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.
Google was famously really resistant to ads at first. They wanted to do a subscription service of some kind, but honestly ads just brought in so much more revenue even back then that it was a nigh-inevitable decision. It produces a crazy amount of economy.
> 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)
Yeah I have done similar evasive maneuvers a bunch of times. Also people run stop signs constantly, a competent defensively driving human may have just not started driving forward yet when they saw the other car driving towards the stop sign with some speed. I’m not sure of the exact timing in the story but I’ve waited at a stop sign when I saw another car driving towards the intersection many, many times, and a small percent of the time they don’t stop.
Which isn’t to say that the average driver wouldn’t have hit it, it’s just not obviously superhuman.
In this case, the other car didn't run the stop sign.
It waited at the stop sign like it was making a turn, then suddenly entered the intersection when the Waymo was 5-10 meters away, despite not having the right of way.
Maybe they were trying to commit suicide-by-Waymo?
I can almost guarantee, given the way no traffic laws are enforced anymore in LA and so many cars are breaking them that cutting off a waymo car in dangerous ways, because you know it has the attention and reflexes to yeild, will be a new thing.
If it becomes a new thing, it would be fairly trivial for Waymo to respond by sending footage of incidents involving reckless driving to local authorities.
To prevent over-reporting, they could even make a system which logs number plates and only reports it if (for example) the same car is involved in incidents 3+ times in a week.
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!
Its is corporate fiefdom. Everybody trying to one up other executives to show impact instead of working together towards a unified goal. The bigger the company the more we see this phenomena. Nobody gets promotion if you just used existing internal service.
Waymo supporting music services at all is stupid to me. They should just let you bluetooth your phone and play your own music. I don't want any of those services. I want my own music.
There is a large group of people (maybe even the majority?) who, as soon as they get in a car, MUST immediately turn on the radio or some kind of extra noise source. Is this some kind of a Pavlovian reflex?
I'm always amazed by this, as my car is one of the few places where I have actual control over my environment (unlike on public transport, or at my workplace, or even in my home - neighbours can be noisy...). We are living in a sea of unwanted noise, bombarded by constant ads and "music", so it is nice to have a place of "quiet".
There are even people who listen to music at home! They even buy expensive speakers just for this purpose =) I listen to music pretty much all the time except when I talk to other people and sleep.
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.
> 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.
> when suddenly a car ignored its stop sign and drove into the road.
I vaguely recall reading at some point that this is something human drivers learn to do around robot cars because the robots are so timid. Is that still the case, was it never the case, or has it stopped being the case?
If it's still the case, one could argue that if you were not in a robot, the situation would never have occurred in the first place! (On the other hand, if you were both in robots, maybe it also wouldn't have...)
Great reaction time from the Waymo. But a 50 km/h side impact in a 5 star crash-rated Jaguar with curtain airbags is not "saved my life" territory, more
like an insurance claim. The fact that you instinctively framed it as a near-death experience is Waymo most impressive engineering achievement.
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!
>Waymo was going too fast for the intersection IMO.
This is literally impossible without slowing to single digit speeds for every intersections. At some point you just have to rely on the other traffic honoring signaling and signage or having some desire for self preservation.
Waymo uses new cars which probably have 100km/h (62mph) to zero stopping distance around 36m (120ft) — that's what my 2020 car quotes and tests at. As stopping distance grows quadratically, from 50km/h it would have stopped in 8m. Two lane street is usually at least 8m wide.
The claim was that after braking for 10m, it was still going at 40km/h. It'd take another 6-7m to come to a full stop. If it was a full 18m stopping distance (half the one from 100km/h), that'd mean a bit over 70km/h, so over 60km/h anyway for 16-17m.
I do not know of any country where there are intersections you can go through at 60+ km/h legally.
This does not mean that Waymo in question was going too fast, but something is off in the claim (maybe it did not react on time and really brake for 10m; maybe the collision speed was not a full 40km/h; or maybe it was going too fast...).
>I do not know of any country where there are intersections you can go through at 60+ km/h legally.
The US is dotted with "real highways" (i.e. designed as such, not a former main street that's seen a bunch of upgrades) with 50+mph speed limits and low traffic streets that tee into them with nothing more than a stop. And this isn't some middle america thing that can be dismissed as backwards flyover states. The rich coastal states have them too. Divided medians and T-junction type are fairly common. 2-way stops and cross type junctions less so in my experience but in more rural areas they're more common.
>This does not mean that Waymo in question was going too fast, but something is off in the claim (maybe it did not react on time and really brake for 10m; maybe the collision speed was not a full 40km/h; or maybe it was going too fast...).
Seems like someone pulled out from a residential road onto a main road with no f's to give and the waymo went around and OP is messing up the numbers a little. No matter how fast you're going it always feels faster from the passenger seat.
Pulling out randomly, I see it all the time. I beat the computer by a) anticipating; and b) assuming other drivers are idiots who don't see me. I don't have to calculate trajectories and whatnot, people aren't computers, and they can do some things better than a computer can, especially a solely-reactive one.
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!
