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

New Study: Waymo is reducing serious crashes and making streets safer

waymo.com

Post reply on HN
Print view View original

271–280 of 413 posts

Re: New Study: Waymo is reducing serious crashes and making streets safer

#271

Post by globular-toast »

johnfn wrote: post #37

As someone who is often on SF city streets without a car - I bike and run a lot - I absolutely love Waymo. I am continuously seeing human drivers cut me off, perform illegal maneuvers (i.e. run red lights when I'm going through a crosswalk), and break various other traffic laws. All these things genuinely put people in danger. Just the other day, a guy started running a "no right turn on red" lane in SF, and when I p…

They are now because winning trust is their biggest hurdle. They've got the "public risk" slider turned all the way down. Let's hope they don't later start to optimise for speed and realise that people probably won't just step out due to fear of death and it's in their best interest to nurture that fear like human drivers currently do.

Re: New Study: Waymo is reducing serious crashes and making streets safer

#272

Post by Mawr »

sneak wrote: post #149
wffurr wrote: post #31

Earlier quoted context omitted.

This also came up recently in a thread about speed governors: https://news.ycombinator.com/item?id=43812856 Just a few governor-equipped vehicles (e.g. from public fleets or prior speeding offenders) can reduce speeding on the whole road.

On most roads, speeding isn’t a safety issue and shouldn’t be discouraged, as road throughput is important. Speed limits are usually set too low due to standardized inflexible policies or desire for revenue. On some roads, however, it is a massive safety issue, and everyone is driving unsafely because the road is designed badly for its intended purpose. (So-called “stroads” are the canonical example.)

Yeah and we should definitely leave deciding the appropriate speed to the drivers. Because, uh, throughput is important, got it. Is safety important? Who knows, but throughput definitely is!

Re: New Study: Waymo is reducing serious crashes and making streets safer

#273

Post by rlue »

johnfn wrote: post #37

As someone who is often on SF city streets without a car - I bike and run a lot - I absolutely love Waymo. I am continuously seeing human drivers cut me off, perform illegal maneuvers (i.e. run red lights when I'm going through a crosswalk), and break various other traffic laws. All these things genuinely put people in danger. Just the other day, a guy started running a "no right turn on red" lane in SF, and when I p…

The one time I ever rode in a waymo (in Los Angeles), I had a contradictory experience. My Waymo was attempting to make a right turn at a red light. We were stopped behind a human driver who was waiting for pedestrians to finish crossing before proceeding to make the turn. This was a college campus (UCLA), so there were lots of pedestrians. After a few seconds of waiting, the Waymo decided that the driver ahead of us was an immobile obstacle, and cut left around this car to complete the right turn in front of it. There was only one lane to turn into.

Luckily, no one was hurt, and I generally trust a waymo not to plow into a pedestrian when it makes a maneuver like that. I also understand the argument that autonomous vehicles are easily safer on average than human drivers, and that’s what matters when making policy decisions.

But they are not perfect, and when they make mistakes, they tend to be particularly egregious.

Re: New Study: Waymo is reducing serious crashes and making streets safer

#274

Post by Mawr »

indigodaddy wrote: post #34

How do AI cars fare with those instinctual "decide which way to swerve in a split-second" scenarios that come up maybe once every 10-20 years over the course of a driver's history?? It's happened to me about 2-3 times and I've always made the "(assumedly) correct" split second swerve decision. Wondering if that is a "human/instinctual" skill and if AI cars do just as well or better, or perhaps not as good? I don't ha…

- Humans have awful reaction times, so the "AI" cars should fare 10-100x better. "split-second" is laughably slow to a computer.

- How seldom these scenarios come up for human drivers is a huge disadvantage for them. For self driving, it doesn't matter, the cars' reactions can be simulated in arbitrary scenarios as many times as needed, so even the rarest of scenarios can be ensured to be handled properly.

- There's nothing special about the decision to swerve vs to say, brake. I'd expect self driving cars to not need to swerve nearly as often because the need to swerve probably only ever exists due to excessive speed and/or poor following distance to the vehicle ahead.

> It's happened to me about 2-3 times and I've always made the "(assumedly) correct" split second swerve decision.

Easy question: Did you make that decision with full awareness that you would not end up in a collision path with another vehicle by swerving? Oops.

Even if you did, how many drivers do you think would "instinctively" swerve into another lane and get hit by an oncoming vehicle because they do not maintain constant situational awareness around their vehicle? The majority, at least.

Re: New Study: Waymo is reducing serious crashes and making streets safer

#275

Post by babyent »

johnfn wrote: post #37

As someone who is often on SF city streets without a car - I bike and run a lot - I absolutely love Waymo. I am continuously seeing human drivers cut me off, perform illegal maneuvers (i.e. run red lights when I'm going through a crosswalk), and break various other traffic laws. All these things genuinely put people in danger. Just the other day, a guy started running a "no right turn on red" lane in SF, and when I p…

I saw a waymo break a red light yesterday in Nob Hill. I think they’re cool but I exercise extra caution around them.

