Anecdotally, both from riding in them and walking/driving next to/around them, this feels obvious. They never get distracted. Sure, they sometimes make mistakes, but the mistakes are never "I didn't see that". They see better than humans in all cases (where they operate). They react faster than humans.
The one case where they hit a child, it was because the child jumped in front of the car. And they showed that they hit the child at a lower speed than a human would have because of the reaction time.
I would rather be in an area where only Waymo's are allowed than an area where they are banned.
When I visited LA, I rode in a Waymo going the speed limit in the right lane on a very busy street. The Waymo approached an intersection where it had the right of way, when suddenly a car ignored its stop sign and drove into the road.
In less than a second, the Waymo moved into the left lane and kept going. I didn't even realize what was happening until after it was over.
Most human drivers would've t-boned the car at 50+ km/h. Maybe they would've braked and reduced the impact, which would be the right move. A human swerving probably would've overshot into oncoming traffic. Only a robot could've safely swerved into another lane and avoid the crash entirely.
Unfortunately, the Waymo only supported Spotify and did not work with my YouTube Music subscription, so I was listening to an advertisement at the time of my near-death experience. 4.5 stars overall.
> Unfortunately, the Waymo only supported Spotify and did not work with my YouTube Music subscription, so I was listening to an advertisement at the time of my near-death experience. 4.5 stars overall.
This detail sent me, it's crazy that we can pay $25 to have a life saving robot take us across the city yet Spotify is going to blast ads at us the whole time for the sake of making an extra $0.18 (yes that's the actual number) per hour of listening time.
That's around 44.64 (0.18831) per month, no wonder ads are preferrable to companies over subscriptions! That's actually a lot for people that listen to music all day every day at work.
> it's crazy that we can pay $25 to have a life saving robot take us across the city yet Spotify is going to blast ads at us the whole time for the sake of making an extra $0.18 (yes that's the actual number) per hour of listening time.
Oh, the self driving car business will get there, believe me. This is just the first iteration. Designed to get everybody on board with the idea.
I wonder if Waymo gets a cut. I also wonder if riding in a Waymo at the time signals that you're in a demographic that can afford a Waymo and thus get more expensive ads.
How ironic that an Alphabet company, Waymo, only works with a competitor streaming music service, Spotify, and not their own, YouTube Music. I guess that shows how separate they are.
I think it's also a privacy thing; you have to go into the Waymo app and “connect” your YouTube Music account (even though both have the same @gmail.com address), because otherwise the terms of service of one do not allow sharing data with the other without user consent. (Contrary to popular perception Google is very finicky about privacy, at least privacy as defined as conforming to the terms of service.)
That's good news, if I can't use the Youtube Music I've paid for in the Waymo then I'm not going to put up with Spotify Ads instead, better to sit in silence (or use my headphones and my own music)
Nearly got T-boned in a Lyft in LA. I am lucky to still be alive as the driver was not aware and should not have been driving. Where available I've stopped using human driven rideshare.
Thankfully they've now shipped their own product, YouTube Music.
And Google Pay, imagine my surprise back in 2021 when I signed up for Waymo and realized I had to manually type in my credit card. No Google Pay??? C'mon y'all, you're Alphabet!
> I was listening to an advertisement at the time of my near-death experience.
You'll probably never forget that advertisement, which is an exciting business opportunity for Waymo.
They could partner with Spotify and other media content partners so that the Waymo can generate an adrenaline-rush near crash experience when a premium advertiser's ad is playing. /s (hopefully)
This is one of those comments that made me laugh nervously. It's straight out of Ubik or another PKD novel, which probably means it's less than 5 years away from being real.
I hope you are misremembering. Swerving is most often the wrong choice, and I would be disappointed if Waymo were opting for that. By far the best option is to panic stop. Human or robot, physics is a harsh mistress and swerving is more likely to make you lose control and end up in a much more unforgiving wreck.
It wasn't possible to stop at the speed the Waymo was moving at.
The Waymo didn't have the stop sign, the other driver did, at a three way intersection.
The other driver decided to suddenly enter the intersection, when the Waymo was like 5-10 meters away. This was after having stopped at the stop sign.
Either they weren't looking or intentionally trying to cause an accident. Swerving prevented the Waymo from crashing at 40 km/h into the driver's door.
For a human this advice is true. But what if a computer can near-instantly calculate a perfect swerve within the performance envelope of the car and driving conditions?
The optimally regulated system doesn't know the road conditions that well. When the road surface is more slippery, it has the most profound effect on lateral friction, way more than braking.
The Waymo driver can measure the speed and the acceleration of the offending car and calculate, within at most tens of ms, its range of likely future trajectories. And it can calculate its own likely trajectories under maximum braking. And it can track exactly where all obstacles are that would matter if it swerves. All at once. And it can execute that emergency lane change with the control input that is least likely to cause a loss of control and most likely to successfully avoid the other car. It even has processing power to spare to keep playing that Spotify ad!
This depends a huge amount on car, driver and situation. It was the right advice for a learner driver in the 90s with no stability control, no experience and no side airbags, because if you’re going to hit something, hitting it front on is the least risky way. I’m not convinced it’s the right advice for a competent driver in a modern vehicle.
It is still standard advice today, as far as I know. Tires are better, stability control is better, but all else being equal you are still much more dynamically stable and have a lot more friction with the road when the car is stopping in a straight line than when swerving.
Also, in the case of someone running a stop sign, it is far from a sure bet they are going to hit their brakes at all, so by swerving you increase the odds that you will still hit them, but now it will be while you have exhausted all your adhesion on lateral control. So now instead of a front end collision with all the benefits of airbags and crumple zones, you are at a significant risk of rolling the car or spinning off the road and hitting something immovable with a part of your car lacking crumple zones.
The common mistake is people swerve and brake, which is a terrible combination - you should accelerate through a sudden manoeuvre, as it maintains control through it, much as you should accelerate through corners in general.
Absolutely. I was recently driving on a motorway in Portugal when a boulder (giant chunk of granite, 10+ tonnes) fell off the back of a truck - right in front of us, in a heavily laden (7 pax and luggage) car. Immediate massive cloud of dust, I checked my blind spot, veered across two lanes, and continued our journey, unscathed. I looked in the rear view, to see the car behind us jump on the brakes instead of evading. They caught the boulder.
Nobody killed, according to the news, but several taken to hospital in critical condition.
Oh, I say unscathed but our tyre exploded the next day, as apparently we caught a fragment, and again, that’s not a “slam on the brakes” moment, but rather “trundle to a stop on the shoulder and walk to the conveniently nearby tyre shop”.
I think you're over-playing how decisively a Waymo will move and under-playing how decent the average human is.
I've ridden in Waymos. They don't exactly slap on the blinker and move at the limit of traction like someone about to miss their exit. If cut off they absolutely will go full brake rather than perform any sort of spicy lane change or turn.
> If cut off they absolutely will go full brake rather than perform any sort of spicy lane change or turn.
Essentially, a meat driver was waiting at a stop sign to make a turn onto the main road. I was in a Waymo driving on the main road and did not have a stop sign.
When we were 10 meters away from the intersection, the meat driver suddenly started to enter the intersection. I have no idea why.
Full brake would've hit the other car in the driver's side door at 40 km/h.
> under-playing how decent the average human is.
I got to SMFC in CSGO which means I'm in the top 3% of players in clicking on heads within 500 ms of them appearing on my screen. I have never reacted as fast as that Waymo did.
If instant (<50ms) reaction would have lowered the speed only to 40km/h in 10m, Waymo was going too fast for the intersection IMO.
My experience is that for a human driver to react quickly in city driving conditions, style and prep are more important than reaction time: in the case you describe (entering an intersection with another car waiting on a stop sign perpendicular to your path), I'd have my foot hanging over the brake and off the gas pedal — this has helped me avoid hitting many other cars with inattentive/distracted/bad drivers, and even pedestrians running over the road or a red light on a crosswalk. When you are prepared and looking, you slam the brakes much faster!
To be fair, we are not provided with the sensors to swerve safely.
If we had some sort of 360 constant recording in the car (on screen?) it would be safer for humans to swerve. Instead we have to move our head, which is cheaper but lacks info.
That's why we now have rear cameras
We have rear cameras because people DONT move their head. And because regulations have made cars way taller than they need to be, meaning there is a big blind spot close to the ground
I mean, even in low cars you cannot see a small enough kid walking behind your car. That's why you back slowly.
Back when I just got my driver license, there is a big lesson many drivers go through (in Italy) which is you back off a parking and there is an obstacle that's so low that cannot be see through the back window and it's small enough that cannot be seen through the mirror. You hit it and if you followed the "go slow part" you only damaged the paint.
So I'm not opposing the ideas of rear cameras, but I'm totally against tall cars, because you cannot see kids IN FRONT either now.
Depends on how you define "average driver": what if 95% of the crashes are caused by 5% of the drivers?
My reading of all the human crash stats has been that majority of them happen when human drivers are impaired (drunk, drugged or too tired): as this is something we could (in theory, at least) control, I'd like to see and compare with stats for non-impaired human drivers too.
Then, I'd like to see it compared to attentive, non-distracted drivers too (but we won't have crash data for this, as they would avoid most potential crashes).
Note that I am only talking things under every human driver's control, and not things like skill, reaction time, etc.
Also, modern cars (like Waymos) will have a much lower braking distance compared to "average": eg. my Volvo has 35m braking distance from 100km/h or 62mph compared to 50m (45% more) listed as average (excluding reaction distance) — so from 50km/h, it should be around 8m!
In Houston the stats suggest that every driver should get into a crash at least once every. But many ppl haven’t been an crash all their lives and more have been in multiple
> Waymo saved my life in LA.
When I visited LA, I rode in a Waymo going the speed limit in the right lane on a very busy street. The Waymo approached an intersection where it had the right of way, when suddenly a car ignored its stop sign and drove into the road.
> ...
> Unfortunately, the Waymo only supported Spotify and did not work with my YouTube Music subscription, so I was listening to an advertisement at the time of my near-death experience. 4.5 stars overall.
Here is a photo I took inside of the Waymo outside of an Erewhon. Going to Erewhon and experiencing the $20 Hailey Bieber smoothie was on my brother's bucket list and riding in a Waymo was on mine.
He said it was overpriced but bought it again from a different Erewhon so I assumed he liked it (Canadian understatement).
There's apparently a quality gap between locations. The pre-Waymo one was from Erewhon Grove and was freshly blended. Erewhon Beverly Hills on Rodeo Drive premade a bunch of them and left them lying around for a while before selling.
My brother's theory is that Erewhon Grove customers are people who legitimately wanted a smoothie and Erewhon Beverly Hills customers just want photos with the smoothie since it was very popular on Instagram at the time.
Most surprising fact was despite being a licensed product, it was better than the best non-licensed smoothie (coconut cloud).
Licensing deals should make the product worse because the royalties cut into the product margin. The company cuts costs or doesn't take creative risks as a result. But somehow Erewhon resisted these pressures when designing the Hailey Bieber smoothie. We had a discussion about why that was the case but couldn't come up with an answer.
Waymo as a system has crossed the threshold where I trust them more than average driver, but all this hardware is relatively new, well maintained, and their software is closely tied to it.
I’m way less confident of self driving in the hands of the general public when differed maintenance often results in people and even companies driving with squealing breaks and balding tires etc.
I am also not looking forward to the system transitioning from "big experiment, burn money to make it good" to "established business unit, tweak it to death for incrementally more money / personal promotion." We're still in the honeymoon period and I very much expect to hate Waymo in 10 or 15 years when they reach a steady state.
What levers are there, really? Waymo has a monopoly and it seems like they will for a while, so they have a lot of power, but all I really see them doing is making it expensive. Anything that makes the experience worse takes away from their ability to take market share away from Uber/Lyft.
Self-driving vehicles need aircraft-type maintenance. Yet there's nothing like the FAA to enforce a minimum equipment list, maintenance intervals, or signoffs by approved mechanics.
Is there a scratch or chip in the scanner dome? Are both the primary and backup steering actuators working? Is there any damage to the vehicle fender sensors? Is dispatch allowed with some redundant components not working? If so, for how long?
Here's the FAA's Minimum Equipment List for single-engine aircraft.[1] For each item, you can see if it has to be working to take off, and, if not, how long is allowed to fix it.
There's nothing like that for self-driving land vehicles.
What's the fleet going to look like at 8 years of wear and tear?
> Self-driving vehicles need aircraft-type maintenance.
That's a hyperbolic false equivalence.
Aircraft typically carry hundreds of people and can crash to the ground. As long as a self-driving car can detect when it is degraded, it can just stop with the blinkers on. Usually with 0 - 2 people inside.
The question is how broken can a car be when dispatched. What's the safe floor? See the other article today about a Tesla getting into an accident because of undetected sensor degradation.
Waymo's software has crossed multiple generations of sensors and vehicles over almost two decades. It does not seem to be tightly coupled to a particular device.
Not tightly coupled in obvious ways, but as I understand it they aren’t putting it on pickup trucks, convertibles, or anything toeing a boat etc. Their vehicles don’t have aftermarket suspension systems dramatically changing handling characteristics, or turned one into a stretched limo etc.
