Apple engineer killed in Tesla crash had previously complained about autopilot • Hacker News Board

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

#851

Post by snowwrestler » Wed, Feb 12, 2020, 11:43 PM UTC

Reason077 wrote: post #790

snowwrestler wrote: post #586

Earlier quoted context omitted.

I think autonomous driving advocates would do well to look at the history of computer handwriting recognition, an easier technical problem with lower consequences that received significant investment over decades. But it has never gotten good enough to succeed in the marketplace against alternatives. Why? It never exceeded consumer expectations, which are extremely high for automated systems. Even a correctness rate…

> "handwriting recognition... has never gotten good enough to succeed in the marketplace against alternatives." There is little market demand for handwriting recognition, and thus little active research goes into it. Not because it is a difficult or problematic technology, but because better alternatives exist that make it irrelevant. Even if someone were to come up with an absolutely perfect handwriting recognition…

This is all true today, but it was not true in the past. There was a time in the computing industry when everyone believed that pen interfaces with handwriting recognition would be a crucial enabler for highly mobile computing. Both Apple and Microsoft built major product launches around this idea in the early 1990s.

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

#852

Post by GaryNumanVevo » Wed, Feb 12, 2020, 11:47 PM UTC

greedo wrote: post #812

ReptileMan wrote: post #507

Earlier quoted context omitted.

It is not suicidal, but probably worthy of a honorable mention in the annual Darwin Awards roundup since it was a very bad judgement on his part. As an engineer - he should have known better than to continue to use the feature the moment it became unreliable for reasons unknown. Those kind of failure modes are extremely devious killers.

So all the engineers who experience a BSOD, they discontinue using Windows immediately after?

Totally not the same thing. If my car's steering locks up and I can't turn because of a software glitch, I'm returning that thing to the dealer ASAP.

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

#853

Post by s_y_n_t_a_x » Wed, Feb 12, 2020, 11:58 PM UTC

cjhopman wrote: post #456

s_y_n_t_a_x wrote: post #288

Earlier quoted context omitted.

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 im…

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 indust…

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.

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

#854

Post by dkonofalski » Thu, Feb 13, 2020, 12:07 AM UTC

Piskvorrr wrote: post #836

dkonofalski wrote: post #831

Earlier quoted context omitted.

That's not what happened. The AI was not programmed to drive through anything. It was incorrectly programmed to dump previous data when the categorization changed. It correctly identified at each point that it was meant to slow down and/or stop but, by the time it determined what the obstacle was, the previous data was thrown out and it didn't have enough time to stop properly. In your example, it was more like "Ther…

The car did "see" an obstacle for over 6 seconds and did not brake for it, now someone is dead. You are haggling over semantics to make it look like this did not happen and/or this is not a bug. Atrocious. (Or, more charitably, "oops, somebody forgot that object persistence is a thing" does not excuse the result)

What? That's not at all what I'm doing and you're being extremely disingenuous to suggest that. I'm simply correcting misinformation. The car wasn't programmed to drive through anything. It was programmed to throw away information. Either way, it's an atrocious mistake and I've even said, elsewhere in these comments, that the people responsible for that code should be held liable for criminal negligence. There's no need to lie about my point or my position to defend yourself. That's just silly.

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

#855

Post by farisjarrah » Thu, Feb 13, 2020, 12:49 AM UTC

_pmf_ wrote: post #401

Alupis wrote: post #388

Earlier quoted context omitted.

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…

Comma.ai kind of follows the "improving lane detection until it drives you completely" line, at least according to what I get from this interview with Hotz: https://lexfridman.com/george-hotz/ He's still a bit too optimistic for my taste.

I bought in to Comma.ai at first and even started porting my car over. But I quickly realized that the whole comma.ai crowd is way way to cavalier with safety. One day I was driving down the freeway and saw a small fender bender occur in front of me. I looked up at my dash and it was all ripped up with wires hanging everywhere because I was working on the Comma.ai stuff. If it had been me in that fender bender you know that insurance would definitely blame me for that given that I was messing with the OBDII and radar sensors.

As cool as Comma.ai is, I really believe that their approach to allowing so much community involvement with little to no oversight is highly irresponsible.

That being said.... If they do succeed, and get some sort of government approval or oversight.... You bet I'm putting that stuff back in. Its cool A.F.

