Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> Guy was using his phone not paying attention to the road when testing experimental equipment on a public road. He was there specifically to mitigate potential malfunctions of said experimental equipment, but he was reading reddit or whatever instead of keeping lookout.

> I can't see how this is not clear cut negligence.

1) The driver is a woman.

2) While I think the driver does bear some fault here, they don't bear all the fault. Uber designed an unsafe system that relied on an unnatural amount of vigilance from a single person while simultaneously discouraging that vigilance [1]. They didn't design the car to shut down when one of it's critical safety components (the driver) was not operating correctly, and they didn't even give that driver amphetamines or something to increase their vigilance to artificial levels.

[1] Basically: pay close attention to a boring process while doing absolutely nothing for hours on end. I'm pretty sure that's a classic "humans suck at this" task.



That "unnatural amount of vigilance" is just sitting in a comfy seat and looking at the road ahead of you go by. There is nothing unnatural about that. Lots of people do similar or more boring tasks just fine.

This isn't a truck driver falling asleep from exhaustion caused by aggressive scheduling. You don't "accidentally" take out your phone when your job is looking ahead. That's deliberate negligence.

If you want to require eye sensors to detect distraction, by all means, pass a law about it. Maybe include regular non-AI cars too. #1 cause of accidents over here.


> That "unnatural amount of vigilance" is just sitting in a comfy seat and looking at the road ahead of you go by. There is nothing unnatural about that.

Statements like this seem to indicate a failure to understand what the difficult part of the task was. In short: a boring monitoring task that was practically unnecessary 99.99000% of the time, but absolutely critical maybe 0.00001% of the time. The typical-case lack of necessity would strongly re-enforce engaging with distractions over time (i.e. distraction more rewarding, and nearly all of the time no negative feedback of any kind).

> This isn't a truck driver falling asleep from exhaustion caused by aggressive scheduling. You don't "accidentally" take out your phone when your job is looking ahead. That's deliberate negligence.

I'm not saying there was no negligence here on the part of the driver, just that Uber itself was at least as negligent if not more so for designing the system in the way it did. Focusing too much on the driver is an error.

Honestly, watching a show was egregious, but I wouldn't be surprised if the drivers would end up falling asleep on a regular basis if all distractions were removed from their environment. Half paying attention to something I wasn't interested in has always been the best way to get me to fall asleep.


> That "unnatural amount of vigilance" is just sitting in a comfy seat and looking at the road ahead of you go by.

Doing that for five minutes is easy.

Can you do that for eight hours? Day after day? Without a single lapse in attention?

This is a much harder job then the truck driver has - because he constantly has to make microadjustments, to correct for road conditions.

People's brains don't work the way you think they do.


The driver wouldn't be prosecuted if she just had a lapse of attention. She had taken her phone and was watching a TV show. That's a deliberate act.


I'm not saying the driver wasn't negligent in watching TV, instead of the road. In fact, I don't think anyone in this thread is saying that.

I am saying that passively watching the road for hours on end is much harder than actively driving.


Many people write this notion in this thread. Do you have any data for this BTW? How are you claiming this confidently? Have you driven a self-driving car? Many people (myself included) find it quite easy to monitor the road in it for long periods of time. If the driver was not one of them, she shouldn't have taken the job and risked innocent lives.


https://www.mitre.org/sites/default/files/pdf/pr-3426-lesson... is a 2019 literature review of the phenomenon - section 2, reaction times in response to deviations in automated tasks, is of particular relevance here. (Summary: humans are bad at this)


How are you so confident that you are different from all the Tesla owners whose self-crashing autopilots drove their vehicles into stationary objects, fire trucks, semis, etc?

What makes you confident that you are actually good at it, and are not the victim of Dunning-Kruger? Do you regularly find yourself in the process of stopping your self-serving car from crashing into things?

Or has your car simply not crashed yet?