It certainly matters: "average driver" does not exist, and 95% of the drivers beat the average.
So a claim how autonomous driving system beats the average would only tell us that it beats 5% of the human drivers.
Now, the way stats are massaged here is not even about "drivers", but miles driven, and this language is even worse. We'd need to make sure we are looking at human-driven miles in the same area, same roads, with similar cars.
>To be fair, if 5% of drivers cause 95% of crashes then the average driver is still terrible.
>The median one might be better, but does it even matter? The average driver is still wreaking havoc.
Yes it matters. To be acceptable this technology needs to be at least in the same ballpark as a median-ish person on a median-ish day. Not some nonexistent average that is pulled down by the 1/X people who are drunk and the 1/Y who are from Socal and driving in Maine in a blizzard.
The fact that you basically never hear of "average non criminal driver" or "median law abiding driver" and that there is no real attempt at even standardizing a concept of normal drivers not engaged in bad behavior just reeks.
It's like the door is intentionally being left open for the same slight of hand as when people peddle some policy goal having to do with school shootings and back it up with statistics that are mostly normal crime. Or they are peddling some devious tax that will screw a whole lot of people, and they justify it with an average that's dragged way up by a few oddballs, or dragged way down by a bunch of zeros. Seems like the safety crowd and and self-interested industry are setting up to play off each other in a "recyclable plastic" sort of way.
Second off, what are you talking about that the "average driver is wreaking havoc"? The average driver is filing a collision claim every 15-20yr depending on who's numbers you believe. While I don't know the distance between average and median, either is a fairly high bar that Waymo and friends have to meet.
But the point is if we get all the median and better drivers off the road and replace them with autonomous vehicles, yet keep the worst 5% on the roads too, we are potentially worse off.
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
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.
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!
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?
There are many cases where swerving will avoid an accident that braking cannot and cars unexpectedly pulling out from the side are often among these. It’s not a majority, but it’s not at all rare.
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.
Wunibald Kamm begs to differ. For his circle, it doesn't matter if the additional force that causes the friction to be insufficient is forward or backwards on top of the side force. In critical situations either use your friction for lateral xor longitudinal action, never both at the same time. Brake hard, but then sail through the curve. You want that vector to move along the circle and never leave it. As that is very difficult for an untrained driver, better switch hard between both modes.
True if we all drove unicycles - but in the real world, tyre wear is uneven, brake wear is uneven, loading is uneven, the surface is uneven, and those differential forces are what modern ABS seeks to control.
The key difference between braking and accelerating is that in the former case, independent, potentially differentially worn brakes, apply force unevenly, making the chance of a loss of traction on one or more wheels higher. With acceleration, that force is applied through a differential, meaning it will be far more likely to be appropriately distributed.
If you want to decelerate while swerving it can be done, but it should be done through engine braking - and the tricky bit there is matching revs as you drop the clutch back in, otherwise you have too much retarding force and overcome the coefficient of friction, resulting in a skid.
Easier for those of us who grew up with double de-clutching and no synchromeshes, but when you’re in a critical situation, it’s still an awful lot easier to apply acceleration.
In theory, it depends. In practice, slamming on your brakes is the correct call 99% of the time. To a large extent that is because of the "competent driver" part. I'd expect 80%+ of drivers to consider themselves just that, whereas the truth is of course the opposite.
So, the correct advice is to say "brake, don't swerve", so that drivers internalize that their first thought and reaction in any emergency should be to brake. Teach them to actually brake—fully press on the pedal—while you're at it.
A slightly more nuanced advice would be "brake first, swerve as needed as a follow-up".
But I would never in good conscience be able to give anyone advice to swerve instead of braking.
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”.
> 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.
I think they have rotating specials. These are pre-made, cheaper, and smaller. When you have a membership this is the one you get for free (once a month).
They used to have a “Dr. Paul's Raw Animal-Based Smoothie” that I looooved, but unfortunately they stopped making it. Still sad about that one :(
Mixing saving one's life with rating for a drink in the same comment makes feel very weird. Not funny, TBH, but actually it reads like a millions of generated bad review comments on amazon, expedia etc.
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.
> Aircraft typically carry hundreds of people and can crash to the ground.
Cars are more numerous and could spontaneously either plow into pedestrians, or rear-end someone, causing chain damage and, quite often, a spillage of toxic chemicals (e.g., a cistern carrying acid/fuel/pesticide).
Plus, you have a problem of hostile actors having easier access to cars compared to planes.
The way I see it, self-driving cars have the potential to deliver us from the burden of ownership altogether--maintenance, insurance, liability, parking, and all the rest. This hinges on availability, quality of service, pricing, and a rather large shift in the culture around cars and driving but I have hope that we can get there with time.
It's not obvious that will exist in the near future, anyway. Waymo aren't planning on selling their cars, and the economics and liability structure of self-driving strongly bias towards just running a taxi service.
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 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 mean, I do talk to them and I do have this skill, but it's a skill that I only ever seem to employ in talking to Uber drivers, so I'm not sure it's of any great benefit.
