Live data from Hacker News

The High-Stakes Race to Rid the World of Human Drivers

theatlantic.com

1–10 of 156 posts

Re: The High-Stakes Race to Rid the World of Human Drivers

#2
I just bought a 2016 Volkswagen that has laser cruise control and lane assist. I ordered the fully loaded package because the incremental cost was not large, and I figured I might as well have all the toys if I was going to the expense of buying a brand-new car.

I consider myself an enthusiast; I have a fair bit of time on the racetrack in lapping days and amateur races, sometimes I volunteer at car control clinics and teen instruction days. All of my cars have manual transmissions and I ride a motorcycle. I definitely did not expect to become dependent on any driving nannies.

What I've found is that it's absolutely astounding how fast you come to trust the automated systems. In my car, the laser cruise control will drive the car for you at virtually any speed. It does a great job of keeping up in traffic, has no issues with hills or going around bends, it's always able to track the car in front of me perfectly. It uses the brakes to slow the car. I've got the manual transmission so it only slows the car down to about 15mph and then prompts me to take over and press the brake; if I had the automatic, it would come to a full stop in traffic and only require a tap of the accelerator to resume once traffic ahead started moving.

Even though I only have 700 miles on the car and have only used the laser cruise for perhaps 20 or 30 miles in total (mostly just playing with it), I find I still expect the system to brake the car once disengaged. For example, I'll be coming up to a red light, lift off the throttle, start braking, and then think to myself "huh, I wonder why the car didn't start slowing down for me."

The lane assist feature will correct the steering once or twice if you start to come too close to the lane marker (within maybe 2 feet) but it's very subtle and you barely feel it while your hands are on the wheel. After a correction or two (and maybe 10 seconds, I'm not sure exactly how it decides), the system tells you "hey, idiot, you're supposed to be steering". I have, on multiple occasions, been surprised that it "allowed" me to get as close as I was to adjacent cars, before realizing that it only is meant to keep me within my lane, it knows nothing about the vehicles around me. But something about the fact that it has lane assist (and, incidentally, ultrasonic parking sensors) fools me into thinking it knows more than it does. I'm not saying I was letting the car drive for me, I was fully in control, I was simply surprised that it didn't override my judgement of how close is too close.

It's a brave new world out there. It's surprising how quickly you cede mental control, if not physical control, even to systems that have clearly defined limits and features that you don't even intend on using, let alone relying on. I imagine things will get worse before they'll get better. That said, humans are bad enough drivers that it might not get MUCH worse!

Re: The High-Stakes Race to Rid the World of Human Drivers

#3
anyone work on artificial intelligence technologies here? i have some lay questions - is anyone testing driverless cars in an environment where most cars are driverless? does that make a difference?

are there any weird effects / properties that can emerge?

does it have the potential to solve the persistent traffic jam issue?

Re: The High-Stakes Race to Rid the World of Human Drivers

#4

anyone work on artificial intelligence technologies here? i have some lay questions - is anyone testing driverless cars in an environment where most cars are driverless? does that make a difference? are there any weird effects / properties that can emerge? does it have the potential to solve the persistent traffic jam issue?

I believe so.

As driverless cars take over more and more, the weird (aka dangerous) behaviors should become less common. Many forms of distracted driving will simply vanish, not even considering the obvious issues like drunk driving.

Once you start automating dispatch, I think it simplifies and improves things further. For one, you're going to be less likely to be late.. which means less stress/anxiety and less distractions which should mean less accidents and near-accidents that cause traffic issues.

IF you can eliminate many - but unlikely all - of the little issues, problems, etc that compound to cause big problems, why would there be traffic jams?

(And that's not considering traffic updates to re-route around existing problems.)

My post on the topic: http://caseysoftware.com/blog/future-transportation-today

Re: The High-Stakes Race to Rid the World of Human Drivers

#5

anyone work on artificial intelligence technologies here? i have some lay questions - is anyone testing driverless cars in an environment where most cars are driverless? does that make a difference? are there any weird effects / properties that can emerge? does it have the potential to solve the persistent traffic jam issue?