Which means the software can safely assume the vehicle will behave within a relatively narrow operating range.
I suppose owners will be motivated to have the thing do the driving (and so seek defeat devices and such), but at least the software can have "do nothing" as a safety mode if it manages to detect that the vehicle is not configured as expected.
And maybe the software can be designed to be coupled to a vehicle dynamics model that can be updated.
I don't think the vehicle performance really matters in the typical case. They're using like 20% of what the vehicle "can" do. They're probably hedging against the long tail of variance on the road somehow. Kinda like how private people can tow whatever the f they want with their pickups but in a work setting you need to keep it fairly stupid proof.
The new (as of now than a year ago) Waymo cars still had human safety drivers last I saw one (a month or two ago). I also don't see them taking customers. So they do seem to slow roll hardware rollouts.
We're not even a decade beyond some poorly conceived software crashing two otherwise functional aircraft into the ground and now it's going to save us all...
There is also a different kind of increased safety. There is no driver. No weird conversations about slaughtering goats, no sexual advances. No worrying that your driver is going to assault you or attempt to kidnap you. I know, it's all very far fetched, and Uber/Lyft drivers are almost always nice, courteous and professional, but I have experienced a few times when that hasn't been the case. With Waymo, it's not even an issue.
> There is also a different kind of increased safety. There is no driver. No weird conversations about slaughtering goats, no sexual advances. No worrying that your driver is going to assault you or attempt to kidnap you.
There are also new risks that weren't possible before. A software error can send you into oncoming traffic. Hackers can gain control of your vehicle either directly/remotely or by cleverly designed signage placed on the roadside. A disgruntled waymo contractor in the Philippines can remote drive you into a crowd of people. A flashing stoplight can leave you stranded at an intersection. The car may not see or react appropriately any number of uncommon hazards that human drivers would recognize and avoid. Only a relatively small number of these cars have been on the road, in limited conditions, and only for a small number years. There will be failures and risks we haven't even imagined yet.
One of these sets of risk is mostly theoretical (aside from the large scale stoplight outage), one of them is happening often enough that anyone who takes rideshare repeatedly will have a story.
If we limit ourselves to risks that have actually manifested, not hypothetical risks, I'd rather risk getting stuck at an intersection if there is a city wide power outage than deal with the weird conversations I've had on rideshares (not even counting the countless drivers who demonstrated that it is possible to drive a car without crashing for the duration of one rideshare ride without taking your eyes off the phone for more than a few seconds at a time).
> A disgruntled waymo contractor in the Philippines can remote drive you into a crowd of people.
They cannot. The remote drivers for Waymo offer "nudges" to the robot driver, but they cannot do full remote control.
They can effectively mark a dot in the middle of a crowd of people on their tablet and say "Your best course of action is to drive here", and the waymo very well might decide to try and follow that suggestion, but they cannot override Waymo's brakes nor coded-in "do not hit humans" mandate, and the waymo would stop before hitting anyone.
> Only a relatively small number of these cars have been on the road, in limited conditions, and only for a small number years.
The average uber driver has driven fewer miles on the road than Waymo's software, and hasn't seen all the conditions either. Most uber drivers have cumulatively like 5-20 years driving experience in the city they're driving in.
Waymo has racked up waaaay more miles than the average single human ever gets, and unlike humans, all the Waymos benefit from improvements to the software.
> There will be failures and risks we haven't even imagined yet.
This is pointless fearmongering. Like, ketchup could cause cancer, but we have no meaningful evidence in that direction, so saying "ketchup has unknown risks we haven't imagined yet" is silly.
We know now that waymo is statistically safer than human drivers, I personally know that I haven't had a waymo driver play make me feel unsafe yet, but uber drivers often did, so you know, waymo seems to have some pretty nice improvements already.
I'll wait for actual evidence of these "unimaginable risks and failures" before I evaluate them. At this point, it would have to be a pretty bad failure to change the math though.
Yes. "Weird" people are somewhat rare opportunity to build certain social skills.
I enjoy the challenge of finding creative ways to guide the discussion and understand their headspace for a little while. I am not even trying to control the level of weirdness, but just keep them talking and comfortable.
Unfortunately, most of the time they're not even weird people and it was just a weird first impression. They vent for like 3 minutes and then it gets boring again.
I realize it is hard to do this, but please understand that other people have different perspectives on personal safety. For example, try and image how things might be different if you were a woman alone in an Uber with a driver who starts saying weird things.
There are second order effects though. Once Waymo kills the Uber driver/taxi jobs, what are the chances your Waymo is attacked by a roving band of jobless drivers? It's surely nonzero.
This seems a little silly. Did mobs of jobless taxi drivers attack the Uber drivers who took their jobs? No. No offense, but if you have a girlfriend, wife, or female friend, you might want to ask them about safety and security of ride sharing services. I suspect their answer will be an eye opener for you.
The one case where they hit a kid, they should have been driving slower to begin with. Their stopping distance exceeded their visibility in a school zone during pickup time. They might have done better than a bad human driver, and had good reflexes on the brakes, but a good human driver would have evaluated the conditions and not have been going that fast.
Riding a motorcycle or even a bicycle around Waymos feels surprisingly safer. You can reliably predict so many things about how it will behave and to an extent even its traffic calming effect on other cars.
"The one case where they hit a child, it was because the child jumped in front of the car. And they showed that they hit the child at a lower speed than a human would have because of the reaction time."
Was this the case that was featured on here a few months ago? Where they voluntarily "disclosed" it? I seem to remember noticing at the time that they never said this was the only time they hit a child/someone. Which made me wonder how representative this case was. I might be mis-remembering though.
The one major mistake I've seen is where they recently repainted a road from 2 lanes to 1 with some somewhat nonstandard markings indicating a merge, and the Waymo just drove through the merge as if the 2nd lane was still there
If only Waymo's were on the road I wouldn't worry about bike path dividers at all.
I sometimes pace them to act as a moving shield.
Nothing else comes close, not even eye contact and being waved on by a human. The other autonomous cars that have been introduced are at least just as scary to be around as people.
Nobody will. In fact, most car fatalities that are caused by humans involve zero criminal charges for anybody involved. In America, everybody from the courts to the media to society at large is primed to think of car accidents as normal. If you want to murder somebody in the middle of town in broad daylight, you can actually get away with it, as long as you do it with your car.
When a piece of construction equipment falls over and kills someone, the person or company who owns the equipment is liable. I image it would be the same thing here.
Sometimes that person then counter-sues the manufacturer of the equipment if they think it was faulty. I image that would also happen here if there were personal ownership of self driving cars.
I was recently in a Waymo in SF. It was turning right from a busy street onto a narrow street. Mid-turn, the car slammed on the brakes. I sat there for a couple seconds like “???” wondering if we'd hit something. Then a dude on an e-bike _flies_ past the car in the bike lane.
The car saw this dude coming from way down the street, flying, and was like “yeah, better stop.” Probably saved the biker from serious injury, or worse. I wouldn't have seen him if I was driving.
I've been observing their behavior in Atlanta for about the past year. Our roads here are fairly curvy, hilly, and lacking of expected markings, yet I haven't seen a driverless Waymo vehicle make a single odd move. One thing that brought a smile to my face was when I came to a 4-way stop at the same time as a Waymo vehicle at night & I flash my brights to tell the other vehicle to go ahead (southern hospitality) and I see the Waymo immediately begin its course through the intersection. I was so jolted that I began to tail it in order to pull up next to it to see if there was a human behind the wheel. Watching it drive down this slowly descending hilly road with intermittent speed humps and cars parked alongside the main right lane gave me a close up view of its slightly curving trajectory and braking behavior with regard to the humps. My thought on human or not was inconclusive until we reached a red light, and as I shot my eyes over and saw an empty driver seat, I smiled widely knowing that the software responds to brights flashed at 4-way stops (please don't tell me it doesn't and it just saw me indecisively not initiate at the stop). Thanks for reading
It definitely does not respond to flashing headlights in that manner. You’re observing its default behavior when at a 4 way stop with other vehicles not moving.
You're right I don't have inside information, but we've been interacting with them on the street for years in SF. Waymos don't wait for human subjective guidance to give them clearance to pass, as evidenced by tons of videos and IRL experience. As soon as they come to a required stop, and if a vehicle or other object's linear travel path does not intersect it, it will go. Flashing lights will not change this behavior. (Yes you're right there is a regulatory requirement to respond to safety officer guidance, but compliance is spotty as evidenced by a lot of videos of vehicles entering active crime zones, etc.)
Unlike the traffic cops directing traffic that would likely require special programming, "proceed if the other car flashes its lights at you" is completely the kind of thing that could just accidentally fall out of a neural network learning to imitate humans.
Hopefully if they ever go to Sri Lanka they get localised tuning because I was surprised to find out flashing your lights over there doesn't mean "go ahead", it means "if you don't get out of my way I will ram you"
And then there's trucks flashing an indicator to say it's safe to overtake if you're behind them. In the UK it's the nearside indicator, which makes sense: it's a bit like the truck is pulling over to let you pass. In Aotearo, it's often the off-side indicator, so you think the truck is going to pull out in front of you. I've never understood what the Aotearoa drivers are thinking there
It also doesn't make sense because "get out of my way or I will ram you" is the default state of operating a motor vehicle. Not the goal but the physical reality of it.
Waymo has published a ton about the imitation learning they've been using since 2018. They're not imitating random cars but their drivers who are paid to drive around and follow traffic laws.
It's not enough so they use heavy reinforcement learning etc. but it's still a huge foundation to build on.
Waymo immitates humans insofar as its neural net trained on avoiding collisions after millions of miles of video footage and LIDAR data on roads shared with humans causes it to immitate humans.
It's likely manually programmed not to (incorrectly) turn the wheel to the left while stopped and waiting for an opportunity to turn. If you get rear-ended, you'll end up in the lane of oncoming traffic. It's certainly programmed to use its turn signals to indicate when it is going to turn. But after driving around thousands of cars without turn signals on but with their wheels pointed left, it "knows" to predict that they're about to turn, and might immitate humans by anticipating that action and moving to pass the stopped car on the right.
Anecdote from 1000s of miles biking: I bike a lot in the Bay, for fun, exercise, commute, all of the above (I'm a friendly one, I promise!) and the comfort I feel when I see a Waymo alongside me or at a stop sign is immediately apparent. I have been hit 5-10x riding in NYC and SF (nothing serious, gratefully, mostly just people turning right not knowing/caring I was there), and the Waymo's awareness that I exist is immediately obvious and so different from a large percentage of human drivers. I hope the meaningful improvement in safety continues to convince people this should be a part of the future.
+1. I bike in the Bay and have been hit 3 times (1 broken bike and 1 broken helmet, but thankfully okay otherwise). I can’t wait for the day when I can be confident that all vehicles see me and I don’t have to worry about being the victim of a moment’s lapse in attention.
13X is way more impressive than it seems at first glance.
Let's take a simplistic model of accidents: that the average driver is at fault in an accident 50% of the time. So a perfect driver would only halve the number of accidents -- they only eliminate the accidents where they would otherwise have been at fault.
But Waymo's numbers are better than the "perfect" driver above. How is that possible? Because in most accidents the blame is not split 0%/100%. You can avoid a lot of accidents with defensive and safe driving.
13x compared to what ? The average driver is such a bullshit statistic - accidents are highly correlated with stuff like alcohol/drugs/lack of sleep/lack of experience/physical issues, then the other huge behavioral factor distraction and driving style, and on top of that car performance matters a lot too. I don't see any attempt to correct for that in their "human benchmark". Heck the least they could have done is compare to human taxi drivers which would be apples to apples. If it's 13x compared to that I'm sold for using it as a taxi service !
But individual driving - you can eliminate all those factors assuming you're a healthy, expericed driver with a new car. Nothing against self driving in principle but the failure cases I've seen look so bizarre - I'm way more comfortable with my limitations.
Is it a bullshit stat though? it's not like you or I can go to a different dimension where all drivers are healthy, fully awake, undistracted, sober, competent, etc.
You don't need to. You just need to be in good condition yourself and actually paying attention. Professional delivery drivers routinely achieve seemingly absurd mileages per incident.
More than 1/2 of roadway fatalities involve alcohol or drugs. An oversized fraction of fatalities are represented by young men under 24. 1/6 of all fatalities are motorcycles. 1/6 of all fatalities are pedestrians being struck by a vehicle.
I live in LA and Waymos are the only cars I don't have to play chicken with when crossing the street. Even the drivers that see you will just give you a "sorry, I'm in a rush" wave as they nearly run you over.
I did find the Waymos in SF disconcerting as they approached pedestrian crossings while I was waiting - with a human driver there's many different cues that they've seen you and noticed you - whether it be looking directly at you, or slowing down in preparation for you to cross.
I'm sure if I had just started walking across the crossing it would have reacted perfectly, but I wasn't willing - based on the lack of observable "I have noticed you" cues - to test that theory.
Make sure to never be in a hurry to get anywhere because you might then get stuck behind a fleet of them going exactly the speed limit, grid locking you in.
In Italy several cities lowered the maximum speed from 50 to 30 km/h.
There was a huge fight over it, car drivers in those cities were mad. Plenty of politicians opposed it.
One year later stats were super clear: streets got way safer and the number of fatal accidents dropped to near 0. Time to traverse cities didn't change much, as it was already limited mostly by traffic and lights.
Autonomous vehicles following proper signalling before lane changes can be safe at arbitrary speeds (see Autobahns working at all). Humans, we should limit passing speed to roughly ~5 mph delta between adjacent lanes and leave it at that.
Humans with adequate following distance in the entire lane can probably manage 10 mph delta. I routinely travel dozens of miles very safely at ~80 with the flow of traffic (including the cops), and been stressed out at 55 in the carpool lane through stop and go traffic in the right-hand lanes due to on ramps/offramps.
I think 75 is memorable and roughly in the region where the tradeoff between increased kinetic energy and decreased time to arrival per additional unit of velocity becomes untenable.
Because American drivers have normalized always driving 10 mph (16 km/h) over the speed limit.
Cops won't pull you over or write tickets if you're not at least 15 mph over, we basically don't have speed cameras, everyone's trying to win the rat race and dehumanizing other cars around them, and it's not considered morally wrong (by most) to break that specific part of the law.
So a single vehicle obeying the law will quickly get a long line of tailgaters and tailgaters of tailgaters trying to "push" the vehicle to go faster.
They can suck it, I'm not late or in a hurry, and my ancient truck, steel bumper, and class 5 receiver hitch will not be badly harmed by your plastic grille. I get better gas mileage and have a longer stopping distance when I drive the limit, and I don't care if others are honking or riding my ass because they think I should drive faster.
> my ancient truck, steel bumper, and class 5 receiver hitch will not be badly harmed by your plastic grille
I've been rear-ended in my truck, and the receiver punched a nice hole right through the radiator of the guy who hit me. Definitely fucked his car up way more than it did my truck ... except man, that is definitely one of the hardest impacts I've ever felt in my body. I now appreciate how hard the head rests really are, despite looking a little soft. I think I'd rather have crumpled crumple zones and a new truck next time.
On most US highways (i.e. multi-lane limited access roads), it's customary to leave a path in the left 'passing lane' for any traffic that wants/needs to go faster than you. If cars match speeds across lanes, it impedes faster traffic.
The speed limit itself is a separate convention and regulation. In some places you can be cited for obstructing traffic by going the speed limit in the passing lane if you are matching the speed of cars to your right, effectively blocking the road.
Sometimes I an appreciate wanting to cruise in the middle lane, because ADAS level 2 systems common on cars today is far more comfortable when it does not have to deal with regular merging traffic. But aside from that, I really don't like it when people camp in the middle lane because they tend to form a pretty tight line and manage to effectively turn a three-lane highway into two single-lane highways -- hard to get through from one side to the other.
> If cars match speeds across lanes, it impedes faster traffic.
I think this undersells it a little. It does not just impede faster traffic, when the lanes are pacing each other it makes navigating harder -- simply switching lanes is more difficult. The highway moves so much more efficiently with a small but steady difference in speed between each lane.
> How do you get “stuck” behind someone doing the speed limit?
"Only 46.5 percent of U.S. drivers consider going more than 15 miles per hour over the speed limit on the freeway to be "extremely" or "very" dangerous — with 40.6 percent openly admitting to doing it at least "a few times" in the last 30 days" [1].
We have a lot of freeway speed limits that are holdovers from the last oil crisis decades ago. Cars have gotten quieter, smoother, more capable, to the point where 55 mph is kind of hilariously slow. When the legal speed limit does not reflect what most drivers think is reasonable, then we can stamp our feet and insist that the law must be right, or we could redesign the road or adjust the speed limit to more closely reflect conventional wisdom.
I'd love to cycle more outdoors, but I'm always wary of the risks. How cool would it be if you could hire a waymo as a "team car" and have it follow you around? It could also carry extra equipment...and act as a ride home in case of emergencies.
If you ride conservatively (use lights at night, use good judgement at intersections) and stay away from buses and trucks, the exercise vastly outweighs any risk.
To turn the colourful phrasing against you, I too would like a death robot to protect me from all the murder monkeys having a death race around me if I was out biking on the road/death race track.
Wouldn't the better solution be to prevent those murder monkeys from doing the death races around where people would want to cycle? Some kind of grade-separated trail? More regulations on those murder monkey death races?
The grade separated trail sounds great in theory, but in practice it's surprisingly ineffective for some unknown reason. In Sweden we have a ton of those and for some reason bikers insist on participating in the death race with the other murder monkeys. Which is annoying because it can really slow the race down! But I guess once they have a good alternative, at least their death race participation is voluntary, and so I guess they're fair game like the rest of us
I've said before that Waymo is already vastly, incredibly safer than some of my older relatives who refuse to give up their keys. Ever been around The Villages in Florida and seen a man leaning forward behind the wheel to squint at what they're driving toward, with their wife shouting at them to turn left!, turn left!? That's just kind of tolerated in some places where the cops don't want to make waves with the wealthy older community.
A self-driving car never gets tired and sleepy after driving for many hours straight. A highway-bound Waymo would be safer than a few instances of distant past me who stayed on the road longer than I was safe to. They also never get drunk, and are safer than approximately 100% of impaired drivers.
I genuinely think we'll all be safer when lots of people collectively realize that someone other than themselves should be driving.
"For example, the current cities Waymo operates in do not have appreciable snow fall, and as a result neither the Waymo nor the human benchmark data include this type of inclement weather."
I'm happy to see this acknowledged, and hope it's a sign that they appreciate the difficulties of winter driving.
It’s wild to me how people are so fixated on this. Yes, obviously winter driving has challenges. And also obviously, the leading self-driving company has thought about that.
They’re preparing to launch and have already been testing in Chicago, detroit, Minneapolis, Denver, Philadelphia, Boston, NYC, and London. I think it’s safe to assume they’ve considered winter driving.
I was sold on Waymo when in San Francisco I saw it treat a human holding a Stop sign in a construction zone just like a human driver did.
For anyone who doesn't know this, in a construction zone if a human is holding a stop sign, it means stay stopped until they flip the sign and suggest you to move slowly. Waymo just handled this as a human would
A city with driverless cars is still a city dominated by cars. Bikes, trains, busses, and walking support the human soul and a city worth living in so much more than a robot car that requires car infrastructure and mentality.
I think the average is ballpark 1 per 100 million miles, so my guess is that someday they might show this, but they probably want to wait til they have more miles driven.
Also a motorcyclist. I've seen perhaps 6 or 7 Waymos whilst riding. One pulled out of curbside parking space right in front me.
It wasn't life or death or anything like that, but I was close enough that it was a real "dick move" and I had to get on the anchors a lot harder than I'd have liked. Not sure what sensor or whatever it was missing for that to happen, but I can assure you it did.
(I'm not suggesting my anecdotal evidence says anything particularly worthy around autonomous vehicle safety, just sharing a surprising incident)
The cynic in me says this is a moral hazard waiting to happen, perhaps we'll raise speed limits and reduce traffic regulations until the stats match the pre-robo-taxi days.
My question is: is safer than average human good enough?
When I drive I have the option to choose to be safe or not. When a computer drives I lose that option. So for 49% of the people, safer than the average human is less safe than before.
I think we need to reach "Safer than the safest 10% of humans".
Also these reports should be done by a government agency.
Yes, it's good enough. Because you cannot control who else is on the street around you. Having cars around you that are driving safer than the average is better than them driving average.
Even the most visible academic skeptic of Waymo (Phil Koopman) had to throw in the towel and admit that they've cleared every conceivable statistical hurdle to conclusively demonstrating that they are better than humans on injuries and airbag deployments. They have moved the goalposts to aesthetic arguments, for example: if it's so safe why does it sometimes do weird stuff? But to principled systems thinkers they have already shown what needed to be shown. It's safer.
Great, what's more concerning than crash rates though are impeding emergency vehicles, blocking roads, stalling when traffic lights are down, rolling past school buses with stop arms out, navigating in construction zones and lots of other issues that are left out of this report.
Blaming the kid here is absurd. The kid lives in a system where pedestrians are second class citizens in a world dictated by the auto-petro industrial complex. An industry that has co-opted unelected traffic "engineer" in the US and completely changed the way we live for the last 70 years and have made Americans fatter and less connected.
If the child lived in a neighborhood where cars went slower (it was a 25mph zone) he wouldn't have gotten hit in the first place. Praising Waymo here is like praising a priest for not molesting a child. Yes it's good that the waymo slowed down more than the average car, but really the whole system should be completely rethought. Instead, we're pouring billions into single occupancy vehicles, when we should've been pouring billions into high speed rail, subways, etc.
I'm hopeful that waymos converge on a more efficient design and improve cities in general. As it stands, they are a way for the rich to commute without having to exchange pleasantries with the underclass.
I was probably a bit too harsh on the OP. The OP was probably not blaming the kid. But if Waymo isn't being sued and the city isn't being sued, then society has collectively placed blame on the kid and their parents.
Blame isn't really zero sum is it. Like you can be criticized for leaving your laptop unatended and your doors unlocked, but it doesn't really reduce blame for the thief.
What you're talking about has nothing to do with Waymo at all though. It's ostensibly off topic here. You're talking about car culture in general.
Yes, I blame the parents or the adults that were supposed to supervise the child (but not the child). I teach my kids not to run into the street. I also watch them like a hawk near streets because kids are dumb.
I agree with you that we have too strong of a car culture. But we do. So until that changes, we need to teach our kids and adults to be vigilant.
But while we do that, I'd still rather have Waymos around than human drivers.
Imagine creating a brand that became renowned world wide and even created its own verb.
And then throwing all that away for the genius brand name of... "x". Brought to you from the same 50 year old that decided that having car models that spell S3XY is cool.
Nothing has been thrown away, evident by billions of dollars in the continued success of many flagship products that leave the competition in the dust.
Someone once said that this is because Waymos are novelty, and they still behave a bit weird, like being slow and undecisive. Which leads to humans being super-careful around them. So the Waymo safety record is actually not their own achievement.
I guess we'll have to wait to one of the two things to happen to really assess Waymo's performance:
1. They need to lose their markings and easily distinguishable features (like a big lidar on top), so they don't get any special treatment from other drivers.
2. They need to be majority of vehicles on the road.
That would make sense a while ago, but definitely not in SF for locals who have lived here a while. For me as a pedestrian/bicyclist/motorcyclist I actually feel safer around them than any other car.
Someone also once said that the Azores are the remains of Atlantis. I simply didn't put any credence in it.
While behavioral changes around a self-driving car are plausible; they're common enough now that, at least where I live in San Francisco, regular human drivers should be pretty well acclimated to them.
The other day a human driver in front of me was doing 30 km/h under the speed limit down the middle of two lanes.
On that same drive, another driver doing around 15 under clipped a roundabout on the way in and on the way out. Guess they couldn’t decide to turn the wheel fast enough.
I refuse to believe everybody is hammered all of the time, but I’m starting to wonder.
It is less than 10km round trip, in the ‘burbs. Driving with humans scares me anymore. Bring on the robots.
Ugh - either the commons is an unregulated 3D space or we actually tag and separate moving bodies regulated by size/weight, purpose, owner, occupant type, etc. I don't necessarily hate commercial vehicles utilizing the various rights-of-way but clearly there is a difference in momentum, agency, and general "value" between some human wandering around and a heavy robot.
recently (past couple of months) they've been much more aggressive in the ways that make a good driver a good driver - confident and assertive when they should be. for me this has anecdotally been a massive improvement
That info is pretty outdated: they were slow and indecisive in 2024, but now they behave pretty much like any top-decile human driver. I don’t think they get special treatment from other drivers either, I can’t read anyone else’s mind but I treat them like just another car and it seems like everyone else does as well.
one of the things that i noticed in a recent trip to austin is that the waymo vehicles were far more assertive and quick than the human drivers so maybe that has been addressed.
If someone drives badly they might go to jail if they hurt or kill someone. If a machine does it who pays? I want to see waymo and other CEO for decades for each mistake.
I feel like an important thing here is that we are very much not good at imprisoning bad drivers at all, even if they injure or kill multiple people. We rarely even take away their licenses!
The only type of car crash that consistently gets some level of enforcement is drunk driving, basically everything else is written off as an accident
The great deal: let's redesign our cities to be car free. Consider the economic boom that amount of renovation would produce. Consider the increased economic activity from happier and more productive people. Consider the increased space for nature, parks, real estate, development.
Cars are the worst thing to have been invented. Optimizing the personal automobile leads to optimizing for a horrible living experience in the city. Let us reconsider all of this. This is bad. We can do better. We must.
> Optimizing the personal automobile leads to optimizing for a horrible living experience in the city. Let us reconsider all of this. This is bad. We can do better. We must.
I agree with you insofar as I am always in support of making cities more friendly for pedestrians and cyclists, and like the idea of closing off parts of cities to cars.
But to not even acknowledge the benefits to society of a technology which can reduce serious traffic accidents by 90% just feels hopelessly extreme to me.