If anything the fact that most of them are immigrants puts the conversation on easy mode if you're a native speaker. They're doing twice the mental work you are so it's easy to orchestrate the conversation.
Not really transferrable to native-speaking workers. Like speaking to a barista is very different. Speaking to a construction worker different again.
That's interesting. Cultural differences and language barriers aren't what I would consider weird.
I was thinking of those people who have wild stories and/or mountains of narcissism to overcome. They have a fascinating worldview like an artist would if they had those ambitions.
They get bonus points in my book the more genuinely unhinged and confused they seem to be. They got that way by questioning things into absurdity and I don't mind listening.
Well there's a virtuous cycle for immigrants whereby if you integrate, you improve the language, you get a better job, and you integrate more, thus often ironing out any weirdness wrt to the host culture. Uber driver is pretty dead-end and isolated. You work constant hours but all your interactions tend to be very surface level.
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.
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 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.
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.
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
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.
Well with an autonomous system, you can actually audit what the machine was "thinking" and how it came to think that way, so you actually can distinguish between an accident and criminal negligence better than a he-said-she-said situation with humans
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'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.
I think we're not interpreting the original comment in the same way.
In most places, I think, when driving on the highway, flashing your lights when behind someone means basically 'I would like to overtake you'. Same here in the UK. But that's very specific to that context. You would never see a 'go ahead' context that would mean 'get out of my way', right?
But what the original comment means is there are some countries where you'd think it was 'go ahead' but it really means 'get out of the way'. Like if you're both on a main road, and you are signaling to turn into a side road, the opposing car flashes the lights and that means you can turn. I assume the same in Serbia.
But in some places that can actually mean don't turn, I'm going first. Which I think is what the parent is describing.
You are right that I did not read it the same way, and yes, the unwritten rules are matching in Serbia. FWIW, I've mostly switched to using left-turn signal to indicate "I'd like to overtake", which I've seen done on EU highways.
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.
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 would largely agree with that assessment, yeah. Dangerous place to bike, too. I've even seen pedestrians get clobbered by bikers because they stepped into the bike lane not realizing a bike was barreling toward them at 20mph+. This is part of why Waymo and Uber warn you when the dropoff is next to a bike lane.
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.
That combined with the parent's post is, perhaps counterintuitively, somewhat concerning.
The proper technique for yielding to pedestrians wishing to cross is to start slowing down early, as if you were planning to stop before the crossing. That sends a clear signal to the pedestrian they're good to start crossing. Then you're free to speed back up. This is very comfortable for the pedestrian and the vehicle never needs to stop, so the slowdown is minimal.
That Waymos apparently don't act this way and seem to need to send an explicit signal to pedestrians sounds concerning to me, even if its ultimately safe.
Waymo does slow down as it approaches stop signs (usually where crosswalks are) and it will slow down if there is a pedestrian entering the roadway (crosswalk or not) since it doesn't want to crash into them.
The explicit signal of a driver noticing you (eye contact) is replaced by the signal above the vehicle. Are you not equally concerned that pedestrians have to get an explicit signal from drivers who are legally required to yield or stop??
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.
I think this ignores the argument the high speed limit people make which basically boils down to "sure some people will die or get injured but its worth it because driving faster is fun"
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.
I've actually been involved in the commissioning of an FMVSS 202a headrest strength tester.
A lot of science and work goes into the construction of those headrests - if it was less firm, you'd get a concussion from the rotational forces in the whiplash or just break your neck, more firm and you'll get a concussion from the linear impact. It's not at all arbitrary, there's a reason they are exactly as firm as they are.
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.
> (b) An operator of a vehicle on a roadway moving more slowly than the normal speed of other vehicles at the time and place under the existing conditions shall drive in the right-hand lane available for vehicles, or as close as practicable to the
right-hand curb or edge of the roadway, unless the operator is:
> (1) passing another vehicle; or
> (2) preparing for a left turn at an intersection or into a private road or driveway.
Note this law specifically mentions "normal speed of other vehicles at the time and place" and doesn't directly mention speed limits. So by the text of this law, if you're driving the speed limit and hanging out in the left lane while the normal speed at that time is like 10 over you're technically breaking this law.
We have specific signage for highways where this is supposed to be the law.
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.
> Results: A 5-mph increase in the maximum state speed limit was associated with an 8.5% increase in fatality rates on interstates/freeways and a 2.8% increase on other roads. In total during the 25-year study period, there were an estimated 36,760 more traffic fatalities than would have been expected if maximum speed limits had not increased—13,638 on interstates/freeways and 23,122 on other roads.
That doesn't make sense to me. If you want to change lanes, and worst case scenario you're right next to someone, go 2mph slower for 20 seconds and they'll be shifted by 60 feet. I'm sure you can plan your lane changes on a freeway 20 seconds in advance.
It is the dynamics. When lanes are pacing each other the gaps all tighten up. So sure, you can slow down a bit to find the gap behind you, except that gap is not big enough to fit in. So you turn on your signal and wait for someone polite enough to let you in, meanwhile the guy behind you is riding you like a pony because you are no longer keeping up with traffic.
When traffic isn't balled up so tight, you can plan for a lane change in advance and accomplish it without having to slow down traffic. Everything flows better.
> 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.
> 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
Most Americans ignore speed limits. This stems from it being socially and legally problematic to permanently revoke our driver’s licenses. We should raise a lot of limits. But many others are fine and still sped through.
Where do you guys have 55 mph where it's not appropriate? A LLM is telling me that 55 mph is your random urban interstate. The equivalent where I live is 80 km/h on motorways inside city/town limits, which is 50 mph, and it feels very appropriate, cars make a ton of noise and you don't want the full motorway speed in the city. And that's basically the fastest you'll ever go in a city, when it's an actual motorway (often elevated, ramps, sound barriers). Other roads (even big and important) are ~43 mph and the general urban speed limit is ~31 mph.
> A LLM is telling me that 55 mph is your random urban interstate.
Yes, this sounds about right. In the metro area, 55 mph on a limited access interstate freeway. Arterial surface streets typically 40-45 mph, lower level surface streets commonly 25 mph and sometimes 20 mph depending on locality.
In the US, in particular out west where I live, 'urban' does not have the same meaning as it does somewhere much more dense, so it amounts to 55 mph in many places you might regard as rural.
On the stretch of motorway that I frequent in Italy, the speed limit is mostly 130 km/h, but the majority of people drive at about 100 to 110 km/h (including me).
But there are also people who drive in the left lane, who will tailgate you at 1 or 2 meters because you're doing 130 km/h. These people are idiots, but you get these sorts of people everywhere.
On American freeways, you don't have a choice, every lane is doing about 10 mph over the limit (or in LA way under) and it is disruptive or dangerous not to. These freeways tend to be running at full capacity so it actually makes sense since it improves capacity.
Road capacity does not increase with speed above 50 km/h on urban roads or 70 km/h on highways. Following distance scales with speed, so more speed can actually mean fewer cars per unit of time.
In theory, braking distance scales quadratically with speed. In practice, people leave less room on highways, because they rely on others driving predictably, but spacing still increases faster than linear.
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.
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.
Do you have a solution for preventing humans from driving while under the influence of alcohol or drugs? Or even preventing humans from driving while texting?
Most accidents are caused by drunk drivers, distracted drivers, jackass drivers and others, who together still constitute a minority, and there are plenty single-vehicle crashes. 13x improvement over average may still constitute diminished safety for a diligent sober driver.
13x better than an average that includes drunk, fatigued and distracted drivers would not be very impressive on a closed course.
On a real world course where the only way to achieve those kind of numbers is to avoid getting hit by those drunk, fatigued and distracted drivers? Very impressive.
Yes, but Waymo also has to drive on the road with those drivers, and these stats include crashes that are their fault. Diligent drivers get hit by drunk/distracted drivers all the time.
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.
Well, if we were talking about forcing people to stop driving and transition to current waymos it's plausible that diligent sober drivers would be facing greater risk. Would that be acceptable to improve average statistics?
If you're arguing to the regulator that "look these cars are not as bad as humans overall" I guess that's fine, but if you're trying to sell me on using self-driving by comparing to something that's completely unrelated to my use-cases that's just BS.
Unless their message is "if you're drunk turn on self driving" which I could get behind, I sincerely doubt current self driving is better than humans - simply because of the data they chose to compare to. If they were better than professional taxi drivers I'm sure they would tout that data widely.
This perspective makes the statistic even more impressive. Every day driving they interact with probably thousands of people who are impaired in some way and manage to drive safely far more often
No not really - because it doesn't compare to other comparable drivers. You're assuming that human drivers can't drive defensively.
Like a realistic comparison would be comparing it to taxi drivers/ride share drivers - then you'd see the risk vs using a cab/uber/whatever. Not a drunk 18 year old with a clunker with no brakes.
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.
This is very dependent on the environment. I would not be so confident saying this about the USA in particular.
If you can stick to at least half-decent dedicated cycling infrastructure separated from cars, sure. If riding alongside traffic, I wouldn't be surprised that, depending on the exact route, the gains wouldn't outweigh the risks.
Generally speaking, operating a bike safely is considerably more difficult than a car and the margins of error are tiny.
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
Sorry, yes, I didn't necessarily mean full grade separation, I was mostly thinking about just separated bike lanes in general. Having some kind of real buffer between the cyclists and the murder monkeys.
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.
Ooh, that's a worthy challenge. Of course, I can imagine getting enough data on all of those cities and deciding to launch everywhere else but not Boston "because your roads are garbage and you all drive like you're impaired 24/7" :-)
They’re in Denver now (not available to the public yet) for this exact reason. Unfortunately we’ve had our least snowy winter in recorded history this year - so they only got a few days of real snow.
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
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.
OK right, so they're at 170 million now, maybe ~1 life saved. I see it might be a bit early to be making claims about that.
Also in comparison the cases of disruption/blocking intersections/emergency-services do seem significant - seems plausible in the right circumstances you could lose lives.