Besides, this is a study on Waymo probably influenced by them too to publish on their blog.

Re: New Study: Waymo is reducing serious crashes and making streets safer

#276

Post by Mawr »

indigodaddy wrote: post #71
JumpCrisscross wrote: post #45

Earlier quoted context omitted.

> those instinctual "decide which way to swerve in a split-second" scenarios that come up maybe once every 10-20 years The correct answer is almost always to hit the brakes. Not to swerve. And Waymo will hit the brakes earlier than you or me.

I disagree. High percentage of these scenarios are at speeds and scenarios where hitting the brake would not prevent a collision and a quick swerve is the only option to not collide. Obviously your quick swerve could cause another collision so I guess it's which decision on average causes the least harm/death? I agree though that most people's extinct would be to brake. Mine never has been though.

Right, so you've created a dangerous scenario by speeding and by the direct consequences of your actions you're now forced to execute a dangerous manoeuvre of swerving. Maybe, just maybe, the solution is not speeding.

> I agree though that most people's extinct would be to brake. Mine never has been though.

...

You operate a motorized vehicle and your first instinct when seeing anything dangerous ahead is to do something other than braking?

Re: New Study: Waymo is reducing serious crashes and making streets safer

#277

Post by notTooFarGone »

riquito wrote: post #204
manquer wrote: post #150

Earlier quoted context omitted.

Here is one, Humans drive in all weather conditions on all types of roads and also many types of personal vehicles of varying ages and conditions. Waymo is limited to few specific locations with decent roads and does not drive in poor weather and is limited to a relatively large and safer expensive SUV that is maintained professionally in a fleet. Studies like this rarely account for such factors , they are compare o…

> Waymo is limited to few specific locations with decent roads and does not drive in poor weather the study is comparing Waymo to accidents occurred in the same cities where Waymo operates, and my understanding is that Waymo drives 7 days a week, 24h a day in those cities, so same roads, same weather. Seems a legit comparison

Also there is some sort of bias not accounted for: People drive when most people drive and most people are stuck in the most dangerous area: traffic. Waymo driving at night on empty streets is not a good indicator for accident prevention when measured against the average human, who is stuck mostly in traffic.

Re: New Study: Waymo is reducing serious crashes and making streets safer

#278

Post by Mawr »

umbra07 wrote: post #173
bluGill wrote: post #135

Earlier quoted context omitted.

If you cannot brake you are following too close. I know everyone else does it (except 3/4ths of semis!), but you can be the one who maintains a long following distance and thus can stop in time. Every time setup the long following distance 5 cars jump into the gap - but then the gap remains as no more do.

what about when there's an idiot slightly ahead and in the next lane who decides to randomly swerve into my lane? do you also maintain a long following distance when there's a car right next to you in the next lane? I try to, because I don't want to stay in someone's blind spot, but sometimes it's not really possible to fall back.

I watch a lot of car accident videos and I've seen that happen exactly once.

Perhaps you mean the far more common scenario when the car in the next lane simply decides to merge into yours? Nothing about that is random[1] and the response in 90%+ of cases is just to let off the gas for a few seconds. That's it.

[1]: In most cases, it's because your lane is open and theirs is about to be backed up. You'd want to switch lanes too, so it really shouldn't be surprising they do.

The rest are mostly people realizing at the last second they want to turn right/left at an upcoming intersection (or highway exit). Again, predictable.

Re: New Study: Waymo is reducing serious crashes and making streets safer

#279

Post by _bin_ »

Lammy wrote: post #116
sollewitt wrote: post #70

Earlier quoted context omitted.

Fellow SF cyclist: Even setting aside the malicious SF stuff, Waymo's have enormous advantages over humans relying on mirrors and accounting for blindspots. I never have to be concerned a Waymo hasn't seen me. I can't wait until the technology is just standard on cars, and they won't let drivers side-swipe or door cyclists.

> I never have to be concerned a Waymo hasn't seen me. Funnily enough that's exactly why I don't like them. Every time one rolls by me I know that tens of photos of me and even my 3D LIDAR scan get piled in to some fucking Google dataset where it will live forever :/ Their site is even proud of it: https://waymo.com/waymo-driver/ section titled “Keeping an eye on everything, all at once” “The Waymo Driver's perceptio…

Second this gripe. At this rate they're going to turn American cities into copies of beijing or london with cameras every other place you look. "Oh but the police will then be able to subpoena footage to catch criminals more easily" yes I don't want a world in which the government can instantly do that. It sucks.

Re: New Study: Waymo is reducing serious crashes and making streets safer

#280

Post by ringeryless »

i am suspicious of all the anti-human-driver comments and all the dismissal of any concerns about Waymo in the comments here.

I am not convinced that public testing of such services is safe, let alone commercial service. One cannot punish a self driving vehicle in any meaningful sense. Corporate incentives vs the public commons, is a general concern that cannot be sweettalked away.

The metaphors about human drivers recording you also seem like reductio ad absurdum.

puff pieces like this should not be well received on HN or it discredits any pretence at separation of concerns with regards to HN and ycomb.

Post reply on HN