Americans will vehemently deny this, but you're absolutely right. Decades of car industry propaganda has convinced people that the ability to drive anywhere is true freedom, and they can't see that the freedom not to need a car (all the time) is better for everyone; cities are quieter, more comfortable, and less polluted with fewer cars (no, electric cars don't fix this). It leads to other absurdities, like US cities frequently having parking minimums for bars. That's insane!
There's also the classic problem of people wildly misinterpreting statements and getting mad about it. You can say "we should design cities not centred around cars" and people will hear "I'm going to make it illegal for you to own a car". Or my favourite exchange, "Let's improve public transit" followed by "but public transit is bad for me, I can't take it".
Yes, this debate comes up every time someone mentions the word "car" on the internet, and there are crazies on both sides. But I don't think it's fair to frame either side of the debate by what the crazies are saying. Or to assume that just because someone disagrees with you they have fallen victim to propaganda.
I think most Americans just like their cars and are reasonably happy with the status quo. They can be receptive to incremental improvements to public transportation, cycling, and pedestrian infrastucture, but they bristle at the idea of turning their city into a "car free city" (which is what the parent is suggesting) or being told they are wrong for liking their car.
But it objectively makes cities worse. People love visiting Europe in part because they don't do this to nearly the same extent (obviously this varies by country/city). People aren't entitled to not having their opinions be proven wrong, nor are they entitled to ignore negative externalities (pollution, noise, danger, unpleasant city centres, and so on).
> are reasonably happy with the status quo.
They're not, except for the having a car part. Road maintenance, especially in the suburbs, is hideously expensive and is falling further and further behind. Cars are the least efficient mode of transit, so traffic gets worse and worse. "Just one more lane" always makes it worse (induced demand), but that's the only solution being tried. The only way to make traffic better is to get significant numbers of people to switch to other modes, and you're simply not going to do that with "incremental improvements" because the status quo is so abysmal for anything other than a car. Cars themselves are horribly expensive and yet are required in most US cities; most are in effect paying a tax to car companies to participate in society.
“Waymo is using around four NVIDIA H100 GPUSs at a unit price of 10,000 dollars per vehicle to cover the necessary computing requirements. The number of sensors – five lidars, 29 cameras, 4 radars”
“Safer” == “Safer than all other human drivers in the same city.” By their own admission, this is not a straightforward comparison. If they could do the math for the same routes, times of day, and conditions … maybe I’d believe it. Otherwise, this data is trivial to cherrypick, and they have every reason to present it as well as possible.
I believe Waymos are pretty safe, and that’s a great thing. “Safer than humans (for selected rides inside this area)” is still very good, but it’s not at all “Safer than humans (period).”
In essentially all cases where a Waymo and a human-driven car have collided, the human driver has clearly been at fault. This seems definitive and not susceptible to cherry picking.
That could just be, and seems to be in some cases at least, because Waymo doesn't behave like a human would, and people gets tripped up.
I don't doubt Waymos are very safe, but I always irk at these comparisons. Majority of human accidents are due to gross negligence and/or driving under some influence or serious fatigue. A system incapable of alcohol etc. is better than that? Well that is a substantially lower bar than you can possibly imagine. Add to that that all systems have constraints on how and where they are able to go. Combined even Tesla can be made to look good.
Depending on the context and question it might still be the question to pose. But people often make the leap to assume that a typical Waymo is x better than a typical human driver which is an entirely different question entirely.
Waymo is for sure one of the (if not the only) good players out there though, gives me some hope.
> could just be, and seems to be in some cases at least, because Waymo doesn't behave like a human would, and people gets tripped up
Driving conventions vary wildly across states and even within them. And foreign drivers are a thing. A human who gets tripped up by a Waymo acting unusually will also get confused by someone getting used to no turns on right in Manhattan, driving on the right side of the road if coming in from the Commonwealth or adapting from California's protected left turners can turn into any lane, not just the leftmost. They'll also get confused by children and pets, who aren't bound by social custom, and deer, who aren't bound by physics.
A year or two ago, would have agreed with you. I used to think they were cherry picking as well.
But Waymos have driven so many miles by this point, if they are hiding some data that would tip the scales back towards human drivers I have yet to see it. If there is a way to slice the data that makes Waymo’s look less safe I would welcome the correction.
If Waymo truly has 80-90% fewer crashes in the conditions they drove in, then it still has policy implications for places like Phoenix that do have good conditions.
I've heard that Waymo relies on having very accurate map data for the areas where they operate, so perhaps they could perform worse than human drivers in areas where they don't have good map data.
But I also trust that the company wouldn't deploy them in those areas until the quality data they need is available. So perhaps "safer in the environments where they are actually deployed" would be more accurate, but that's also the only thing that matters.
Speculating about what would happen if they were used in ways they are neither intended to be used nor are actually used feels a little silly. Most machines can be unsafe if you use them in ways they're not intended to be used.
> By their own admission, this is not a straightforward comparison.
If they wanted to cherry pick, would they not omit that admission?
In any case, it seems plausible to me that the routes that Waymo drives are above average in human incidents, given that Waymo is probably overrepresented in high stress/traffic, inner city scenarios.
The comparison gets picked up as the headline; the admission does not. This is exploited quite often, e.g. in science reporting. I'm not saying this is what Waymo did - they don't seem to be bad actors - but absolutely, the pattern does occur.
If you were choosing between getting into a Waymo or a car driven by a human driver (where Waymo operates, for a route that Waymo would do), the data shows that the Waymo is safer.
No, it does not. For one thing, we don’t have access to all the data, just what’s being told us. For another, it at best shows that Waymo is safer than average. Safer than an attentive London Cabby? I bet not.
> For one thing, we don’t have access to all the data
In the US, we do have access to all the data [1]. They're required to report every incident with an injury or any amount of property damage, and it's all available for download as CSV.
> For another, it at best shows that Waymo is safer than average.
No, it shows that Waymo is 6 to 12x safer than average.
I know that anecdotal experience is definitionally just that, anecdotal. But I've had a handful of attentive London cabby experiences (and enough in-Waymo experiences) that give me conviction that Waymos are far safer than them. They're out there driving all day every day, it's obvious to me that a Waymo driver is going to be safer than even a professional.
One cabby pulled out of a t junction to end up alongside me on a motorbike – a Waymo would never do that.
I appreciate the skepticism. While I suspect motor vehicles that cannot be distracted would be safer than motor vehicles that can be, it shouldn't be claimed without real evidence.
If I were Google, I'd partner with some insurance carriers to compare the number of claim events normalized to the number of drivers on the road (approximated with Android data) in a city (same time of year, etc) before and after introducing Waymo. If claims per driver decreases, then I would be more inclined to support the claim that they're actually safer and that they don't just "seem safer"
Perhaps the comparison should only be to other taxis. Since I cannot buy a Waymo, it is not really relevant whether it is better than an average driver (including all the drunk ones, and the speed racers, etc).
There are other safety differences with human-driven vehicles… interpersonal violence does happen with taxis (e.g. drivers sexually harassing/assaulting passengers, passengers robbing drivers) - by definition those things cannot happen with a Waymo
accidents are not equally distributed across humans. more serious accidents will be caused by people who are habitually doing things that are unsafe but, for various reasons, most places lack effective ways to stop these people driving so they keep causing accidents.
the metric is not some nebulous aspect of skill but the bottom decile of human drivers causing accidents. it is not difficult to believe that an AI can drive better than this group, it is not a high bar, below the 10th percentile are people who should not be driving but cause most of the accidents.
I just watched a short that said some (all) Waymo drivers are not autonomous, but remote controlled by humans in the Phillipines.
I'm sure it's a combination of both since the latency would mean immediate reactions are impossible, but the presenter raised an interesting point, and that was that the remote drivers are not licensed to drive in the states that Waymo operated in, which would make it illegal.
except for the overwhelming evidence from video we already have that waymos are reacting than a human in the driver’s seat would, much less someone 200ms away.
Anecdotally, both from riding in them and walking/driving next to/around them, this feels obvious. They never get distracted. Sure, they sometimes make mistakes, but the mistakes are never "I didn't see that". They see better than humans in all cases (where they operate). They react faster than humans.
The one case where they hit a child, it was because the child jumped in front of the car. And they showed that they hit the child at a lower speed than a human would have because of the reaction time.
I would rather be in an area where only Waymo's are allowed than an area where they are banned.
Waymo saved my life in LA.
When I visited LA, I rode in a Waymo going the speed limit in the right lane on a very busy street. The Waymo approached an intersection where it had the right of way, when suddenly a car ignored its stop sign and drove into the road.
In less than a second, the Waymo moved into the left lane and kept going. I didn't even realize what was happening until after it was over.
Most human drivers would've t-boned the car at 50+ km/h. Maybe they would've braked and reduced the impact, which would be the right move. A human swerving probably would've overshot into oncoming traffic. Only a robot could've safely swerved into another lane and avoid the crash entirely.
Unfortunately, the Waymo only supported Spotify and did not work with my YouTube Music subscription, so I was listening to an advertisement at the time of my near-death experience. 4.5 stars overall.
> Unfortunately, the Waymo only supported Spotify and did not work with my YouTube Music subscription, so I was listening to an advertisement at the time of my near-death experience. 4.5 stars overall.
This detail sent me, it's crazy that we can pay $25 to have a life saving robot take us across the city yet Spotify is going to blast ads at us the whole time for the sake of making an extra $0.18 (yes that's the actual number) per hour of listening time.
That's around 44.64 (0.18831) per month, no wonder ads are preferrable to companies over subscriptions! That's actually a lot for people that listen to music all day every day at work.
Imagine your last thing in your mind being an ad about mongoDB.
I actually find those amusing because they just make me remember the 'web scale' meme.
It's kind of wild how you have so many ads targeted at devs in SF.
It’s like all the ads at airports clearly aimed at C-level execs.
And those in Brussels are all by American giants that want EU bureaucrats to know they take privacy seriously.
Thanks for that thought. Horrible.
> it's crazy that we can pay $25 to have a life saving robot take us across the city yet Spotify is going to blast ads at us the whole time for the sake of making an extra $0.18 (yes that's the actual number) per hour of listening time.
Oh, the self driving car business will get there, believe me. This is just the first iteration. Designed to get everybody on board with the idea.
I wonder if Waymo gets a cut. I also wonder if riding in a Waymo at the time signals that you're in a demographic that can afford a Waymo and thus get more expensive ads.
How ironic that an Alphabet company, Waymo, only works with a competitor streaming music service, Spotify, and not their own, YouTube Music. I guess that shows how separate they are.
We do support YouTube Music and actually supported that before Spotify. But we only do ad-supported on Spotify and iHeartRadio (also paid Spotify).
It looks like YouTube Music was only added in October? I took the ride in September.
https://support.google.com/waymo/answer/16623742?hl=en
How about some games to pass the time? Make some exclusives so I look forwards to a 20 minute cross-town drive!
I think it's also a privacy thing; you have to go into the Waymo app and “connect” your YouTube Music account (even though both have the same @gmail.com address), because otherwise the terms of service of one do not allow sharing data with the other without user consent. (Contrary to popular perception Google is very finicky about privacy, at least privacy as defined as conforming to the terms of service.)
In January YouTube music worked fine when I took Waymo in Menlo Park.
This was in September, so I'm happy to see the change!
That's good news, if I can't use the Youtube Music I've paid for in the Waymo then I'm not going to put up with Spotify Ads instead, better to sit in silence (or use my headphones and my own music)
Can you not steam arbitrary audio to it from your phone?
Nearly got T-boned in a Lyft in LA. I am lucky to still be alive as the driver was not aware and should not have been driving. Where available I've stopped using human driven rideshare.
Thankfully they've now shipped their own product, YouTube Music.
And Google Pay, imagine my surprise back in 2021 when I signed up for Waymo and realized I had to manually type in my credit card. No Google Pay??? C'mon y'all, you're Alphabet!
> Only a robot could've safely swerved into another lane and avoid the crash entirely.
Do you drive a lot? I feel like humans take evasive actions like this all the time.
Some do and some more successful than others.
Waymos have since added support for YouTube Music thankfully.
> Waymo saved my life... Unfortunately the Waymo only supported Spotify
I chuckled
> I was listening to an advertisement at the time of my near-death experience.
You'll probably never forget that advertisement, which is an exciting business opportunity for Waymo.
They could partner with Spotify and other media content partners so that the Waymo can generate an adrenaline-rush near crash experience when a premium advertiser's ad is playing. /s (hopefully)
This is one of those comments that made me laugh nervously. It's straight out of Ubik or another PKD novel, which probably means it's less than 5 years away from being real.
If there’s a torment nexus to be built they’ll build it.
Might be Orhan Pamuk or JG Ballard's mantle to be picked up
They support YouTube Music now, thank god
I hope you are misremembering. Swerving is most often the wrong choice, and I would be disappointed if Waymo were opting for that. By far the best option is to panic stop. Human or robot, physics is a harsh mistress and swerving is more likely to make you lose control and end up in a much more unforgiving wreck.
It wasn't possible to stop at the speed the Waymo was moving at.
The Waymo didn't have the stop sign, the other driver did, at a three way intersection.