Seems they need to 10x their miles before they can start making confident claims about lives saved.
Wow. >40% of their accidents have a speed delta-V of less than 1mph. They should just sell this thing as a kit to pop onto your existing car, with a major discount on car insurance. I’d buy it in a heartbeat if they can get it under $30k.
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)
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.
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.
You seem to be thinking this is compared to median driver, and not average one: if most accidents are caused by small number of bad drivers, your average will be bad, but even median might still be good.
And this does not even compare the drivers, but simply miles driven.
So I think that 80% of human drivers would likely be safer than Waymo unless they are driving under the influence or extremely tired or distracted.
Note that "13x safer" already implies being in the top 10%, though.
From a purely technical perspective, I would expect a Waymo to react to unexpected external stimuli much, much faster than a human. It gives the Waymo options that are unavailable to a human's slower reaction time.
there was a startup I saw once here on HN selling driverless kits that could be added to general cars.
if Waymo can do the same - then partner with insurance companies - they can easily be a $1 trillion dollar company even if they just reduce accidents by 5X.
The insurance market or cost thereof is their market.
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.
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.
DoorDash has these little cute robots doing delivery. I often seen them followed by a person on a e-bike. This has been going on for more than a year. My recent Lyft driver said one reason is because the Waymo's ignore the other robot and kill them and the bike ensures they don't.
I don’t use Waymo’s because they are really slow. At least for me. I get they’re safer. What’s new is everyone I know is complaining how expensive they are. Why are they so expensive? Is it because they are trying to make a profit on the total cost of ownership of a vehicle? They’re electric so it’s not gas. Uber and Lyft definitely outsource the cost odd maintenance onto the contractor.
I wish the executives at Waymo would go to prison until their cars stopped breaking traffic laws left and right. I get they're safer then humans. I've enjoyed my rides. But they're doing one of this "fuck the laws while we cash in things" because they know they can get away with it.
I guess more realistically, I wish the government would step in demand footage because the cars have records of every law they broke.
So how long until this is enshittified and how will it go? Will people ask for cars to be faster so they will take riskier decisions? Will a cheaper company with a worse training model capture the most market share? Pilot products are usually great, but that rarely lasts into their profitable period.
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.
No, I think those are solidly less concerning than crash and injury rates.
Like, could you imagine yourself seeing a self driving car that's perfect in all the ways like that, no dumb quirks, but it crashes twice as often as a human, and saying "well it got the important stuff right"
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
Maybe at your country. Not at mine.
Driving irresponsibly might have consequences. If you race, advance red light on don't stop for pedestrians, if there is not enough evidence you couldn't do anything about it, you might go to jail.
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.
... no? That children, animals etc. acts differently everyone knows. But what about a self-driving car that looks the same as every other car?
Anyway, it was Waymos own findings when they started out. They got into more accidents, none of which where their own fault, than expected and realized that they had to make it behave more like a human to not confuse human drivers.
> what about a self-driving car that looks the same as every other car?
My Aussie friends, when they visit America, don't put up a sign in their window saying they're Aussie and will occasionally try to turn left at a red.
> They got into more accidents, none of which where their own fault, than expected and realized that they had to make it behave more like a human to not confuse human drivers
Sure. I'm still filing this in the nothingburger file. If anything, it screams that we have a lot of people on the road who should not be.
I think the other way to think about it is that Waymo is probably in the 99th percentile for not being distracted and 99th percentile for reaction time, always, just from a pure sensor and computation standpoint.
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.
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
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"
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.
Google was famously really resistant to ads at first. They wanted to do a subscription service of some kind, but honestly ads just brought in so much more revenue even back then that it was a nigh-inevitable decision. It produces a crazy amount of economy.
I still loathe ads though.
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.
They know their market. :)
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.
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.)
Couldn't it just be a Bluetooth audio device? That way you could play anything you want, be it from YouTube, Spotify or your own music collection.
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
Maybe you could recreate it now, but with better music?
How about some games to pass the time? Make some exclusives so I look forwards to a 20 minute cross-town drive!
why does the car need games? Just use your phone/tablet/laptop
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?
> 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.
Yeah I have done similar evasive maneuvers a bunch of times. Also people run stop signs constantly, a competent defensively driving human may have just not started driving forward yet when they saw the other car driving towards the stop sign with some speed. I’m not sure of the exact timing in the story but I’ve waited at a stop sign when I saw another car driving towards the intersection many, many times, and a small percent of the time they don’t stop.
Which isn’t to say that the average driver wouldn’t have hit it, it’s just not obviously superhuman.
In this case, the other car didn't run the stop sign.
It waited at the stop sign like it was making a turn, then suddenly entered the intersection when the Waymo was 5-10 meters away, despite not having the right of way.
Maybe they were trying to commit suicide-by-Waymo?
Some do and some more successful than others.
I can almost guarantee, given the way no traffic laws are enforced anymore in LA and so many cars are breaking them that cutting off a waymo car in dangerous ways, because you know it has the attention and reflexes to yeild, will be a new thing.