> all the Tesla owners whose self-crashing autopilots drove their vehicles into stationary objects

All the 3 of them? Among millions of Tesla vehicles out there and the hundreds of millions of miles driven? Is that even a considerable risk when compared to the general (non-zero!) risks of driving?

For me personally I simply know when I watch the road and when I don't. For some people this might be hard, for others - not so much. I am aware of when I pull out my phone or distract myself and when not.


I'm sure you're aware of when you've pulled out your phone, or are playing with your infotainment system. (You should also stop doing it, it's negligent and illegal.)

But are you just watching or are you seeing the road? How many times have you taken control from your car doing something stupid and dangerous?

Unless the answer to that second question is 'I do it all[1] the time, and I'm batting 20/20', what makes you confident that you'll catch the next instance?

[1] If that's really the case, you should probably short TSLA, it doesn't sound like their car can safely operate.


It is rather unlikely, statistically, that this was a single lapse. Or that a single lapse would cause an accident. It's not like the driver's finger was hovering over a nuclear launch button.

Yes, I think I would be able to do this job safely. Not everyone has problems with focus to such an extent that you can't help but be on the phone when you need to be paying attention.


I agree with you except for

> and they didn't even give that driver amphetamines

Yo, what? IIRC, most *amphetamines are either illegal in the US or like FDA Class 1 substances or something like that, and you just want to throw them at people operating > 1 ton machinery? That seems like it could have an even higher risk for danger.


> Yo, what? IIRC, most amphetamines are either illegal in the US or like FDA Class 1 substances or something like that, and you just want to throw them at people operating > 1 ton machinery? That seems like it could have an even higher* risk for danger.

That's true, and my comment was mostly sarcastic (but a little serious). Amphetamines are given to military pilots to increase alertness [1]. They're also used to treat attention deficit disorder (stimulants also increase focus in healthy people, see coffee). Some years ago I read that they're also popular in some parts of Asia not for recreation, but to help workers focus on boring, repetitive tasks (and they apparently can even make that kind of work seem "fun"), but I can't find the exact article again.

The kind of job these "car monitors" are expected to do is so unnatural that it's ludicrous that someone could be expected to succeed at it perfectly, unassisted.

[1] https://aviation.stackexchange.com/questions/8885/do-militar...


Modern cars are essentially self driving on the highway, and it doesn't require an an unnatural level vigilance to operate them. To be fair, the driver's assistance features are usually come with an attention monitor, so Uber is negligent in that aspect.

Also, no one said the driver has to do nothing. They can safely listen to audiobooks or talk on the phone.


talking on the phone can be nearly as distracting as watching TV in some cases. it's also illegal in a lot of places...


How is the required amount of vigilance any different than that of normal driving?


In normal driving, the driver is continually taking actions. Making small adjustments, considering hypotheticals, checking mirrors and speed. If the user is continually engaging with the system, there's less of a risk of the user's attention lapsing. Once the user's input is no longer required, humans are very bad at paying attention to something that doesn't change, and very bad at reacting to the change when it does happen.


I fail to see what stops one of these car monitors from “considering hypotheticals, checking mirrors and speed.”

I get that it’s boring but fail to see how that’s an excuse for not paying attention (or watching streaming video) to a dynamic system you’re expected to actively engage with.

Many train systems are largely autonomous and far more static and yet we still expect train operators to not kill someone on the tracks.

How do you feel about lifeguards as a job? Failure of pools and beaches to design a good system?


Please see the top comment for how lifeguard systems avoid these issues - 30-minute short shifts, overlapping responsibilities, and monitored attention.

Train conductors do not require second-level reaction time. A train cannot stop in five seconds, no matter how quick the conductor is. People who are stopped on train tracks get killed, and the train operator is very rarely to blame.


In what world is asking someone to at least not pick up their phone and start watching videos instead of doing one's job and at least looking at the road outside is "unnatural amount of vigilance"?

It's the basic job description. If she couldn't do it, she shouldn't have taken the job.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: