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

591–600 of 919 posts

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

#591

Post by rwmurrayVT »

arethuza wrote: post #424
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 avoid new technology, exactly because I'm an engineer." I would never call myself an engineer, although I have worked in environments where there were lots of engineers (not in the software world) and one definition of "engineering" that I heard was: "Meeting the requirements while doing as little new as possible"

That's a very accurate definition in the mechanical world. Every junior engineer with an overly complicated, cutting edge, unproven idea gets shot down pretty quickly. In our shipyard safety is more important than anything else. After that is cost.

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

#592

Post by JauntyHatAngle »

JauntyHatAngle wrote: post #239
thdrdt wrote: post #236

Earlier quoted context omitted.

Yes. Current self-driving might be safer than the avarage driver, but what most people forget is how unsafe it is when it fails. It fails harder than most avarage drivers and even bad drivers. In my eyes self-driving should still be called driving-assistance.

Can you cite any sources that AI fails harder? I'm not exactly sure what you are referring to.

Why on earth am I getting downed so hard? It was a genuine question.

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

#593

Post by khendron »

ijpsud wrote: post #555
koheripbal wrote: post #545

Earlier quoted context omitted.

Having driven in Canadian winters, I honestly agree that reliable autonomous driving in inclement weather is indeed decades away. The visual recognition needed is well beyond the systems today.

Isn't that just "image de-obfuscation" though? Seems like narrow AI will be able to out-class humans at that in no time. You can generate as much training data for that as you want. Doesn't really require human-type intelligence. Though I guess you might mean that the obfuscation makes the edge cases even harder, which makes sense.

Here's my go-to example about the challenges of driving in a Canadian winter.

I was waiting for my bus to work one morning after a large snowfall. The snow clearing crews were hard at work, but the street was effectively blocked by piles of snow, men, and machines.

Yet, my bus arrived on time *driving down the sidewalk".

I am not sure how any self-driving system could have figured that out :)

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

#595

Post by Pigo »

ipsi wrote: post #570
tim333 wrote: post #460

Earlier quoted context omitted.

Yeah but speaking as someone kinda on the other side there are things like this 'actual Level 4' delivering stuff for real in Wuhan https://kr-asia.com/jd-com-uses-l4-autonomous-driving-soluti... Fair enough it's not very good - that one just went 600m - but it's hard to argue it won't exist for decades when it exists now. And historically going form sorta works but is rubbish in info tech eg. early cellphones, inter…

The issue is that there are so, so many edge cases to worry about. Simple example: someone decides to troll you in your self-driving car and steps out in front of it. You don't have controls in the car, so you can't go around them - you have to wait for them to move. In reality, it seems like it would resolve quickly - you get out and yell at them, call the police if that doesn't work, etc. But it can get more sinist…

Have you seen the roads in Alabama? They have this weird, red asphalt and rarely any edge lines. In Louisiana they have these elevated roads over the bayou that don't seem to have break-down lanes. In Tennessee they have roads that go through mountains with shocking curves and gradients. Pot holes, broken lights & signs, weird parking lots, new construction, and seeing a small ball roll near the road and knowing that a kid might be coming after it soon.

It seems like it would take an endless list to cover every new edge case. Our technology is a amazing, but I almost think the edge case is places where autonomous driving makes sense.

The thing I trust the least is the operators they want to put in these cars. They better be completely autonomous, self-maintaining, and somehow tamper-proof. It's a really tall order, which I hope we do fulfill one day. But maybe I'm a pessimist, and they have it all figured out already.

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

#596

Post by choward »

PopeRigby wrote: post #146
salimmadjd wrote: post #12

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…

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.

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?

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

#597

Post by JauntyHatAngle »

thdrdt wrote: post #253
JauntyHatAngle wrote: post #239

Earlier quoted context omitted.

Can you cite any sources that AI fails harder? I'm not exactly sure what you are referring to.

Well for example all the YouTube videos that show how Tesla drivers take over the wheel in case of a failure. Those failures include driving straight into barriers. Other systems like OpenPilot show the same. When it fails you better take control of the wheel or you will crash hard.

For clarity, I personally mistrust AI driving and wouldn't be comfortable using it, but the question I have is more along the lines of - if you take the incidence of serious driver error (e.g texting and crashing, speeding and sliding off a road, falling asleep etc) does that happen less often than autopilot going nutso for the demographic driving it? Failing hard seems very possible for both, so stats backing up that AI fails hard regularly seem applicable.

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

#598

Post by overcast »

Alupis wrote: post #388
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…

This whole self driving, fully autonomous thing has become some sort of strange faux-religion. It's followers see no other possible solution as acceptable. Instead of putting our heads together to make really good driver assistance technologies and being satisfied, the darn thing needs to also drive itself everywhere otherwise we're leaving something on the table! Untill we have zero deaths, we cannot stop demanding…

Exactly. For anything but open highway, self driving is so far into the future, that we should be focusing on making smaller pieces safer, ultimately building up to the end goal. If there is a self driving car in my lifetime that will drive up the windy mountain road to Yosemite, in the winter, then I'm the Pope. This all or nothing shit has to stop.

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

#599

Post by dkarl »

Obviously Tesla is going to hide behind the consumer's decision here. Autopilot scared the shit out of this guy on multiple occasions, and he kept turning it on and trusting it with by far the most deadly thing he did every day. Odd choice when you put it that way, but one a lot of people make, I'm sure, maybe one that I would make.

I really hope the government isn't too aggressive in hindering companies from doing self-driving car R&D on public roads, but I would be fine with the government saying consumers shouldn't be used as guinea pigs for systems that are less safe that typical human drivers. Anecdotal evidence can't establish that self-driving cars are less safe than humans (even if it establishes that they make mistakes that humans wouldn't) so I would expect a numbers-driven standard, but self-driving technology shouldn't be unleashed on consumers until it achieves at least a human level of safety.

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

#600

Post by JauntyHatAngle »

freeqaz wrote: post #254
JauntyHatAngle wrote: post #239

Earlier quoted context omitted.

Can you cite any sources that AI fails harder? I'm not exactly sure what you are referring to.

I believe he's alluding to the nature of the accidents. They're high intensity events which are more likely to be fatal (speeds were ~70mph). They're not fender benders when the autopilot fails.

But isn't that what autopilot is used for? High speed Highway traffic? I don't trust AI cars yet but I'd like to know my instinct is true on this and not just my natural inclination to avoid unfamiliar tech.

These are high risk areas, if autopilot is "failing hard" with a regularity equal to or higher than normal than this would be good to demonstrate with stats. Guessing Tesla doesn't really release that info?

Post reply on HN