If it becomes a new thing, it would be fairly trivial for Waymo to respond by sending footage of incidents involving reckless driving to local authorities.
To prevent over-reporting, they could even make a system which logs number plates and only reports it if (for example) the same car is involved in incidents 3+ times in a week.
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!
Google only hires the best of the best.
Its is corporate fiefdom. Everybody trying to one up other executives to show impact instead of working together towards a unified goal. The bigger the company the more we see this phenomena. Nobody gets promotion if you just used existing internal service.
Waymo supporting music services at all is stupid to me. They should just let you bluetooth your phone and play your own music. I don't want any of those services. I want my own music.
Why do you even want music?
There is a large group of people (maybe even the majority?) who, as soon as they get in a car, MUST immediately turn on the radio or some kind of extra noise source. Is this some kind of a Pavlovian reflex?
I'm always amazed by this, as my car is one of the few places where I have actual control over my environment (unlike on public transport, or at my workplace, or even in my home - neighbours can be noisy...). We are living in a sea of unwanted noise, bombarded by constant ads and "music", so it is nice to have a place of "quiet".
There are even people who listen to music at home! They even buy expensive speakers just for this purpose =) I listen to music pretty much all the time except when I talk to other people and sleep.
It maybe a surprise to you, but many people actually enjoy 'music' and don't find it to be just noise.
If I don't have my music then I have to listen to my own thoughts, and nobody wants that.
In seriousness, music is one of the small joys of life. Like a morning coffee or the smell of winter. It makes living a little bit more bearable.
But think of the shareholders and their tasty dividends
If the CEO isn't juicing market cap 110% of the time, the board will prosecute them and they will go to JAIL!
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.
> Waymo saved my life... Unfortunately the Waymo only supported Spotify
I chuckled
Waymos have since added support for YouTube Music thankfully.
> 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
> when suddenly a car ignored its stop sign and drove into the road.
I vaguely recall reading at some point that this is something human drivers learn to do around robot cars because the robots are so timid. Is that still the case, was it never the case, or has it stopped being the case?
If it's still the case, one could argue that if you were not in a robot, the situation would never have occurred in the first place! (On the other hand, if you were both in robots, maybe it also wouldn't have...)
Great reaction time from the Waymo. But a 50 km/h side impact in a 5 star crash-rated Jaguar with curtain airbags is not "saved my life" territory, more like an insurance claim. The fact that you instinctively framed it as a near-death experience is Waymo most impressive engineering achievement.
They support YouTube Music now, thank god
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!
>Waymo was going too fast for the intersection IMO.
This is literally impossible without slowing to single digit speeds for every intersections. At some point you just have to rely on the other traffic honoring signaling and signage or having some desire for self preservation.
Waymo uses new cars which probably have 100km/h (62mph) to zero stopping distance around 36m (120ft) — that's what my 2020 car quotes and tests at. As stopping distance grows quadratically, from 50km/h it would have stopped in 8m. Two lane street is usually at least 8m wide.
The claim was that after braking for 10m, it was still going at 40km/h. It'd take another 6-7m to come to a full stop. If it was a full 18m stopping distance (half the one from 100km/h), that'd mean a bit over 70km/h, so over 60km/h anyway for 16-17m.
I do not know of any country where there are intersections you can go through at 60+ km/h legally.
This does not mean that Waymo in question was going too fast, but something is off in the claim (maybe it did not react on time and really brake for 10m; maybe the collision speed was not a full 40km/h; or maybe it was going too fast...).
>I do not know of any country where there are intersections you can go through at 60+ km/h legally.
The US is dotted with "real highways" (i.e. designed as such, not a former main street that's seen a bunch of upgrades) with 50+mph speed limits and low traffic streets that tee into them with nothing more than a stop. And this isn't some middle america thing that can be dismissed as backwards flyover states. The rich coastal states have them too. Divided medians and T-junction type are fairly common. 2-way stops and cross type junctions less so in my experience but in more rural areas they're more common.
>This does not mean that Waymo in question was going too fast, but something is off in the claim (maybe it did not react on time and really brake for 10m; maybe the collision speed was not a full 40km/h; or maybe it was going too fast...).
Seems like someone pulled out from a residential road onto a main road with no f's to give and the waymo went around and OP is messing up the numbers a little. No matter how fast you're going it always feels faster from the passenger seat.
Pulling out randomly, I see it all the time. I beat the computer by a) anticipating; and b) assuming other drivers are idiots who don't see me. I don't have to calculate trajectories and whatnot, people aren't computers, and they can do some things better than a computer can, especially a solely-reactive one.
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!
To be fair, if 5% of drivers cause 95% of crashes then the average driver is still terrible.
The median one might be better, but does it even matter? The average driver is still wreaking havoc.
It certainly matters: "average driver" does not exist, and 95% of the drivers beat the average.
So a claim how autonomous driving system beats the average would only tell us that it beats 5% of the human drivers.
Now, the way stats are massaged here is not even about "drivers", but miles driven, and this language is even worse. We'd need to make sure we are looking at human-driven miles in the same area, same roads, with similar cars.