The other driver decided to suddenly enter the intersection, when the Waymo was like 5-10 meters away. This was after having stopped at the stop sign.
Either they weren't looking or intentionally trying to cause an accident. Swerving prevented the Waymo from crashing at 40 km/h into the driver's door.
I assume waymo has a constant full picture of what's around, so swerve should be way safer for a machine than a human
For a human this advice is true. But what if a computer can near-instantly calculate a perfect swerve within the performance envelope of the car and driving conditions?
> swerving is more likely to make you lose control
Even if you're not a panicky human but a optimally regulated control system?
The optimally regulated system doesn't know the road conditions that well. When the road surface is more slippery, it has the most profound effect on lateral friction, way more than braking.
The Waymo driver can measure the speed and the acceleration of the offending car and calculate, within at most tens of ms, its range of likely future trajectories. And it can calculate its own likely trajectories under maximum braking. And it can track exactly where all obstacles are that would matter if it swerves. All at once. And it can execute that emergency lane change with the control input that is least likely to cause a loss of control and most likely to successfully avoid the other car. It even has processing power to spare to keep playing that Spotify ad!
Most often, but this seems to describe the rare exception.
This depends a huge amount on car, driver and situation. It was the right advice for a learner driver in the 90s with no stability control, no experience and no side airbags, because if you’re going to hit something, hitting it front on is the least risky way. I’m not convinced it’s the right advice for a competent driver in a modern vehicle.
It is still standard advice today, as far as I know. Tires are better, stability control is better, but all else being equal you are still much more dynamically stable and have a lot more friction with the road when the car is stopping in a straight line than when swerving.
Also, in the case of someone running a stop sign, it is far from a sure bet they are going to hit their brakes at all, so by swerving you increase the odds that you will still hit them, but now it will be while you have exhausted all your adhesion on lateral control. So now instead of a front end collision with all the benefits of airbags and crumple zones, you are at a significant risk of rolling the car or spinning off the road and hitting something immovable with a part of your car lacking crumple zones.
The common mistake is people swerve and brake, which is a terrible combination - you should accelerate through a sudden manoeuvre, as it maintains control through it, much as you should accelerate through corners in general.
Thanks! Now I know why the Waymo didn't slow down.
What about other drivers in that lane? It would have to be 100% sure that any other drivers near it would have enough time to react as well.
Absolutely. I was recently driving on a motorway in Portugal when a boulder (giant chunk of granite, 10+ tonnes) fell off the back of a truck - right in front of us, in a heavily laden (7 pax and luggage) car. Immediate massive cloud of dust, I checked my blind spot, veered across two lanes, and continued our journey, unscathed. I looked in the rear view, to see the car behind us jump on the brakes instead of evading. They caught the boulder.
Nobody killed, according to the news, but several taken to hospital in critical condition.
Oh, I say unscathed but our tyre exploded the next day, as apparently we caught a fragment, and again, that’s not a “slam on the brakes” moment, but rather “trundle to a stop on the shoulder and walk to the conveniently nearby tyre shop”.
I think you're over-playing how decisively a Waymo will move and under-playing how decent the average human is.
I've ridden in Waymos. They don't exactly slap on the blinker and move at the limit of traction like someone about to miss their exit. If cut off they absolutely will go full brake rather than perform any sort of spicy lane change or turn.
> If cut off they absolutely will go full brake rather than perform any sort of spicy lane change or turn.
Essentially, a meat driver was waiting at a stop sign to make a turn onto the main road. I was in a Waymo driving on the main road and did not have a stop sign.
When we were 10 meters away from the intersection, the meat driver suddenly started to enter the intersection. I have no idea why.
Full brake would've hit the other car in the driver's side door at 40 km/h.
> under-playing how decent the average human is.
I got to SMFC in CSGO which means I'm in the top 3% of players in clicking on heads within 500 ms of them appearing on my screen. I have never reacted as fast as that Waymo did.
If instant (<50ms) reaction would have lowered the speed only to 40km/h in 10m, Waymo was going too fast for the intersection IMO.
My experience is that for a human driver to react quickly in city driving conditions, style and prep are more important than reaction time: in the case you describe (entering an intersection with another car waiting on a stop sign perpendicular to your path), I'd have my foot hanging over the brake and off the gas pedal — this has helped me avoid hitting many other cars with inattentive/distracted/bad drivers, and even pedestrians running over the road or a red light on a crosswalk. When you are prepared and looking, you slam the brakes much faster!
To be fair, we are not provided with the sensors to swerve safely. If we had some sort of 360 constant recording in the car (on screen?) it would be safer for humans to swerve. Instead we have to move our head, which is cheaper but lacks info. That's why we now have rear cameras
We have rear cameras because people DONT move their head. And because regulations have made cars way taller than they need to be, meaning there is a big blind spot close to the ground
I mean, even in low cars you cannot see a small enough kid walking behind your car. That's why you back slowly. Back when I just got my driver license, there is a big lesson many drivers go through (in Italy) which is you back off a parking and there is an obstacle that's so low that cannot be see through the back window and it's small enough that cannot be seen through the mirror. You hit it and if you followed the "go slow part" you only damaged the paint.
So I'm not opposing the ideas of rear cameras, but I'm totally against tall cars, because you cannot see kids IN FRONT either now.
I think you're humongously overselling the average driver. I mean, the stats for waymo vs human drivers speak for themselves.
Depends on how you define "average driver": what if 95% of the crashes are caused by 5% of the drivers?
My reading of all the human crash stats has been that majority of them happen when human drivers are impaired (drunk, drugged or too tired): as this is something we could (in theory, at least) control, I'd like to see and compare with stats for non-impaired human drivers too.
Then, I'd like to see it compared to attentive, non-distracted drivers too (but we won't have crash data for this, as they would avoid most potential crashes).
Note that I am only talking things under every human driver's control, and not things like skill, reaction time, etc.
Also, modern cars (like Waymos) will have a much lower braking distance compared to "average": eg. my Volvo has 35m braking distance from 100km/h or 62mph compared to 50m (45% more) listed as average (excluding reaction distance) — so from 50km/h, it should be around 8m!
In Houston the stats suggest that every driver should get into a crash at least once every. But many ppl haven’t been an crash all their lives and more have been in multiple
> Waymo saved my life in LA. When I visited LA, I rode in a Waymo going the speed limit in the right lane on a very busy street. The Waymo approached an intersection where it had the right of way, when suddenly a car ignored its stop sign and drove into the road.
> ...
> Unfortunately, the Waymo only supported Spotify and did not work with my YouTube Music subscription, so I was listening to an advertisement at the time of my near-death experience. 4.5 stars overall.
What?! Is this a generated comment?
Here is a photo I took inside of the Waymo outside of an Erewhon. Going to Erewhon and experiencing the $20 Hailey Bieber smoothie was on my brother's bucket list and riding in a Waymo was on mine.
https://files.catbox.moe/jdjwy5.jpg
https://files.catbox.moe/mh4ivw.jpg
I have included EXIF data in an attempt to prove this really happened and I'm not an AI commenting bot.
What was the verdict on the smoothie?
He said it was overpriced but bought it again from a different Erewhon so I assumed he liked it (Canadian understatement).
There's apparently a quality gap between locations. The pre-Waymo one was from Erewhon Grove and was freshly blended. Erewhon Beverly Hills on Rodeo Drive premade a bunch of them and left them lying around for a while before selling.
My brother's theory is that Erewhon Grove customers are people who legitimately wanted a smoothie and Erewhon Beverly Hills customers just want photos with the smoothie since it was very popular on Instagram at the time.
Most surprising fact was despite being a licensed product, it was better than the best non-licensed smoothie (coconut cloud).
Licensing deals should make the product worse because the royalties cut into the product margin. The company cuts costs or doesn't take creative risks as a result. But somehow Erewhon resisted these pressures when designing the Hailey Bieber smoothie. We had a discussion about why that was the case but couldn't come up with an answer.
Pffft, like a bot couldn't fake metadata.
Waymo as a system has crossed the threshold where I trust them more than average driver, but all this hardware is relatively new, well maintained, and their software is closely tied to it.
I’m way less confident of self driving in the hands of the general public when differed maintenance often results in people and even companies driving with squealing breaks and balding tires etc.
I am also not looking forward to the system transitioning from "big experiment, burn money to make it good" to "established business unit, tweak it to death for incrementally more money / personal promotion." We're still in the honeymoon period and I very much expect to hate Waymo in 10 or 15 years when they reach a steady state.
What levers are there, really? Waymo has a monopoly and it seems like they will for a while, so they have a lot of power, but all I really see them doing is making it expensive. Anything that makes the experience worse takes away from their ability to take market share away from Uber/Lyft.
enshitification should be a new certainty along with death and taxes
That worries me.
Self-driving vehicles need aircraft-type maintenance. Yet there's nothing like the FAA to enforce a minimum equipment list, maintenance intervals, or signoffs by approved mechanics.
Is there a scratch or chip in the scanner dome? Are both the primary and backup steering actuators working? Is there any damage to the vehicle fender sensors? Is dispatch allowed with some redundant components not working? If so, for how long?
Here's the FAA's Minimum Equipment List for single-engine aircraft.[1] For each item, you can see if it has to be working to take off, and, if not, how long is allowed to fix it. There's nothing like that for self-driving land vehicles.
What's the fleet going to look like at 8 years of wear and tear?
[1] https://www.faa.gov/sites/faa.gov/files/MMEL_SE_Rev_2_Draft....
> Self-driving vehicles need aircraft-type maintenance.
That's a hyperbolic false equivalence.
Aircraft typically carry hundreds of people and can crash to the ground. As long as a self-driving car can detect when it is degraded, it can just stop with the blinkers on. Usually with 0 - 2 people inside.
The question is how broken can a car be when dispatched. What's the safe floor? See the other article today about a Tesla getting into an accident because of undetected sensor degradation.
It’s just death and taxes combined.
Waymo's software has crossed multiple generations of sensors and vehicles over almost two decades. It does not seem to be tightly coupled to a particular device.
Not tightly coupled in obvious ways, but as I understand it they aren’t putting it on pickup trucks, convertibles, or anything toeing a boat etc. Their vehicles don’t have aftermarket suspension systems dramatically changing handling characteristics, or turned one into a stretched limo etc.
Which means the software can safely assume the vehicle will behave within a relatively narrow operating range.
I suppose owners will be motivated to have the thing do the driving (and so seek defeat devices and such), but at least the software can have "do nothing" as a safety mode if it manages to detect that the vehicle is not configured as expected.
And maybe the software can be designed to be coupled to a vehicle dynamics model that can be updated.
I don't think the vehicle performance really matters in the typical case. They're using like 20% of what the vehicle "can" do. They're probably hedging against the long tail of variance on the road somehow. Kinda like how private people can tow whatever the f they want with their pickups but in a work setting you need to keep it fairly stupid proof.
The only thing an autonomous system should do with janky modified cars is drive them very slowly to the state police barracks for destruction.
Perhaps, but you can do a lot to a car while it remains street legal.
The new (as of now than a year ago) Waymo cars still had human safety drivers last I saw one (a month or two ago). I also don't see them taking customers. So they do seem to slow roll hardware rollouts.
A new version rolling out fast and starting to crash will likely kill the program altogether (like it did for some competitors).
10 years down the line, they won't have that risk.
Which to me is a really good, encouraging thing.
Overall I feel safer in a Waymo than a rideshare now and I only spent a few days being able to use Waymo...
The self-driving software could detect that the unmaintained car isn't responding correctly to the controls and refuse to drive.
We're not even a decade beyond some poorly conceived software crashing two otherwise functional aircraft into the ground and now it's going to save us all...
There is also a different kind of increased safety. There is no driver. No weird conversations about slaughtering goats, no sexual advances. No worrying that your driver is going to assault you or attempt to kidnap you. I know, it's all very far fetched, and Uber/Lyft drivers are almost always nice, courteous and professional, but I have experienced a few times when that hasn't been the case. With Waymo, it's not even an issue.
> There is also a different kind of increased safety. There is no driver. No weird conversations about slaughtering goats, no sexual advances. No worrying that your driver is going to assault you or attempt to kidnap you.
There are also new risks that weren't possible before. A software error can send you into oncoming traffic. Hackers can gain control of your vehicle either directly/remotely or by cleverly designed signage placed on the roadside. A disgruntled waymo contractor in the Philippines can remote drive you into a crowd of people. A flashing stoplight can leave you stranded at an intersection. The car may not see or react appropriately any number of uncommon hazards that human drivers would recognize and avoid. Only a relatively small number of these cars have been on the road, in limited conditions, and only for a small number years. There will be failures and risks we haven't even imagined yet.
Frequency matters.
One of these sets of risk is mostly theoretical (aside from the large scale stoplight outage), one of them is happening often enough that anyone who takes rideshare repeatedly will have a story.
If we limit ourselves to risks that have actually manifested, not hypothetical risks, I'd rather risk getting stuck at an intersection if there is a city wide power outage than deal with the weird conversations I've had on rideshares (not even counting the countless drivers who demonstrated that it is possible to drive a car without crashing for the duration of one rideshare ride without taking your eyes off the phone for more than a few seconds at a time).
