

I also own a Subaru. The assistance is very unwelcome for the crash detection at times. Living in Minnesota the exhaust from a car in front of you can set it off.

This can happen on start off of a green light and can completely decelerate the Subaru. The eyesight detects the exhaust as obstacle the car behind you is now at a loss why you are now immediately breaking off the green light. I’m not a fan at all and eye sight.

Additionally, you have to turn off the backup sensor to leave my drive way or it will slam on the brakes as it detects flowers that may be leaning into the drive way.

Car smarts don’t seem all that smart. I get a kid could be saved here, but driving a car crying wolf all the time leaves a consumer very frustrated. Dealership has been zero help in both regards.

I don't like the idea of automatic braking, but arguably plenty of people had irrational fears about seat belts and air bags (although the fears about air bags seem to be coming back)...so I will seriously consider the new safety features once my car insurance offers a significant discount for them.

Good ones only engage when you don’t act fast enough. They’ll also press the brake pedal harder if it thinks you aren’t pressing hard enough fast enough.

Ours only engages when it is truly justified. It has yet to flip its shit incorrectly. It’s adaptive cruise control on the other hand will sometimes freak out about cars in adjacent lanes and slow down more than it should.

Our Volvo V40 saved our butt once on the highway. Late at night, cars piling up all of a sudden in front of us. The car slammed the brakes and stopped us a few meters away from the car in front :-)

I also have a V40 and have also had an experience exactly like that on a motorway - bright red lights flashed at the bottom of the windscreen along to a warning sound and the autonomous braking kicked in and stopped me from hitting the car in front. Excellent system.

When I was a kid - this was quite a few years ago - I had some older relatives who refused to wear a seat belt. Being somewhat of a brat, I asked them why. They answered like this:

"If I get in a wreck, I don't want to be trapped in the car. I want to be thrown clear!"

This might be just me, but I don't think it's bratty to question truly idiotic behavior.

I think it's obnoxious to call people "truly idiotic" if you would not apply the phrase to everyone who takes "unnecessary" risks. I am conditioned from childhood to wear a seatbelt in a car, but I also have owned a motorcycle in the past, and currently own a pre-airbag car (which I assume also lacks more subtle features like modern crumple zones and high strength steel, etc.). What makes one risk reasonable and another not?

The Subaru doesn’t have a forward radar does it? I heard it uses two optical camera instead. I bet smoke/water vapor can trip that up...

Our Mazda gets tripped up when adaptive cruise is engaged and someone pulls into the turn lane to make a left. It also can sometimes get “spooked” on the interstate when going on turns and trucks are in adjacent lanes.

Hasn’t freaked out with collision warnings though. Every time it has it was totally justified.

My Mazda has the same issue with cars in front taking exits. I’ve also had it brake unexpectedly on the highway on Rt 95 in RI. I was coming over the crest of a small hill and in front was a stone overpass. The angle of the radar was still elevated so it thought that I was headed into a stone wall. Except the road dips down well before the overpass so in reality that isn’t a possibility.

The warning lights and alarm went off and then the car slammed on the brakes. Thankfully no one was behind me.

Subaru EyeSight uses two stereoscopic video cameras only: no radar, lidar, or ultrasonics. It was engineered to achieve the maximum possible safety improvement at minimum cost. And it basically succeeded at that, but it's very limited. Can only sense straight ahead, doesn't work at all in fog or heavy precipitation, etc.

What year is your Subaru? My understanding was that early versions of the eyesight system were pretty bad, but they fixed a lot of the kinks in recent versions. It would be disappointing to learn that wasn’t the case...

2018 Outback. Which is fairly new’ish.

My Acura’s collision alarm always beeps before coming out of the I90 bridge tunnel on the way to Seattle. It’s not like I’m relying on auto breaking to break or anything, and it’s never actually done anything beyond warn spuriously.

I like Eyesight. It's pretty cool and I like the emergency braking feature. The warning triggers before the emergency brake quite often spuriously but it's nice. I like it.

I think the difference between me and the rest of these guys is that I see it as just an assistive technology and they're trying not to drive. I still have my eyes on the road and everything. It's just there to help, just like I still use my mirrors even when I have the blind-spot detector active. And it does a good job, especially since I can override the adaptive cruise and lane assist at any time pretty easily.