>To be fair, if 5% of drivers cause 95% of crashes then the average driver is still terrible.
>The median one might be better, but does it even matter? The average driver is still wreaking havoc.
Yes it matters. To be acceptable this technology needs to be at least in the same ballpark as a median-ish person on a median-ish day. Not some nonexistent average that is pulled down by the 1/X people who are drunk and the 1/Y who are from Socal and driving in Maine in a blizzard.
The fact that you basically never hear of "average non criminal driver" or "median law abiding driver" and that there is no real attempt at even standardizing a concept of normal drivers not engaged in bad behavior just reeks.
It's like the door is intentionally being left open for the same slight of hand as when people peddle some policy goal having to do with school shootings and back it up with statistics that are mostly normal crime. Or they are peddling some devious tax that will screw a whole lot of people, and they justify it with an average that's dragged way up by a few oddballs, or dragged way down by a bunch of zeros. Seems like the safety crowd and and self-interested industry are setting up to play off each other in a "recyclable plastic" sort of way.
Second off, what are you talking about that the "average driver is wreaking havoc"? The average driver is filing a collision claim every 15-20yr depending on who's numbers you believe. While I don't know the distance between average and median, either is a fairly high bar that Waymo and friends have to meet.
I was just making a point re: average versus median. We’re not very good at getting bad drivers off the road
But the point is if we get all the median and better drivers off the road and replace them with autonomous vehicles, yet keep the worst 5% on the roads too, we are potentially worse off.
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
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
> 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!
> The optimally regulated system doesn't know the road conditions that well.
I'd like to introduce you to what autonomous cars were already able to do in 2018: https://www.youtube.com/watch?v=khX0UCqcR3M
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?
There are many cases where swerving will avoid an accident that braking cannot and cars unexpectedly pulling out from the side are often among these. It’s not a majority, but it’s not at all rare.
Anecdotally it does appear that Waymo's do default to braking in many situations where a normal human would choose to swerve.
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.
Wunibald Kamm begs to differ. For his circle, it doesn't matter if the additional force that causes the friction to be insufficient is forward or backwards on top of the side force. In critical situations either use your friction for lateral xor longitudinal action, never both at the same time. Brake hard, but then sail through the curve. You want that vector to move along the circle and never leave it. As that is very difficult for an untrained driver, better switch hard between both modes.
True if we all drove unicycles - but in the real world, tyre wear is uneven, brake wear is uneven, loading is uneven, the surface is uneven, and those differential forces are what modern ABS seeks to control.
The key difference between braking and accelerating is that in the former case, independent, potentially differentially worn brakes, apply force unevenly, making the chance of a loss of traction on one or more wheels higher. With acceleration, that force is applied through a differential, meaning it will be far more likely to be appropriately distributed.
If you want to decelerate while swerving it can be done, but it should be done through engine braking - and the tricky bit there is matching revs as you drop the clutch back in, otherwise you have too much retarding force and overcome the coefficient of friction, resulting in a skid.
Easier for those of us who grew up with double de-clutching and no synchromeshes, but when you’re in a critical situation, it’s still an awful lot easier to apply acceleration.
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.
In theory, it depends. In practice, slamming on your brakes is the correct call 99% of the time. To a large extent that is because of the "competent driver" part. I'd expect 80%+ of drivers to consider themselves just that, whereas the truth is of course the opposite.
So, the correct advice is to say "brake, don't swerve", so that drivers internalize that their first thought and reaction in any emergency should be to brake. Teach them to actually brake—fully press on the pedal—while you're at it.
A slightly more nuanced advice would be "brake first, swerve as needed as a follow-up".
But I would never in good conscience be able to give anyone advice to swerve instead of braking.
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”.
> 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.
I think they have rotating specials. These are pre-made, cheaper, and smaller. When you have a membership this is the one you get for free (once a month).
They used to have a “Dr. Paul's Raw Animal-Based Smoothie” that I looooved, but unfortunately they stopped making it. Still sad about that one :(
Pffft, like a bot couldn't fake metadata.
I believe it is, in fact, humor.
Do you not like people being funny, or is there something else you're reacting so strongly to?
Mixing saving one's life with rating for a drink in the same comment makes feel very weird. Not funny, TBH, but actually it reads like a millions of generated bad review comments on amazon, expedia etc.
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.
Ads in the car.
Forced “safety breaks” due to the newly proven dangers of sitting in a car for more than 20 minutes. Taking place at our safety parter McDonalds.
Deliberately taking certain routes and encouraging you to stop at partner stores.
Making you pay rent for the self driving.
Increasing the subscription costs continuously.
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.
> Aircraft typically carry hundreds of people and can crash to the ground.
Cars are more numerous and could spontaneously either plow into pedestrians, or rear-end someone, causing chain damage and, quite often, a spillage of toxic chemicals (e.g., a cistern carrying acid/fuel/pesticide).
Plus, you have a problem of hostile actors having easier access to cars compared to planes.
It’s just death and taxes combined.
