Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

theinformation.com

Post reply on HN
Print view View original

41–50 of 196 posts

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#41

Post by dymk »

outworlder wrote: post #32
Judgmentality wrote: post #23

Earlier quoted context omitted.

This is incorrect. The hardest part of developing a self-driving car is predicting the world around you in the immediate future. Knowing whether or not that object is a person is a lot easier than guessing whether or not that person is going to jump out in front of the car 1 second into the future. You have to know who is going to run stop signs, when cyclists are about to cut you off, when someone is about to back u…

You can probabilistically predict those events with machines much better than humans can. You don't really know someone will decide to run a stop sign, but you do know when the vehicle is past the point it was supposed to start slowing down. That's relatively "easy", we have been predicting physical object motions with analog computers, even. As the parent says, accurate data from sensors is a much bigger problem. Bu…

> You can probabilistically predict those events with machines much better than humans can.

This is an assumption and has not been shown to be correct or even probable

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#42

Post by obernaru »

dumbfoundded wrote: post #8
ttul wrote: post #4

The slow progress is not an indictment of self driving. This is one of the toughest engineering challenges ever mounted by humanity.

Yeah, that's a bit of an overstatement. We created nukes, landed on the moon, took sludge out of the ground and used it to power the world. We connected this world with wires and glass fibers to build a real-time global communication system that also gathered all of the world's information in a singular and immediately accessible place. Building a self driving car is hard but really not as tough. How quickly do think…

Depends what you mean by "self driving car". IMO, building a fully self driving car requires artificial general intelligence, which is so hard I'm not sure humanity will ever achieve it. If we ever manage to create the kind of AI that fully self driving cars require, self driving cars will be one of the most boring and trivial things that is done with that AI.

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#43

Post by ForrestN »

kahnjw wrote: post #40
ForrestN wrote: post #13

Earlier quoted context omitted.

I'm not sure if I agree with the writer's statement here, and I'm not sure the right outcome is for Waymo to share data, but I don't think your analogy works. Your individual phone calls are made as part of your general participation as an individual in society. One of the largest companies in the world, which does its best to avoid paying taxes, is using public roads as a fundamental part of the infrastructure for a…

How about UPS/FedEx? They certainly test new products on public roads and have no requirement to make public the resulting data.

As you can read above, I'm not arguing for the mandatory disclosure of data.

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#44

Post by ghaff »

smelendez wrote: post #34
netinstructions wrote: post #3

I wonder how many people are upset that the vehicles had to drop them off in a valid passenger load/unload zone as opposed to the usual Lyft / Uber tactic of parking in a no parking / no stopping zone, bike lane, crosswalk, etc because it's most convenient for the drivers and passengers (at the safety and expense of everyone else sharing the space)... That quote at the end: > I guess Lyft has me spoiled. I like getti…

It seems like there's an overall issue where people feel they can't do anything if the car takes a weird route or drops them off in the wrong place. Another comment mentioned in the story says the car skipped the drop-off location and inched passed a bus stop, and other people mentioned inefficient routing. I don't know how the system works, so this might be user error of some kind, but plenty of people are not going…

This is one of the issues with machines vs. people. For an able-bodied person, like I am most of the time, sure I'm fine with being dropped off half a block away. The driver may ask me and I'll be "fine." If someone is using a walker--not so much.

I have to believe the last block or two problem will be a big issue with self-driving whenever it eventually arrives.

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#45

Post by scarmig »

avocado4 wrote: post #19

10500 rides, 70% are rated 5 stars (perfect). 10% increase from last year.

5 stars isn't perfect. 5 stars is "there's nothing so bad that I want to take the time to complain about it."

Waymo employees, who are encouraged to be especially tough, give reviews that are 47% negative. That's likely closer to the metric for perfect.

We don't get perfection from human drivers either, of course. Though part of the promise of Waymo is much better performance than humans, which it apparently isn't close to yet. And this data is for relatively common cases: for the long tail, one-in-a-million case, the conventional wisdom is that humans would do better.

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#46

Post by tlb »

CloudYeller wrote: post #22

When are they going to just seal off an area so it's self-driving vehicles only, and test them there? In a 100% self driving environment, edge cases are probably much easier to reason about. And if the future is 100% self driving, they should probably get some training data from that environment anyway.