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

#856

Post by ijpsud » Thu, Feb 13, 2020, 1:06 AM UTC

saalweachter wrote: post #627

ijpsud wrote: post #555

Earlier quoted context omitted.

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.

There's like fifty caveats that go along with this statement, but this is the internet so I'm just going to skip all that. Something like half your human brain is devoted to visual processing. There's a tendency to think that things like language is what makes the human brain special, or our ability to plan or think abstractly, and we talk about things like "eagle eyes", but the truth is humans are seeing machines wi…

> Vision is not the easy part of AI.

I think it is, actually. Going from raw pixels to objects is the (relatively speaking) easy part. It's the next part (using that for planning and common-sense reasoning) that's the hard part. Machine learning has already advanced past humans in this regard for many classes of problems - which is part of the reason why captchas are getting so hard.

This was several years ago, hence the move away from obfuscated text (which was getting harder and harder to read): https://spectrum.ieee.org/tech-talk/artificial-intelligence/...

I'd be surprised if basic perception tasks as human-ness tests last more than a few more years.

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

#857

Post by Alupis » Thu, Feb 13, 2020, 1:30 AM UTC

emiliobumachar wrote: post #508

Alupis wrote: post #388

Earlier quoted context omitted.

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…

Would Tesla be a heretic in this religion you just modeled? They're all about boring driver assistance, and acceptance of a low number of deaths. They don't even try Level 5 for now, and get a lot of criticism for requiring that the human pays attention at all times.

> and get a lot of criticism for requiring that the human pays attention at all times

Because that's the law... and...

> They don't even try Level 5 for now

That's not what Elon has been telling us for years... Full Level 5 is just months away!

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

#858

Post by Alupis » Thu, Feb 13, 2020, 1:34 AM UTC

ghaff wrote: post #668

contravariant wrote: post #625

Earlier quoted context omitted.

One of the problems with merely fewer deaths than a human operated car is that technology tends to fail in 'silly' ways. Also at the very least an self-driving car should reach the level of a good driver, having self-driving cars cause as many deaths as drunk or inattentive drivers do nowadays isn't defensible. Especially since there's usually no explanation and nobody to hold accountable.

>nobody to hold accountable Well, one of the issues is that someone more or less has to be held accountable. And that someone pretty much has to be the manufacturer. No one is going to hand over full control to a vehicle and accept the responsibility if that vehicle commits vehicular manslaughter because "software isn't perfect." It's actually an interesting legal situation because, other than maybe drug side effects…

What about simple speeding tickets? According to Wikipedia[1], Tesla Autopilot max speed is a whopping 90 Mph!!!

Who's responsible if you get pulled over for going 75 in a 65 mph zone?

[1] https://en.wikipedia.org/wiki/Tesla_Autopilot

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

#859

Post by Alupis » Thu, Feb 13, 2020, 1:36 AM UTC

vkou wrote: post #797

heedlessly3 wrote: post #755

Earlier quoted context omitted.

It is laughable that you don't think the goal wasn't already zero. It has been zero the whole time. Real life and ideals are different thing. You can't promise that accidents will never happen. But you can promise that accidents will be substantially reduced. In the US, about 35k people die per year from motor vehicle related deaths. If you get it down to 10k, then that would be a major success. Of course, you will s…

> In the US, about 35k people die per year from motor vehicle related deaths. If you get it down to 10k, then that would be a major success. If we were actually serious about reducing motor vehicle deaths, we would mandate that every car be equipped with a breathalyzer device. No fancy new technology is necessary, and there's plenty of low-hanging fruit (Impaired driving) that we can deal with. For some reason, thoug…

To take this further, it's been shown every year the numbers are released, that of that ~35k motor vehicle related deaths, ~20k are alcohol related.

On average, humans are actually pretty good at not dying in motor vehicle related accidents - or avoiding them altogether, given the sheer number of miles traveled per day in the US.

That, however, just isn't the narrative Self-driving followers want everyone to know.

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

#860

Post by perl4ever » Thu, Feb 13, 2020, 1:36 AM UTC

Piskvorrr wrote: post #306

perl4ever wrote: post #91

Earlier quoted context omitted.

"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" sy…

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 "…

"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?