> A disgruntled waymo contractor in the Philippines can remote drive you into a crowd of people.
They cannot. The remote drivers for Waymo offer "nudges" to the robot driver, but they cannot do full remote control.
They can effectively mark a dot in the middle of a crowd of people on their tablet and say "Your best course of action is to drive here", and the waymo very well might decide to try and follow that suggestion, but they cannot override Waymo's brakes nor coded-in "do not hit humans" mandate, and the waymo would stop before hitting anyone.
> Only a relatively small number of these cars have been on the road, in limited conditions, and only for a small number years.
The average uber driver has driven fewer miles on the road than Waymo's software, and hasn't seen all the conditions either. Most uber drivers have cumulatively like 5-20 years driving experience in the city they're driving in.
Waymo has racked up waaaay more miles than the average single human ever gets, and unlike humans, all the Waymos benefit from improvements to the software.
> There will be failures and risks we haven't even imagined yet.
This is pointless fearmongering. Like, ketchup could cause cancer, but we have no meaningful evidence in that direction, so saying "ketchup has unknown risks we haven't imagined yet" is silly.
We know now that waymo is statistically safer than human drivers, I personally know that I haven't had a waymo driver play make me feel unsafe yet, but uber drivers often did, so you know, waymo seems to have some pretty nice improvements already.
I'll wait for actual evidence of these "unimaginable risks and failures" before I evaluate them. At this point, it would have to be a pretty bad failure to change the math though.
This is like keeping your kids inside in case something bad happens to them.
If your kids never leave the house, something bad definitely happens to them, they stay kids.
Is there some benefit to talking to weird Uber drivers I've yet to discover that's comparable with 'going outside at all'?
Interaction with the common person is great. I wouldn't have know one could trim their toenails while driving otherwise.
Or that a taxi driver in Wuhan could answer his phone while shifting his manual transmission and smoking a cigarette.
There are probably better places to interact with other people than rideshares, like at a public establishment. There's significantly less risk
Yes. "Weird" people are somewhat rare opportunity to build certain social skills.
I enjoy the challenge of finding creative ways to guide the discussion and understand their headspace for a little while. I am not even trying to control the level of weirdness, but just keep them talking and comfortable.
Unfortunately, most of the time they're not even weird people and it was just a weird first impression. They vent for like 3 minutes and then it gets boring again.
I realize it is hard to do this, but please understand that other people have different perspectives on personal safety. For example, try and image how things might be different if you were a woman alone in an Uber with a driver who starts saying weird things.
I would rather say they develop crippling anxiety and agoraphobia. This is happening right now even to adults working from home.
There are second order effects though. Once Waymo kills the Uber driver/taxi jobs, what are the chances your Waymo is attacked by a roving band of jobless drivers? It's surely nonzero.
This seems a little silly. Did mobs of jobless taxi drivers attack the Uber drivers who took their jobs? No. No offense, but if you have a girlfriend, wife, or female friend, you might want to ask them about safety and security of ride sharing services. I suspect their answer will be an eye opener for you.
Using "second order effects" because big words sound cool without understanding the whole point of "second order"...
The one case where they hit a kid, they should have been driving slower to begin with. Their stopping distance exceeded their visibility in a school zone during pickup time. They might have done better than a bad human driver, and had good reflexes on the brakes, but a good human driver would have evaluated the conditions and not have been going that fast.
Riding a motorcycle or even a bicycle around Waymos feels surprisingly safer. You can reliably predict so many things about how it will behave and to an extent even its traffic calming effect on other cars.
Ya and they're the only ones I can count on being polite during rush hour
There was however a detail that explained the car was in a school area during pickup time and should have been on high alert exactly for that
> They see better than humans in all cases (where they operate). They react faster than humans.
You're absolutely right!
https://www.youtube.com/watch?v=yp0W5v8GOPc&t=520s
"The one case where they hit a child, it was because the child jumped in front of the car. And they showed that they hit the child at a lower speed than a human would have because of the reaction time."
Was this the case that was featured on here a few months ago? Where they voluntarily "disclosed" it? I seem to remember noticing at the time that they never said this was the only time they hit a child/someone. Which made me wonder how representative this case was. I might be mis-remembering though.
The one major mistake I've seen is where they recently repainted a road from 2 lanes to 1 with some somewhat nonstandard markings indicating a merge, and the Waymo just drove through the merge as if the 2nd lane was still there
I ride my bike and rollerblade around Austin.
If only Waymo's were on the road I wouldn't worry about bike path dividers at all.
I sometimes pace them to act as a moving shield.
Nothing else comes close, not even eye contact and being waved on by a human. The other autonomous cars that have been introduced are at least just as scary to be around as people.
sounds like you enjoy the predicability of Waymo vehicles. humans are unpredictable.
When it happens, who will go to jail?
Nobody will. In fact, most car fatalities that are caused by humans involve zero criminal charges for anybody involved. In America, everybody from the courts to the media to society at large is primed to think of car accidents as normal. If you want to murder somebody in the middle of town in broad daylight, you can actually get away with it, as long as you do it with your car.
At least with Waymo, it's much less frequent.
Crashes are not accidents. Language matters. People should go to jail for harming others
When a piece of construction equipment falls over and kills someone, the person or company who owns the equipment is liable. I image it would be the same thing here.
Sometimes that person then counter-sues the manufacturer of the equipment if they think it was faulty. I image that would also happen here if there were personal ownership of self driving cars.
Nobody. But you will be offered a voucher.
I was recently in a Waymo in SF. It was turning right from a busy street onto a narrow street. Mid-turn, the car slammed on the brakes. I sat there for a couple seconds like “???” wondering if we'd hit something. Then a dude on an e-bike _flies_ past the car in the bike lane.
The car saw this dude coming from way down the street, flying, and was like “yeah, better stop.” Probably saved the biker from serious injury, or worse. I wouldn't have seen him if I was driving.
I've been observing their behavior in Atlanta for about the past year. Our roads here are fairly curvy, hilly, and lacking of expected markings, yet I haven't seen a driverless Waymo vehicle make a single odd move. One thing that brought a smile to my face was when I came to a 4-way stop at the same time as a Waymo vehicle at night & I flash my brights to tell the other vehicle to go ahead (southern hospitality) and I see the Waymo immediately begin its course through the intersection. I was so jolted that I began to tail it in order to pull up next to it to see if there was a human behind the wheel. Watching it drive down this slowly descending hilly road with intermittent speed humps and cars parked alongside the main right lane gave me a close up view of its slightly curving trajectory and braking behavior with regard to the humps. My thought on human or not was inconclusive until we reached a red light, and as I shot my eyes over and saw an empty driver seat, I smiled widely knowing that the software responds to brights flashed at 4-way stops (please don't tell me it doesn't and it just saw me indecisively not initiate at the stop). Thanks for reading
It definitely does not respond to flashing headlights in that manner. You’re observing its default behavior when at a 4 way stop with other vehicles not moving.
How are you saying that so confidently? Waymos respond to traffic cops directing traffic manually
You're right I don't have inside information, but we've been interacting with them on the street for years in SF. Waymos don't wait for human subjective guidance to give them clearance to pass, as evidenced by tons of videos and IRL experience. As soon as they come to a required stop, and if a vehicle or other object's linear travel path does not intersect it, it will go. Flashing lights will not change this behavior. (Yes you're right there is a regulatory requirement to respond to safety officer guidance, but compliance is spotty as evidenced by a lot of videos of vehicles entering active crime zones, etc.)
Unlike the traffic cops directing traffic that would likely require special programming, "proceed if the other car flashes its lights at you" is completely the kind of thing that could just accidentally fall out of a neural network learning to imitate humans.
Hopefully if they ever go to Sri Lanka they get localised tuning because I was surprised to find out flashing your lights over there doesn't mean "go ahead", it means "if you don't get out of my way I will ram you"
And then there's trucks flashing an indicator to say it's safe to overtake if you're behind them. In the UK it's the nearside indicator, which makes sense: it's a bit like the truck is pulling over to let you pass. In Aotearo, it's often the off-side indicator, so you think the truck is going to pull out in front of you. I've never understood what the Aotearoa drivers are thinking there
This is true for India too though traffic there isn't known for its rules.
I hate the countries that do this because it doesn't even make sense as a signal. We already have a horn. They are wasting a channel!
It also doesn't make sense because "get out of my way or I will ram you" is the default state of operating a motor vehicle. Not the goal but the physical reality of it.
That's not how Waymo works, though. Waymo doesn't imitate humans. Waymo is trained to obey traffic laws and avoid collisions.
Waymo has published a ton about the imitation learning they've been using since 2018. They're not imitating random cars but their drivers who are paid to drive around and follow traffic laws.
It's not enough so they use heavy reinforcement learning etc. but it's still a huge foundation to build on.
Waymo immitates humans insofar as its neural net trained on avoiding collisions after millions of miles of video footage and LIDAR data on roads shared with humans causes it to immitate humans.
It's likely manually programmed not to (incorrectly) turn the wheel to the left while stopped and waiting for an opportunity to turn. If you get rear-ended, you'll end up in the lane of oncoming traffic. It's certainly programmed to use its turn signals to indicate when it is going to turn. But after driving around thousands of cars without turn signals on but with their wheels pointed left, it "knows" to predict that they're about to turn, and might immitate humans by anticipating that action and moving to pass the stopped car on the right.
> It's likely manually programmed not to (incorrectly) turn the wheel to the left while stopped and waiting for an opportunity to turn.
I'm both surprised and not surprised that people do this. You'll hit the divider.
A quaint, positive anecdotal comment?? On MY internet?!?!
How do you know? It’s trained on videos where it might see that happen often.
Anecdote from 1000s of miles biking: I bike a lot in the Bay, for fun, exercise, commute, all of the above (I'm a friendly one, I promise!) and the comfort I feel when I see a Waymo alongside me or at a stop sign is immediately apparent. I have been hit 5-10x riding in NYC and SF (nothing serious, gratefully, mostly just people turning right not knowing/caring I was there), and the Waymo's awareness that I exist is immediately obvious and so different from a large percentage of human drivers. I hope the meaningful improvement in safety continues to convince people this should be a part of the future.
+1. I bike in the Bay and have been hit 3 times (1 broken bike and 1 broken helmet, but thankfully okay otherwise). I can’t wait for the day when I can be confident that all vehicles see me and I don’t have to worry about being the victim of a moment’s lapse in attention.
13X is way more impressive than it seems at first glance.
Let's take a simplistic model of accidents: that the average driver is at fault in an accident 50% of the time. So a perfect driver would only halve the number of accidents -- they only eliminate the accidents where they would otherwise have been at fault.
But Waymo's numbers are better than the "perfect" driver above. How is that possible? Because in most accidents the blame is not split 0%/100%. You can avoid a lot of accidents with defensive and safe driving.
13x compared to what ? The average driver is such a bullshit statistic - accidents are highly correlated with stuff like alcohol/drugs/lack of sleep/lack of experience/physical issues, then the other huge behavioral factor distraction and driving style, and on top of that car performance matters a lot too. I don't see any attempt to correct for that in their "human benchmark". Heck the least they could have done is compare to human taxi drivers which would be apples to apples. If it's 13x compared to that I'm sold for using it as a taxi service !
But individual driving - you can eliminate all those factors assuming you're a healthy, expericed driver with a new car. Nothing against self driving in principle but the failure cases I've seen look so bizarre - I'm way more comfortable with my limitations.
Is it a bullshit stat though? it's not like you or I can go to a different dimension where all drivers are healthy, fully awake, undistracted, sober, competent, etc.
You don't need to. You just need to be in good condition yourself and actually paying attention. Professional delivery drivers routinely achieve seemingly absurd mileages per incident.
> You can avoid a lot of accidents
More than 1/2 of roadway fatalities involve alcohol or drugs. An oversized fraction of fatalities are represented by young men under 24. 1/6 of all fatalities are motorcycles. 1/6 of all fatalities are pedestrians being struck by a vehicle.
I live in LA and Waymos are the only cars I don't have to play chicken with when crossing the street. Even the drivers that see you will just give you a "sorry, I'm in a rush" wave as they nearly run you over.
I did find the Waymos in SF disconcerting as they approached pedestrian crossings while I was waiting - with a human driver there's many different cues that they've seen you and noticed you - whether it be looking directly at you, or slowing down in preparation for you to cross.
I'm sure if I had just started walking across the crossing it would have reacted perfectly, but I wasn't willing - based on the lack of observable "I have noticed you" cues - to test that theory.
The lights up on top will show a walk sign when it's waiting for a pedestrian to cross before proceeding.
Are these at pedestrian crossings or jaywalking?
Make sure to never be in a hurry to get anywhere because you might then get stuck behind a fleet of them going exactly the speed limit, grid locking you in.
Isn't the correct answer to this, lobbying for higher speed limits? Rather than chastizing obedience to current rules.
In Italy several cities lowered the maximum speed from 50 to 30 km/h.
There was a huge fight over it, car drivers in those cities were mad. Plenty of politicians opposed it.