I own a 2019 Forrester with EyeSight. While I dont hate it, and I like the adaptive cruise, I'm curious..

How exactly do you easily override the crash detection logic? I've had it slam on the brakes spuriously due to steam from manhole vents, and short of disabling the feature (which cant be done in a way that sticks between starts) I dont know how I would avoid this behavior, particularly easily.

Oh sorry, I meant that part to refer to the lane detection and adaptive cruise braking. Let me clarify.

I haven't had a fake brake yet and I've driven over quite a bit of steam myself so I guess I got lucky.

Read the manual. You should be able to explicitly turn these off for car wash, etc. They’ll be on again the next time you start.

Subaru EyeSight will do automatic braking under limited circumstances. It only works to reduce the risk or severity of forward collisions when the cameras can see clearly. Overall it's better than nothing and worth the minor extra cost.

I refuse to use Toyota's for the same reason. On I-95 in northern Virginia, people drive slow in the middle and left lanes all the time, so people are always weaving in and out of lanes trying to get around. This means you get cut off a lot. And every time someone cut in front, the truck would slam on the brakes well before it was necessary.

I've driven Toyota's all my life and never once been in a vehicle that does anything automatic beyond turning on the lights when it gets dark.

Maybe you just need to avoid the Toyota model you bought.

Every single new Toyota from 2017 on has adaptive cruise control and emergency braking assistance. You can disable the EBA and use reg cruise control but it’s all on by default.

> other driver assistance features are creating problems too. I think it was Subaru I was just reading about having serious problems (and a recall) where automatic braking triggers inappropriately and may cause an accident.

I had a late model C-class Mercedes for a bit and got rid of it for that exact reason. It auto braked twice in conditions that did not warrant it at all (and hard too), the first time causing the car to come very close to skidding into the support posts of a cantilevered bridge, the second time because of a traffic sign in a turn that apparently gave enough of a return to trigger e-braking. Very bad software. According to the company that sold me the car everything was as it should be.

I remember an engineering class at uni in the early 90's discussing how cruise control worked, and was borderline horrified at the childish simplicity of the design. In all this time, I've probably used it 1/2 dozen times, because my trust level is very low.

Just the other day I was a passenger in a Peugot and it was raining very heavily. Cruise control automatically disabled when one of the tyres suddenly bogged down due to heavy water on the road causing the car to lurch.

Until these systems look at and analyse the road and traffic like we do, I see little hope in an automated system that wasn't deadly by design. For me the number 1 litmus test is to read the slope of the road: if about to go uphill- more gas required, about to hit a crest- back off and maybe allow engine braking.

Sorry but using cruise control in heavy rain is just simply a bad idea. It's supposed to be used when conditions are ideal, not when they as bad as can be. That is human error, in my opinion.

In my driver’s education course back in the early 2000s, they explicitly told us that Cruise Control can dangerously malfunction in rain and should only be used on dry roads.

How is it any more idiotic than maintaining that speed manually?

I'm struggling to think of any combination of hardware/software that isn't asinine on first glance (e.g. feed the cruise control off only the rear wheel speed sensors on a FWD vehicle) where cruise control would be any different than a driver manually holding the same speed by putting their foot on the gas.

I don’t think a normal human would try to maintain an unsafe speed going uphill in heavy rain.

It's been almost 30 years, you might want to revisit that decision.

Sometimes simple designs are good. Older cruise control designs are quite primitive but very effective. Nothing but a simple feedback loop. Car not fast enough? More gas. Too fast? Less gas. The smarter ones (probably ones with electronic transmissions) would even downshift to keep under the preset speed.

I bet the earliest ones were all mechanical too.

Of all the features in a vehicle, I don’t think you need to worry about “standard” cruise control.

You kind of validate the other poster's point that it should be called "driving assistance" like the other half-baked solutions and not "autopilot".

There is no current recall for Subaru EyeSight. Perhaps you're thinking of a different manufacturer?

Subaru did have a recall several years ago due to an electrical problem which could prevent automatic braking (not trigger inappropriately).

https://www.consumerreports.org/cro/news/2015/06/subaru-reca...

GM calls theirs Super Cruise which is true hands-free.

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

I like the idea of super cruise -- it seems like a lidar scan of the freeway would have prevented Huang's death. But, obviously the freeway changes over time so I wonder how frequently GM re-scans the freeways. Lidar is expensive of course, but it also seems so obvious that the lidar needs to be onboard the vehicle.