Waymo has quite a large autonomous-only test facility. https://www.theatlantic.com/technology/archive/2017/08/insid... They also use it to populate simulations with realistic traffic models.

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#47

Post by jldugger »

btrettel wrote: post #30
netinstructions wrote: post #3

I wonder how many people are upset that the vehicles had to drop them off in a valid passenger load/unload zone as opposed to the usual Lyft / Uber tactic of parking in a no parking / no stopping zone, bike lane, crosswalk, etc because it's most convenient for the drivers and passengers (at the safety and expense of everyone else sharing the space)... That quote at the end: > I guess Lyft has me spoiled. I like getti…

Cyclist here. In my experience, the majority of the time a driver stops or parks in the bike lane in an urban area in the US (e.g., I live in Austin), there's a legal parking/stopping spot within a reasonable walking distance, often within 50 to 100 feet. (If this isn't true where you live, consider the difference in the location. There are probably exceptions too. I'm told that legal parking isn't typically close in…

I often see waymo vans near San Antonio and El Camino in Mountain View. It's kind of a nightmare drive for all parties. Curbside parking is allowed, there are no demarcated bike lanes, and much of the road is in suboptimal condition. There is often construction going along sidewalks and buildings, and uber dropoffs are common. You occasionally see cyclists, though I suspect most stick to a side street.

What I suspect people are complaining about is that Waymo doesn't do curbside dropoffs at locations with a parking lot -- not common biking routes. I bet Waymo doesn't have the data to know whether a curb is painted yellow, blue, or red, and just avoids them, while a Lyft driver would probably put on hazards and drop people off at yellow curbs and bus stops.

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#48

Post by paxys »

dumbfoundded wrote: post #18
paxys wrote: post #12

Earlier quoted context omitted.

All great achievements, but none of these is in the same ballpark of difficulty as creating a general artificial intelligence (which is pretty much what self driving cars need to achieve true level 5 autonomy).

Level 5 driving doesn't require AGI at all. It's an extremely narrow set of problems that have to be solved incredibly well. It mostly just comes down to creating an accurate 3d representation of the world from a bunch of sensors. You also have to correctly segment and label each object in that 3d representation. If you did those two things extremely well, the actual driving logic can be hardcoded. The problem is tha…

Let's say you do sense and label the entire world (which in itself is an impossible problem). Do you think every single action that a driver takes on the road can be hardcoded?

What happens when a street is temporarily closed but doesn't have the correct signage? What if there's a police office or road worker signalling instructions by waving his hands? What if the lights at an intersection stop working? What happens if there's a car burning on the side of the highway and drivers need to change lanes to go around it?

And these are just some of the problems in a large American city. Think about rural areas, places with more aggressive traffic, places with wildly different written and unwritten traffic rules.

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#49

Post by berdario »

CloudYeller wrote: post #22

When are they going to just seal off an area so it's self-driving vehicles only, and test them there? In a 100% self driving environment, edge cases are probably much easier to reason about. And if the future is 100% self driving, they should probably get some training data from that environment anyway.

They already ran plenty of tests in a controlled environment:

https://www.wired.com/story/google-waymo-self-driving-car-ca...

Re: Waymo’s Backseat Drivers: Confidential Data Reveals Self-Driving Taxi Hurdles

#50

Post by paxys »

ForrestN wrote: post #13
damq wrote: post #6

> The data provide the most detailed view yet of the passenger experience inside vehicles from Waymo, which doesn’t disclose the information even though it uses public roads as its testing grounds. What's up with this statement? Should I be forced to publicize my phone calls just because I made them while driving on public roads? It's utterly bizarre that the author thinks he is entitled to see Waymo's data.

I'm not sure if I agree with the writer's statement here, and I'm not sure the right outcome is for Waymo to share data, but I don't think your analogy works. Your individual phone calls are made as part of your general participation as an individual in society. One of the largest companies in the world, which does its best to avoid paying taxes, is using public roads as a fundamental part of the infrastructure for a…

Companies may or may not have agreements in place with local governments, and those agreements may or may not include provisions for compensation, data sharing etc. However, just saying "you used public roads so you have to make your data public" makes no sense.
Post reply on HN