I believe so. As driverless cars take over more and more, the weird (aka dangerous) behaviors should become less common. Many forms of distracted driving will simply vanish, not even considering the obvious issues like drunk driving. Once you start automating dispatch, I think it simplifies and improves things further. For one, you're going to be less likely to be late.. which means less stress/anxiety and less distr…

On top of that, the cars can communicate more and faster than the drivers, and can be programmed to collaborate with the other vehicles. So you can have stuff like the cars in a packed highway opening up a diagonal for a car in the left lane to take an exit on the right.

Re: The High-Stakes Race to Rid the World of Human Drivers

#6
post #2

I just bought a 2016 Volkswagen that has laser cruise control and lane assist. I ordered the fully loaded package because the incremental cost was not large, and I figured I might as well have all the toys if I was going to the expense of buying a brand-new car. I consider myself an enthusiast; I have a fair bit of time on the racetrack in lapping days and amateur races, sometimes I volunteer at car control clinics a…

I have mixed opinions on the new assistive technology. Blind-spot monitoring is pretty good - it's not intrusive (just a yellow light in your outside mirror) and can save you from an embarrassing accident.

Same with the ultrasonic/radar parking assist. I rarely parallel park (there just aren't that many of that style parking spots any more) so the system really helps me judge where the back corners of the car are. I understand that backup cameras will be required in all new US cars soon.

The lane-keeping systems OTOH are just plain annoying. Whether they vibrate the steering wheel or give it a nudge, they have no situational awareness. Like when I'm crowding the inside lane to ensure I clear a cyclist in the bike lane.

Re: The High-Stakes Race to Rid the World of Human Drivers

#7

anyone work on artificial intelligence technologies here? i have some lay questions - is anyone testing driverless cars in an environment where most cars are driverless? does that make a difference? are there any weird effects / properties that can emerge? does it have the potential to solve the persistent traffic jam issue?

I believe so. As driverless cars take over more and more, the weird (aka dangerous) behaviors should become less common. Many forms of distracted driving will simply vanish, not even considering the obvious issues like drunk driving. Once you start automating dispatch, I think it simplifies and improves things further. For one, you're going to be less likely to be late.. which means less stress/anxiety and less distr…

My offhand anecdata suggests that the bulk of traffic jams are caused by people driving older cars. There is a little-understood cosmic force that causes cars to stop running in proportion to how critical the section of roadway is to regional transportation and in proportion to the age of the car. 1985 Nissan Sentra in a tunnel? Guaranteed to burst into flames.

More seriously, there are obvious reasons why older cars are more likely to just stop running, or to be involved in a fatal accident that snarls traffic. I think the tipping point is not going to come from the deployment of self-driving cars, but will arrive instead when the fleet of junkers nears eradication.

Re: The High-Stakes Race to Rid the World of Human Drivers

#8
One thing I haven't heard much about is how humans and AI drivers will interact.

If we all assume that AI cars are going to be very conservative and very competent drivers, will people take advantage of that? Will human drivers become even bigger assholes that cut off AI drivers at every possibility, knowing that an AI will not hit them? Why wait in a big line for an exit, when you know an AI car will always let you in. The AI driver won't even mind, as it has no emotional state, so why not?

Pedestrians and cyclists may also learn that they can freely cut off an AI car at anytime they please. Why wait for a walk signal? Why walk all the way to the intersection? Just saunter across anywhere you like, so long as you cross in front of an AI car.

Re: The High-Stakes Race to Rid the World of Human Drivers

#10
post #8

One thing I haven't heard much about is how humans and AI drivers will interact. If we all assume that AI cars are going to be very conservative and very competent drivers, will people take advantage of that? Will human drivers become even bigger assholes that cut off AI drivers at every possibility, knowing that an AI will not hit them? Why wait in a big line for an exit, when you know an AI car will always let you…

Yes, people will take advantage of AI drivers, but everything will be recorded and I'm sure the autonomous car companies can figure out how to put those recordings to use. Name&shame, sell the info to insurance companies, push to make it an automatic ticket, etc. After a shakedown period a more or less reasonable compromise will surely emerge.
Post reply on HN