Not sure but apparently they detect road works and are able to update the maps remotely.

Interesting, thanks.

Tesla should put Lidar in their Semi Trucks. Since they'll be constantly driving the whole country, they would have a lot of regularly updated scan data that they could deploy to their lower price point vehicles that only have cameras.

yeah, as long as you're staring straight ahead. It has a camera that watches your face.

Otherwise they're the same. Except Tesla's keeps getting better with each update and GM stays the same until you take it in to the shop...

> GM stays the same until you take it in to the shop...

Over the air software upgrades are supported by GM.

"When the work is done, an over the air software update will turn it back on."

https://www.wired.com/story/cadillac-super-cruise-ct6-lidar-...

GM claims "over 5,000,000 miles of hands free miles driven and counting."

Current estimates of AutoPilot miles driven is ~ 2 billion miles. [1]

That's 400x the number of miles driven under AutoPilot. I think it mainly comes down to number of vehicles. SuperCruise is only availble on CT6 which sells about 1,000 vehicles a quarter, and not all trims even have SuperCruise. The CT5 is supposed to be offered with SuperCruise eventually in 2020, but that just went on-sale without SuperCruise, and sales figures seem to indicate it is selling extremely slowly.

If you look on YouTube it's virtually impossible to find a SuperCruise video of just an owner taking it for a drive offering their thoughts. There are dealer demos, or pro car reviewer demos, but nothing remotely like the thousands of AutoPilot videos.

You'll encounter an AP driving Tesla multiple times a day in major cities. You may never have driven along side SuperCruise. We'll see if Cadillac can deploy the feature more widely. They promised all vehicles in 2020 will offer it as an option but it doesn't look like it's going to happen.

> We'll see if Cadillac can deploy the feature more widely.

"According to GM president Mark Reuss, the feature will be available as an option on 22 vehicles by 2023, with seven vehicles getting the feature next year and 12 more in 2022 and 2023."

https://www.caranddriver.com/news/a30795396/gm-super-cruise-...

Super Cruise is inherently different from AP. It only works on roads that GM has driven and scanned themselves. It’s a lot more cautious and much less ambitious.

> They should call it "Driver assist" or something like "cruise control plus".

Exactly, similar tech has been named Adaptive Cruise Control since the 90s.

This is the ethics of Tesla.

They are marketing primarily, function second.

My biggest concern is that since the company isn't profitable(?) That these highly risky things are acceptable. Other profitable auto companies cannot be this wreckless.

It's not like they can go after Personal assets.

The pilot's name is: marketing. Tesla is well known for rushing engineering and cultivate market cap. Seeing a pattern? Boeing, Uber, ...

You think "autopilot" is bad? Do you realize Tesla has a DLC mode called "full self driving"?

Autopilot never meant full automatic control. In a plane context, which was the most common until now, it's only used to keep a trajectory steady, and anything unusual has to be taken care of by a human pilot.

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

> Autopilots do not replace human operators, but instead they assist them in controlling the vehicle.

Names don't really mean anything, it could have been named anything. If people think it's 99% reliable and it's convenient, they will use it.

> Autopilot never meant full automatic control. In a plane context, which was the most common until now, it's only used to keep a trajectory steady, and anything unusual has to be taken care of by a human pilot.

From what I've seen, the general public has an inaccurate perspective on airplane autopilot systems. Loads of people think pilots simply push a button then sit back and relax as the plane automatically taxis, takes off, flies then lands itself.

In choosing to name their technology after another technology which the general public has misconceptions about, Tesla chose to inherit those misconceptions.

Edit: I'm surprised there is any incredulity here, I've been hearing people say "planes basically fly themselves" and "all it takes to be a pilot is pushing a button" for years. Most often this misconception is surfaced in casual conversation, but here is one example of the misconception getting published: 'A computerized brain known as the autopilot can fly a 787 jet unaided, but irrationally we place human pilots in the cockpit to babysit the autopilot "just in case."' - https://www.wired.com/2012/12/ff-robots-will-take-our-jobs/

Edit 2: Re: Autoland

Modern airliners have autoland, but always take off under human control. Autonomous takeoff is not used by any airliner.

Anyway, to your point, a modern airline's autopilot system allows pilots to take their hands off the controls. Tesla's manual says that is forbidden: "Warning: Autosteer is a hands-on feature. You must keep your hands on the steering wheel at all times." (page 106: https://www.tesla.com/sites/default/files/model_x_owners_man...)

So no matter which way you slice it, Tesla has chosen misleading terminology.

> From what I've seen

Where? How? That's not a scientific statement.

> Loads of people think pilots simply push a button then sit back and relax as the plane automatically taxis, takes off, flies then lands itself.

Who thinks that? That seems like a hyperbole and you can't possibly know.

> Tesla chose to inherit those misconceptions.

Assuming all those above assumptions are true and assuming Tesla reached those same conclusions as the people that didn't know how airplanes work.

Just loads of speculating and attempted mind-reading...

> Autopilot -> Automatic Pilot

Autopilot is one word and has a specific definition in avionics. It keeps the altitude and heading.

If you rip the word apart and magically transform Auto to Automatic you have Automatic and Pilot, but there is no such thing as an automatic pilots, just pilots, that's why autopilot is one word.

Under that same logic automobiles should be automatic mobility devices that self-drive already.

You can't change the definition of a word to meet your argument. You also can't speculate that people probably might think the word may mean this.

Autopilot !== automatic pilot. That's not a thing.

I wouldn't normally use it as a source, but Urban Dictionary shows that colloquially plenty of people consider autopilot to be complete automatic control (worryingly, even to the point of blackout drunkenness).

https://www.urbandictionary.com/define.php?term=autopilot

That's referring to the human condition of "being on autopilot" (aka. your higher cognitive functions aren't doing much and you're brain's operating primarily on reflex). It's generally only used to describe situations where you've done something dumb because you were impaired or weren't really paying attention.

Just expand and define the words, and don't let your knowledge about how the tech works block your ability to reason.

Autopilot -> Automatic Pilot. Automatic: working by itself with little or no direct human control. Pilot: one qualified to control a vehicle. I'd argue if autopilot requires intervention it is hardly automatic and definitely not qualified to operate a vehicle.

Further, consider other automatic devices in most cars, such as the automatic transmission. How do people expect to intervene with a gear shift during normal operation? Never.

Well, I mean, my auto-mobile requires me in order to be mobile so maybe words are more than the sum of their parts.

It's like how 'act' means one thing, 'god' means one thing, but in the right context 'act of god' doesn't mean a supernatural being intervened.

Are you being intentionally caustic? The phrase "running on autopilot" is an established colloquialism for doing something without being mindful of it, and almost every major dictionary definition of "autopilot" states that it is a system used to navigate something without a human.

Demanding a rigorous scientific study on human understanding of a word is simply not necessary when there are multiple recorded videos of people asleep at the wheel with a fucking banana on a string dangling from the steering wheel to fool the system.

So do you have a problem with the name or a problem with people intentionally being stupid?

People trying to fool the system is not Tesla's fault and indicates they are aware of the warnings alerting them they need to pay attention.

If people are dangling bananas they're doing it for attention and you guys are feeding it and blaming the manufacturer.

Changing the product name will not change anything.

Humans will always be a piece of shit sometimes.

"the general public has an inaccurate perspective on airplane autopilot systems. Loads of people think pilots simply push a button then sit back and relax as the plane automatically taxis, takes off, flies then lands itself."

Is this just being pedantic? If you say the autopilot doesn't land the plane, maybe a pilot would agree for all I know, but "autoland" seems to be a thing, right? When I read about "autoland" systems, I also notice they use the phrase "autopilot-controlled landing". I'm not convinced this is about inaccuracy so much as imprecision.

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

No, this is not pedantic. An autopilot is a complex system that needs full time management. It looks like this [1], and you need to correctly twiddle the dials every few minutes to get the plane to do what you want. It's not a button, it doesn't avoid traffic, and it doesn't know about the broader flight plan. Even fancy autopilots that can follow waypoints and intercept a localizer and autoland and all that - managing them looks more like programming a crappy computer than pressing a button.

[1] https://en.m.wikipedia.org/wiki/Autopilot#/media/File%3AFMS_...

Which means that both pilots 1.set the parameters (not a single button push), 2.follow closely that the plane does what they would have, 3.ready to take over at any moment, as the AP cannot even reliably detect it's not on the happy path; this does take off some of the cognitive load off the pilots, but it's very far from the narrative "push button and have a coffee" that's being peddled here.

In fact, it's more of "hold this heading and glideslope" than anything else - a smarter cruise control, if you will.

"In fact, it's more of "hold this heading and glideslope" than anything else - a smarter cruise control, if you will."

Sounds like Autopilot on Teslas is the same thing as an autopilot on an airplane, so why is everyone angry at them again? Somehow they're responsible for the majority of the public having misconceptions about both?

Somebody is promoting and selling "full self-driving", with fine print specifying "in some indeterminate future". Perhaps that could be...nah, not related at all.

The first automated landing on a commercial flight took place before microprocessors were invented, to give you an idea of complexity. To this day, autoland means that the aircraft follows a radio beam down to runway, pitching slightly up and reducing thrust to idle when radar altimeter reading falls under 50 ft. It's really primitive tech, essentially a line following robot.

Yes, the public holds misconceptions about how much the autopilot on a plane does. But some of their beliefs are accurate, and would be dangerous when extended to a car: like how, when the plane AP is engaged, you don’t need sub-second reaction times for possible obstacles. You do need that in a Tesla.

Fairly sure that you do need that for autoland, to engage TO/GA (for go-around) and take control. Most other cases, AP disengagement indeed means "now we think about what next".

No. Required reaction times would be around 2-3 s. Even if you took 10s, you're not going to die -- you might hit the edge lights though

At which point you're arguing semantics: "it's not the short landing that kills you, it's the impact into whatever you hit there!" You're describing Asiana Airlines 214, whose crash killed 3.

https://en.m.wikipedia.org/wiki/Asiana_Airlines_Flight_214

No I'm saying the pilot flying an aircraft correctly configured for autoland in which the autoland system encounters a fault does not require subsecond reaction times to avoid a crash.

Asiana Airlines Flight 214 is irrelevant because (1) it was not correctly configured for approach, (2) the autopilot was switched off over a minute before the crash, and (3) the autopilot did not fail.

The parent of this thread was an argument over whether a aircraft flying on autopilot needs the pilot monitoring it to have subsecond reaction times to avoid a crash. It doesn't.

> Names don't really mean anything, it could have been named anything.

Names absolutely have meaning. Why do you think Tesla named it “auto pilot” and not something else?

Maybe because it literally means what it does?

No, not really. If you're comparing to planes, they have automated systems that will almost always keep them at a completely different altitude to avoid potential collisions.

Plane pilots have to pay very little attention with autopilot on compared to Tesla drivers with Autopilot on.

> No, not really. If you're comparing to planes, they have automated systems that will almost always keep them at a completely different altitude to avoid potential collisions.

Autopilot in VNAV/LNAV modes does not do this, it follows a programmed flight path.

TCAS is the anti-collision system, but it's a secondary system that only gives information to the pilot. It's up to the pilot to make inputs to avoid a collision.

Pilots are the primary anti-collision system - by using their eyeballs to look for other aircraft when flying under visual flight rules (VFR) and also following clearances/information from controllers when flying under instrument flight rules (IFR).

> Plane pilots have to pay very little attention with autopilot on compared to Tesla drivers with Autopilot on.

This is not correct at all.

There are different autopilots for planes. They differ in capabilities. Mostly the pilot sets a heading, altitude and speed and the airplane will maintain that. The main collision avoidance system is Air Traffic Control telling the pilot where to fly when. Then there are other warning systems that only alert the pilot (terrain and other planes).

The real autopilot in planes is called Auto Flight System and is very recent and needs to be monitored closely. It can lift off, fly and land on it's own, but basically only in good weather condition.

“Suicide Mode” would be as successful, I’m sure.

How about "Californian Roulette"?

Every single time this is posted, I have to remind people that the video posted on Tesla's own Autopilot website [0] says this RIGHT at the beginning:

"The person in the driver's seat is only there for legal reasons. He is not doing anything. The car is driving itself"

They are not saying "use with caution, autopilot is an assistance feature", they are saying "The car is driving itself".

[0] https://www.tesla.com/autopilot

Every time you do, I remind you that the page also makes it clear that current systems require supervision.

I have a car with Mobileeye. Mobileye recently demonstrated fsd in Jerusalem. Fortunately, like everyone else who can read, I know that system is not the same system that I have installed on my car.

Tesla’s wording really isn’t that different here.

Did the car salesman showed you video of that FSD demonstration, while they were describing driving assist features of a car you were buying?

No, but they used to run this ad: https://www.youtube.com/watch?v=UTdR91G2L7c

Several interesting things are here: 1. Construction zones -- ProPilot should not be used there. It isn't realiable when lines are incorrect. 2. The wheel is moved automatically to a large degree. In the video it appears to be a larger movement than the system actually supports. 3. This commercial implies that it is actually a solution for inexperienced drivers, but it definitely is not.

I don't know when that went up there, but that's not demonstrating autopilot - that's demonstrating Full-Self-Driving (FSD) capabilities.

Less than 10 seconds in it's demonstrating functionality that Autopilot does not have - turning from one road onto another, stopping at intersections, etc.

Internet Archive suggests it's been up since late 2016, with the video, and only minor changes to copy since then.

https://web.archive.org/web/20161213042716/https://www.tesla...

> Less than 10 seconds in it's demonstrating functionality that Autopilot does not have

It's on a page called "Autopilot", as the hero video. It's not very clear that this is functionality that Autopilot does not have, and it helps add to the confusion about what Autopilot is and what it is not, something seem keen to take advantage of.

They have had 4 years to possibly push that video down the page, perhaps add a title or a disclaimer saying that this is not Autopilot, perhaps clear up the confusion. Whether it's intentionally like this, or just plain old incompetence, either way it's not a great look for Tesla.

> Current Autopilot features require active driver supervision and do not make the vehicle autonomous.

You're conflating FSD with Autopilot.

Since autopilots on airplanes literally fly straight lines between waypoints, it's pretty clear that whatever it means in airplane contexts has zero bearing on what it means in a car context. Having an airplane autopilot fly in proximity to anything else is a recipe for swift death.

> Names don't really mean anything

This is so, so wrong. Names explicitly set people's expectations.

beside the death, the worst part is that Tesla/Musk were all about sound science bringing benefits to users. Now it's becoming a stats game over PR. And I'm not even sure anybody bought a Tesla for AP. It was like a clear coat class option.

"Autopilot" is not misleading. You just don't know what autopilot means.

Autopilot on planes does not mean the plane is self-flying and the pilot doesn't have to pay attention anymore to take over at a moment's notice. Why would you think it's suddenly different in a car?

People in general think this is what autopilot means since that is how it is portrayed in movies and books.

What movies are you watching where they turn on autopilot and the plane lands itself??

If anything, movies portray planes as extremely hard to fly, and the heroic pilot always has to take over from the autopilot.

My airplane can land itself. I always assumed that was normal.

In the movie “Airplane” the autopilot is an inflatable humanlike doll full on manning the plane. :-)

Regardless of what autopilot means to pilots, it means something else to drivers.

Why? Autopilot is a term borrowed directly from aviation.

Most drivers aren't pilots. They don't know what autopilot actually does. They expect it to drive the car for them, and they probably think autopilot does the same on planes. It was a mistake for Tesla to borrow the term.

You're purely speculating on people's knowledge:

> They don't know

> They expect

> they probably think

The alternative is some people are terrible and there will always be people that abuse nice things on purpose.

They know what they are doing, that's why they are trying to trick the warning systems.

These are the same people on the phones and merging without looking in normal cars.

All of that is speculation, but imo it's more likely than yours.

There's no need to speculate.

IIHS did a survey of 2000 people and a driver assistance system named "Autopilot" received the most responses that overestimated the system's capabilities. For example, "Nearly half — 48% — thought it would be safe to take hands off the steering wheel when using the system." In general, about 50% more people overestimated the capabilities of Autopilot compared to the names used by industry leaders (ex. SuperCruise, ProPilot Assist). Other things users were more likely to think they could do with "AutoPilot" included texting and watching a video.

See https://www.iihs.org/news/detail/new-studies-highlight-drive...

Only 6% thought taking a nap was okay though... 16% thought texting was okay.

So clearly not the majority in all aspects and the other names had similar results for those questions.

Because meanings of words can change depending on context.

Speak for yourself. In general, "Autopilot" is misleading. There's no need to just guess about how people interpret it, we know how they do. See https://www.iihs.org/news/detail/new-studies-highlight-drive...

Reply on news.ycombinator.com