One year later stats were super clear: streets got way safer and the number of fatal accidents dropped to near 0. Time to traverse cities didn't change much, as it was already limited mostly by traffic and lights.
I believe you, but do you have a citation?
Yes, agreed. Though speed limits higher than 75 are not something I will ever support.*
* Unless we're talking about removing a speed limit altogether and regulating unsafe driving using other criteria.
Autonomous vehicles following proper signalling before lane changes can be safe at arbitrary speeds (see Autobahns working at all). Humans, we should limit passing speed to roughly ~5 mph delta between adjacent lanes and leave it at that.
Humans with adequate following distance in the entire lane can probably manage 10 mph delta. I routinely travel dozens of miles very safely at ~80 with the flow of traffic (including the cops), and been stressed out at 55 in the carpool lane through stop and go traffic in the right-hand lanes due to on ramps/offramps.
What happens at 76mph?
Same thing that happens at 77mph :)
I think 75 is memorable and roughly in the region where the tradeoff between increased kinetic energy and decreased time to arrival per additional unit of velocity becomes untenable.
> the tradeoff between increased kinetic energy and decreased time to arrival per additional unit of velocity becomes untenable
Sounds like a warning page out of the back of a 94 Geo Metro owner's manual.
Is this something that I’m too European to understand? How do you get “stuck” behind someone doing the speed limit?
Because American drivers have normalized always driving 10 mph (16 km/h) over the speed limit.
Cops won't pull you over or write tickets if you're not at least 15 mph over, we basically don't have speed cameras, everyone's trying to win the rat race and dehumanizing other cars around them, and it's not considered morally wrong (by most) to break that specific part of the law.
So a single vehicle obeying the law will quickly get a long line of tailgaters and tailgaters of tailgaters trying to "push" the vehicle to go faster.
They can suck it, I'm not late or in a hurry, and my ancient truck, steel bumper, and class 5 receiver hitch will not be badly harmed by your plastic grille. I get better gas mileage and have a longer stopping distance when I drive the limit, and I don't care if others are honking or riding my ass because they think I should drive faster.
> my ancient truck, steel bumper, and class 5 receiver hitch will not be badly harmed by your plastic grille
I've been rear-ended in my truck, and the receiver punched a nice hole right through the radiator of the guy who hit me. Definitely fucked his car up way more than it did my truck ... except man, that is definitely one of the hardest impacts I've ever felt in my body. I now appreciate how hard the head rests really are, despite looking a little soft. I think I'd rather have crumpled crumple zones and a new truck next time.
On most US highways (i.e. multi-lane limited access roads), it's customary to leave a path in the left 'passing lane' for any traffic that wants/needs to go faster than you. If cars match speeds across lanes, it impedes faster traffic.
The speed limit itself is a separate convention and regulation. In some places you can be cited for obstructing traffic by going the speed limit in the passing lane if you are matching the speed of cars to your right, effectively blocking the road.
>it's customary to leave a path in the left 'passing lane' for any traffic that wants/needs to go faster than you
It's not just customary in many (most?) states, it's the law. People who sit in the left lane are the problem.
> it's customary to leave a path in the left 'passing lane' for any traffic that wants/needs to go faster than you.
A custom that (where I live) is becoming more honored in the breach than the observance. It makes driving very much more dangerous.
In Britain they have a sardonic nickname for people who do this: CLARAs. "Centre Lane Residency Association".
Sometimes I an appreciate wanting to cruise in the middle lane, because ADAS level 2 systems common on cars today is far more comfortable when it does not have to deal with regular merging traffic. But aside from that, I really don't like it when people camp in the middle lane because they tend to form a pretty tight line and manage to effectively turn a three-lane highway into two single-lane highways -- hard to get through from one side to the other.
I’m from England but I’ve only every heard “middle lane hoggers” for this
> If cars match speeds across lanes, it impedes faster traffic.
I think this undersells it a little. It does not just impede faster traffic, when the lanes are pacing each other it makes navigating harder -- simply switching lanes is more difficult. The highway moves so much more efficiently with a small but steady difference in speed between each lane.
> How do you get “stuck” behind someone doing the speed limit?
"Only 46.5 percent of U.S. drivers consider going more than 15 miles per hour over the speed limit on the freeway to be "extremely" or "very" dangerous — with 40.6 percent openly admitting to doing it at least "a few times" in the last 30 days" [1].
[1] https://usa.streetsblog.org/2023/11/30/why-so-many-u-s-drive...
We have a lot of freeway speed limits that are holdovers from the last oil crisis decades ago. Cars have gotten quieter, smoother, more capable, to the point where 55 mph is kind of hilariously slow. When the legal speed limit does not reflect what most drivers think is reasonable, then we can stamp our feet and insist that the law must be right, or we could redesign the road or adjust the speed limit to more closely reflect conventional wisdom.
You drive an ambulance? Or a fire truck?
Being forced to drive the speed limit isn't that big of a deal
I'd love to cycle more outdoors, but I'm always wary of the risks. How cool would it be if you could hire a waymo as a "team car" and have it follow you around? It could also carry extra equipment...and act as a ride home in case of emergencies.
that's a lot of energy consumption.
For a more ecologically conscious alternative, I recommend carrying a handful of sparkplugs.
If you ride conservatively (use lights at night, use good judgement at intersections) and stay away from buses and trucks, the exercise vastly outweighs any risk.
Personally I avoid riding at night entirely, and use at least a tail light during the day.
I get your idea, but it does rather sound like asking a 4,000lb death robot to follow you around closely and hope that it doesn’t screw up …
To turn the colourful phrasing against you, I too would like a death robot to protect me from all the murder monkeys having a death race around me if I was out biking on the road/death race track.
Wouldn't the better solution be to prevent those murder monkeys from doing the death races around where people would want to cycle? Some kind of grade-separated trail? More regulations on those murder monkey death races?
Crazy thought, I know.
The grade separated trail sounds great in theory, but in practice it's surprisingly ineffective for some unknown reason. In Sweden we have a ton of those and for some reason bikers insist on participating in the death race with the other murder monkeys. Which is annoying because it can really slow the race down! But I guess once they have a good alternative, at least their death race participation is voluntary, and so I guess they're fair game like the rest of us
Seperate is not equal. Grade separated bike routes are often not the best routes.
We don't even fully grade separate rail based murder machines from tire based murder machines.
Living in SF (and dad of a toddler), this seems like a no-brainer. I can't wait for fewer human drivers.
I've said before that Waymo is already vastly, incredibly safer than some of my older relatives who refuse to give up their keys. Ever been around The Villages in Florida and seen a man leaning forward behind the wheel to squint at what they're driving toward, with their wife shouting at them to turn left!, turn left!? That's just kind of tolerated in some places where the cops don't want to make waves with the wealthy older community.
A self-driving car never gets tired and sleepy after driving for many hours straight. A highway-bound Waymo would be safer than a few instances of distant past me who stayed on the road longer than I was safe to. They also never get drunk, and are safer than approximately 100% of impaired drivers.
I genuinely think we'll all be safer when lots of people collectively realize that someone other than themselves should be driving.
I enjoy driving (basically) wherever I want to though.
I don't want to have the freedom to go places determined by some faceless multinational, according to my subscription. Or via some "safety" regime.
"For example, the current cities Waymo operates in do not have appreciable snow fall, and as a result neither the Waymo nor the human benchmark data include this type of inclement weather."
I'm happy to see this acknowledged, and hope it's a sign that they appreciate the difficulties of winter driving.
It’s wild to me how people are so fixated on this. Yes, obviously winter driving has challenges. And also obviously, the leading self-driving company has thought about that.
They’re preparing to launch and have already been testing in Chicago, detroit, Minneapolis, Denver, Philadelphia, Boston, NYC, and London. I think it’s safe to assume they’ve considered winter driving.
Our sixth generation sensor suite is intended for winter driving: https://waymo.com/blog/2025/10/creating-an-all-weather-drive... for some nice visuals.
They're definitely aware and working on it: https://x.com/dmitri_dolgov/status/2028863520037867820
I was sold on Waymo when in San Francisco I saw it treat a human holding a Stop sign in a construction zone just like a human driver did.
For anyone who doesn't know this, in a construction zone if a human is holding a stop sign, it means stay stopped until they flip the sign and suggest you to move slowly. Waymo just handled this as a human would
A city with driverless cars is still a city dominated by cars. Bikes, trains, busses, and walking support the human soul and a city worth living in so much more than a robot car that requires car infrastructure and mentality.
Can they just report directly how many human lives they have saved since beginning operation? (of course, within some error bars)
Maybe that's too much of a statistical stretch.
But would be a good to-the-point number to have on hand for some waymo debates.
"yes they caused some disruption in an intersection in so-and-so scenario, but on the other hand they saved X number of human lives last year"
I think the average is ballpark 1 per 100 million miles, so my guess is that someday they might show this, but they probably want to wait til they have more miles driven.
I give up. I’m not even going to try to read “fade in” pages anymore. We had an article about that just the other day.
It’s horrible and makes reading harder.
I wanted to see this, but I give up.
as a motorcyclist I often feel more comfortable riding near waymos
at this point I trust that they have seen me, know that I'm there, and won't behave unpredictably
Also a motorcyclist. I've seen perhaps 6 or 7 Waymos whilst riding. One pulled out of curbside parking space right in front me.
It wasn't life or death or anything like that, but I was close enough that it was a real "dick move" and I had to get on the anchors a lot harder than I'd have liked. Not sure what sensor or whatever it was missing for that to happen, but I can assure you it did.
(I'm not suggesting my anecdotal evidence says anything particularly worthy around autonomous vehicle safety, just sharing a surprising incident)
The cynic in me says this is a moral hazard waiting to happen, perhaps we'll raise speed limits and reduce traffic regulations until the stats match the pre-robo-taxi days.
This page is old, but they just refreshed the data shows Waymo is 13x safer than human drivers (in the cities it operates in).
Is this an independent study?
They make the data available on that page for you to do your own research.
https://people.com/waymo-exec-reveals-company-uses-operators... might be relevant content as well
Pretty cool to see. But the UI of the visual animations has some weird re-re-rerender bug, at least on mobile safari.
When is waymo going to be available in the north east?
Hard to predict.
Waymo is trialing in several northeast cities. Search for "waymo trial boston" or "waymo trial nyc".
But beyond the technical issues, there are also political issues. Search for "new york bans waymo".
My question is: is safer than average human good enough?
When I drive I have the option to choose to be safe or not. When a computer drives I lose that option. So for 49% of the people, safer than the average human is less safe than before.
I think we need to reach "Safer than the safest 10% of humans".
Also these reports should be done by a government agency.
Yes, it's good enough. Because you cannot control who else is on the street around you. Having cars around you that are driving safer than the average is better than them driving average.
Even the most visible academic skeptic of Waymo (Phil Koopman) had to throw in the towel and admit that they've cleared every conceivable statistical hurdle to conclusively demonstrating that they are better than humans on injuries and airbag deployments. They have moved the goalposts to aesthetic arguments, for example: if it's so safe why does it sometimes do weird stuff? But to principled systems thinkers they have already shown what needed to be shown. It's safer.
The bar is low. I don’t want comparison with an alcoholic with multiple DUIs who still drives and crashes.
The benchmark should be the top decile of drivers.
The number of people who think they're top tier drivers never ceases to amaze me.
Yes statistics are full of drivers with more than one accidents at fault. 15% of the drivers are causing more than 70% of the accidents.
I don’t care about the average driver. I care about the median.
It is not a high bar to expect an autonomous system to be better than 90% of the American drivers.
Great, what's more concerning than crash rates though are impeding emergency vehicles, blocking roads, stalling when traffic lights are down, rolling past school buses with stop arms out, navigating in construction zones and lots of other issues that are left out of this report.
Relevant: https://youtu.be/cJER6WtOkjk?si=pfUMpqgCL8IHewKy&t=28
non-X source: https://waymo.com/blog/shorts/waymo-safety-impact-update-170...
Ah. That's the link I was looking for.
Blaming the kid here is absurd. The kid lives in a system where pedestrians are second class citizens in a world dictated by the auto-petro industrial complex. An industry that has co-opted unelected traffic "engineer" in the US and completely changed the way we live for the last 70 years and have made Americans fatter and less connected.
If the child lived in a neighborhood where cars went slower (it was a 25mph zone) he wouldn't have gotten hit in the first place. Praising Waymo here is like praising a priest for not molesting a child. Yes it's good that the waymo slowed down more than the average car, but really the whole system should be completely rethought. Instead, we're pouring billions into single occupancy vehicles, when we should've been pouring billions into high speed rail, subways, etc.
I'm hopeful that waymos converge on a more efficient design and improve cities in general. As it stands, they are a way for the rich to commute without having to exchange pleasantries with the underclass.
Good thing no one blamed the kid.
>> it was because the child jumped in front of the car.
Did you miss this sentence? How can you read it in any other way?
I was probably a bit too harsh on the OP. The OP was probably not blaming the kid. But if Waymo isn't being sued and the city isn't being sued, then society has collectively placed blame on the kid and their parents.
Blame isn't really zero sum is it. Like you can be criticized for leaving your laptop unatended and your doors unlocked, but it doesn't really reduce blame for the thief.
What you're talking about has nothing to do with Waymo at all though. It's ostensibly off topic here. You're talking about car culture in general.
Yes, I blame the parents or the adults that were supposed to supervise the child (but not the child). I teach my kids not to run into the street. I also watch them like a hawk near streets because kids are dumb.
I agree with you that we have too strong of a car culture. But we do. So until that changes, we need to teach our kids and adults to be vigilant.
But while we do that, I'd still rather have Waymos around than human drivers.
We detached this subthread from https://news.ycombinator.com/item?id=47445836 and marked it off topic.
why does HN still use links to twitter.com and not x.com?
Optimism. Someday the blue bird will be free.
Imagine creating a brand that became renowned world wide and even created its own verb.
And then throwing all that away for the genius brand name of... "x". Brought to you from the same 50 year old that decided that having car models that spell S3XY is cool.
Nothing has been thrown away, evident by billions of dollars in the continued success of many flagship products that leave the competition in the dust.
Presumably that was the submitter's choice.
More boring, too. Can't meet cool people if it's yet again just me left to throw a proverbial tennis ball against the proverbial wall.
Someone once said that this is because Waymos are novelty, and they still behave a bit weird, like being slow and undecisive. Which leads to humans being super-careful around them. So the Waymo safety record is actually not their own achievement.
I guess we'll have to wait to one of the two things to happen to really assess Waymo's performance:
1. They need to lose their markings and easily distinguishable features (like a big lidar on top), so they don't get any special treatment from other drivers.
2. They need to be majority of vehicles on the road.
That would make sense a while ago, but definitely not in SF for locals who have lived here a while. For me as a pedestrian/bicyclist/motorcyclist I actually feel safer around them than any other car.
"Someone once said ..."
Someone also once said that the Azores are the remains of Atlantis. I simply didn't put any credence in it.
While behavioral changes around a self-driving car are plausible; they're common enough now that, at least where I live in San Francisco, regular human drivers should be pretty well acclimated to them.
How slow and indecisive?
The other day a human driver in front of me was doing 30 km/h under the speed limit down the middle of two lanes.
On that same drive, another driver doing around 15 under clipped a roundabout on the way in and on the way out. Guess they couldn’t decide to turn the wheel fast enough.
I refuse to believe everybody is hammered all of the time, but I’m starting to wonder.
It is less than 10km round trip, in the ‘burbs. Driving with humans scares me anymore. Bring on the robots.
Ugh - either the commons is an unregulated 3D space or we actually tag and separate moving bodies regulated by size/weight, purpose, owner, occupant type, etc. I don't necessarily hate commercial vehicles utilizing the various rights-of-way but clearly there is a difference in momentum, agency, and general "value" between some human wandering around and a heavy robot.
I'm only a little weirded out when they're right next to me stopped at a light and that thang is spinnin and making note of me
recently (past couple of months) they've been much more aggressive in the ways that make a good driver a good driver - confident and assertive when they should be. for me this has anecdotally been a massive improvement
That info is pretty outdated: they were slow and indecisive in 2024, but now they behave pretty much like any top-decile human driver. I don’t think they get special treatment from other drivers either, I can’t read anyone else’s mind but I treat them like just another car and it seems like everyone else does as well.
one of the things that i noticed in a recent trip to austin is that the waymo vehicles were far more assertive and quick than the human drivers so maybe that has been addressed.
If someone drives badly they might go to jail if they hurt or kill someone. If a machine does it who pays? I want to see waymo and other CEO for decades for each mistake.
I feel like an important thing here is that we are very much not good at imprisoning bad drivers at all, even if they injure or kill multiple people. We rarely even take away their licenses!
The only type of car crash that consistently gets some level of enforcement is drunk driving, basically everything else is written off as an accident
Car centric design is ruining this country.
The great deal: let's redesign our cities to be car free. Consider the economic boom that amount of renovation would produce. Consider the increased economic activity from happier and more productive people. Consider the increased space for nature, parks, real estate, development.
Cars are the worst thing to have been invented. Optimizing the personal automobile leads to optimizing for a horrible living experience in the city. Let us reconsider all of this. This is bad. We can do better. We must.
> Optimizing the personal automobile leads to optimizing for a horrible living experience in the city. Let us reconsider all of this. This is bad. We can do better. We must.
I agree with you insofar as I am always in support of making cities more friendly for pedestrians and cyclists, and like the idea of closing off parts of cities to cars.
But to not even acknowledge the benefits to society of a technology which can reduce serious traffic accidents by 90% just feels hopelessly extreme to me.
> let's redesign our cities to be car free
Or, let's not spend trillions of dollars on a behavioural experiment and get pedestrian safety with now-proven kit.
Americans will vehemently deny this, but you're absolutely right. Decades of car industry propaganda has convinced people that the ability to drive anywhere is true freedom, and they can't see that the freedom not to need a car (all the time) is better for everyone; cities are quieter, more comfortable, and less polluted with fewer cars (no, electric cars don't fix this). It leads to other absurdities, like US cities frequently having parking minimums for bars. That's insane!
There's also the classic problem of people wildly misinterpreting statements and getting mad about it. You can say "we should design cities not centred around cars" and people will hear "I'm going to make it illegal for you to own a car". Or my favourite exchange, "Let's improve public transit" followed by "but public transit is bad for me, I can't take it".
Yes, this debate comes up every time someone mentions the word "car" on the internet, and there are crazies on both sides. But I don't think it's fair to frame either side of the debate by what the crazies are saying. Or to assume that just because someone disagrees with you they have fallen victim to propaganda.
I think most Americans just like their cars and are reasonably happy with the status quo. They can be receptive to incremental improvements to public transportation, cycling, and pedestrian infrastucture, but they bristle at the idea of turning their city into a "car free city" (which is what the parent is suggesting) or being told they are wrong for liking their car.
But it objectively makes cities worse. People love visiting Europe in part because they don't do this to nearly the same extent (obviously this varies by country/city). People aren't entitled to not having their opinions be proven wrong, nor are they entitled to ignore negative externalities (pollution, noise, danger, unpleasant city centres, and so on).
> are reasonably happy with the status quo.
They're not, except for the having a car part. Road maintenance, especially in the suburbs, is hideously expensive and is falling further and further behind. Cars are the least efficient mode of transit, so traffic gets worse and worse. "Just one more lane" always makes it worse (induced demand), but that's the only solution being tried. The only way to make traffic better is to get significant numbers of people to switch to other modes, and you're simply not going to do that with "incremental improvements" because the status quo is so abysmal for anything other than a car. Cars themselves are horribly expensive and yet are required in most US cities; most are in effect paying a tax to car companies to participate in society.
> being told they are wrong for liking their car.
Who said this?
I more or less agree with you. I just thought your earlier straw man argument was not made in good faith.
Now do Surveillance Impact: https://thelastdriverlicenseholder.com/2024/10/27/waymos-5-6...
“Waymo is using around four NVIDIA H100 GPUSs at a unit price of 10,000 dollars per vehicle to cover the necessary computing requirements. The number of sensors – five lidars, 29 cameras, 4 radars”
“Safer” == “Safer than all other human drivers in the same city.” By their own admission, this is not a straightforward comparison. If they could do the math for the same routes, times of day, and conditions … maybe I’d believe it. Otherwise, this data is trivial to cherrypick, and they have every reason to present it as well as possible.
I believe Waymos are pretty safe, and that’s a great thing. “Safer than humans (for selected rides inside this area)” is still very good, but it’s not at all “Safer than humans (period).”
In essentially all cases where a Waymo and a human-driven car have collided, the human driver has clearly been at fault. This seems definitive and not susceptible to cherry picking.
That could just be, and seems to be in some cases at least, because Waymo doesn't behave like a human would, and people gets tripped up.
I don't doubt Waymos are very safe, but I always irk at these comparisons. Majority of human accidents are due to gross negligence and/or driving under some influence or serious fatigue. A system incapable of alcohol etc. is better than that? Well that is a substantially lower bar than you can possibly imagine. Add to that that all systems have constraints on how and where they are able to go. Combined even Tesla can be made to look good.
Depending on the context and question it might still be the question to pose. But people often make the leap to assume that a typical Waymo is x better than a typical human driver which is an entirely different question entirely.
Waymo is for sure one of the (if not the only) good players out there though, gives me some hope.
> could just be, and seems to be in some cases at least, because Waymo doesn't behave like a human would, and people gets tripped up
Driving conventions vary wildly across states and even within them. And foreign drivers are a thing. A human who gets tripped up by a Waymo acting unusually will also get confused by someone getting used to no turns on right in Manhattan, driving on the right side of the road if coming in from the Commonwealth or adapting from California's protected left turners can turn into any lane, not just the leftmost. They'll also get confused by children and pets, who aren't bound by social custom, and deer, who aren't bound by physics.
A year or two ago, would have agreed with you. I used to think they were cherry picking as well.
But Waymos have driven so many miles by this point, if they are hiding some data that would tip the scales back towards human drivers I have yet to see it. If there is a way to slice the data that makes Waymo’s look less safe I would welcome the correction.
If Waymo truly has 80-90% fewer crashes in the conditions they drove in, then it still has policy implications for places like Phoenix that do have good conditions.
I've heard that Waymo relies on having very accurate map data for the areas where they operate, so perhaps they could perform worse than human drivers in areas where they don't have good map data.
But I also trust that the company wouldn't deploy them in those areas until the quality data they need is available. So perhaps "safer in the environments where they are actually deployed" would be more accurate, but that's also the only thing that matters.
Speculating about what would happen if they were used in ways they are neither intended to be used nor are actually used feels a little silly. Most machines can be unsafe if you use them in ways they're not intended to be used.
> By their own admission, this is not a straightforward comparison.
If they wanted to cherry pick, would they not omit that admission?
In any case, it seems plausible to me that the routes that Waymo drives are above average in human incidents, given that Waymo is probably overrepresented in high stress/traffic, inner city scenarios.
The comparison gets picked up as the headline; the admission does not. This is exploited quite often, e.g. in science reporting. I'm not saying this is what Waymo did - they don't seem to be bad actors - but absolutely, the pattern does occur.
If you were choosing between getting into a Waymo or a car driven by a human driver (where Waymo operates, for a route that Waymo would do), the data shows that the Waymo is safer.
No, it does not. For one thing, we don’t have access to all the data, just what’s being told us. For another, it at best shows that Waymo is safer than average. Safer than an attentive London Cabby? I bet not.
> For one thing, we don’t have access to all the data
In the US, we do have access to all the data [1]. They're required to report every incident with an injury or any amount of property damage, and it's all available for download as CSV.
> For another, it at best shows that Waymo is safer than average.
No, it shows that Waymo is 6 to 12x safer than average.
[1] https://www.nhtsa.gov/laws-regulations/standing-general-orde...
I know that anecdotal experience is definitionally just that, anecdotal. But I've had a handful of attentive London cabby experiences (and enough in-Waymo experiences) that give me conviction that Waymos are far safer than them. They're out there driving all day every day, it's obvious to me that a Waymo driver is going to be safer than even a professional.
One cabby pulled out of a t junction to end up alongside me on a motorbike – a Waymo would never do that.
I appreciate the skepticism. While I suspect motor vehicles that cannot be distracted would be safer than motor vehicles that can be, it shouldn't be claimed without real evidence.
If I were Google, I'd partner with some insurance carriers to compare the number of claim events normalized to the number of drivers on the road (approximated with Android data) in a city (same time of year, etc) before and after introducing Waymo. If claims per driver decreases, then I would be more inclined to support the claim that they're actually safer and that they don't just "seem safer"
Perhaps the comparison should only be to other taxis. Since I cannot buy a Waymo, it is not really relevant whether it is better than an average driver (including all the drunk ones, and the speed racers, etc).
There are other safety differences with human-driven vehicles… interpersonal violence does happen with taxis (e.g. drivers sexually harassing/assaulting passengers, passengers robbing drivers) - by definition those things cannot happen with a Waymo
accidents are not equally distributed across humans. more serious accidents will be caused by people who are habitually doing things that are unsafe but, for various reasons, most places lack effective ways to stop these people driving so they keep causing accidents.
the metric is not some nebulous aspect of skill but the bottom decile of human drivers causing accidents. it is not difficult to believe that an AI can drive better than this group, it is not a high bar, below the 10th percentile are people who should not be driving but cause most of the accidents.
https://waymo.com/safety/impact/#methodology
Worth reviewing the methodology, rather than making stuff up.
I just watched a short that said some (all) Waymo drivers are not autonomous, but remote controlled by humans in the Phillipines.
I'm sure it's a combination of both since the latency would mean immediate reactions are impossible, but the presenter raised an interesting point, and that was that the remote drivers are not licensed to drive in the states that Waymo operated in, which would make it illegal.
Waymo uses remote contractors to hint the Waymo driver when it can't figure out a path forward. They're not being remote piloted.
They are however, very cagy about how often this is necessary.
They don't necessarily even hint them. I think the car mainly asks them yes/no questions and they respond.
except for the overwhelming evidence from video we already have that waymos are reacting than a human in the driver’s seat would, much less someone 200ms away.