The way I see it, self-driving cars have the potential to deliver us from the burden of ownership altogether--maintenance, insurance, liability, parking, and all the rest. This hinges on availability, quality of service, pricing, and a rather large shift in the culture around cars and driving but I have hope that we can get there with time.
Cars are very expensive things to buy and own.
It's not obvious that will exist in the near future, anyway. Waymo aren't planning on selling their cars, and the economics and liability structure of self-driving strongly bias towards just running a taxi service.
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 all know it would end up just as rigged and as useless as Google's Play Integrity system for Android phones.
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...
Yeah it does lead to silly conclusions if all vehicle software of any kind is treated as a homogeneous blob.
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 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.
Until the Chinese or Mossad tell your car to drive into traffic.
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.
Pretty sure that's part of the taxi exam.
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 mean, I do talk to them and I do have this skill, but it's a skill that I only ever seem to employ in talking to Uber drivers, so I'm not sure it's of any great benefit.
If anything the fact that most of them are immigrants puts the conversation on easy mode if you're a native speaker. They're doing twice the mental work you are so it's easy to orchestrate the conversation.
Not really transferrable to native-speaking workers. Like speaking to a barista is very different. Speaking to a construction worker different again.
That's interesting. Cultural differences and language barriers aren't what I would consider weird.
I was thinking of those people who have wild stories and/or mountains of narcissism to overcome. They have a fascinating worldview like an artist would if they had those ambitions.
They get bonus points in my book the more genuinely unhinged and confused they seem to be. They got that way by questioning things into absurdity and I don't mind listening.
Well there's a virtuous cycle for immigrants whereby if you integrate, you improve the language, you get a better job, and you integrate more, thus often ironing out any weirdness wrt to the host culture. Uber driver is pretty dead-end and isolated. You work constant hours but all your interactions tend to be very surface level.
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.
> This seems a little silly.
Yes probably because it's obviously a joke
I despair
Using "second order effects" because big words sound cool without understanding the whole point of "second order"...
Which of 'second', 'order', and 'effects' is the big word?
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.
Yea. Cycling around self-driving cars is obviously much safer and many more people will be encouraged to do it.
> 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
Ya and they're the only ones I can count on being polite during rush hour
"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 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.
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
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.
> I sometimes pace them to act as a moving shield.
Yep, similar concept to using pedestrians crossing the road as shields. Cars reliably yield for them, not so much for cyclists.
> not even eye contact
Okay, DO NOT rely on eye contact. Look at what the vehicle and particularly its wheels, are doing.
> being waved on by a human
That's a potential death trap too. Just because one vehicle is yielding to you doesn't mean every other vehicle you're supposed to yield to is.
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
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
Well with an autonomous system, you can actually audit what the machine was "thinking" and how it came to think that way, so you actually can distinguish between an accident and criminal negligence better than a he-said-she-said situation with humans
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'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.
At highway speeds, engine, road and wind noise usually make horns inaudible.
In Serbia, on top of get-out-of-my-way, it's also used to signal go-ahead, but also "police with speed radars ahead" to incoming traffic.
I think we're not interpreting the original comment in the same way.
In most places, I think, when driving on the highway, flashing your lights when behind someone means basically 'I would like to overtake you'. Same here in the UK. But that's very specific to that context. You would never see a 'go ahead' context that would mean 'get out of my way', right?
But what the original comment means is there are some countries where you'd think it was 'go ahead' but it really means 'get out of the way'. Like if you're both on a main road, and you are signaling to turn into a side road, the opposing car flashes the lights and that means you can turn. I assume the same in Serbia.
But in some places that can actually mean don't turn, I'm going first. Which I think is what the parent is describing.
You are right that I did not read it the same way, and yes, the unwritten rules are matching in Serbia. FWIW, I've mostly switched to using left-turn signal to indicate "I'd like to overtake", which I've seen done on EU highways.
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.
The divider? What divider?
Why wouldn't it be trained to do that? You can easily include that in the training data.
It's not like the people building Waymo have never heard of flashing your brights before.
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.
I stopped biking on Seattle streets after being hit a third time - in a bike lane. I want human drivers off the roads.
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.
The tldr for me from the report was Waymo is great and SF is a dangerous place to drive.
I would largely agree with that assessment, yeah. Dangerous place to bike, too. I've even seen pedestrians get clobbered by bikers because they stepped into the bike lane not realizing a bike was barreling toward them at 20mph+. This is part of why Waymo and Uber warn you when the dropoff is next to a bike lane.
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.
That combined with the parent's post is, perhaps counterintuitively, somewhat concerning.
The proper technique for yielding to pedestrians wishing to cross is to start slowing down early, as if you were planning to stop before the crossing. That sends a clear signal to the pedestrian they're good to start crossing. Then you're free to speed back up. This is very comfortable for the pedestrian and the vehicle never needs to stop, so the slowdown is minimal.
That Waymos apparently don't act this way and seem to need to send an explicit signal to pedestrians sounds concerning to me, even if its ultimately safe.
Waymo does slow down as it approaches stop signs (usually where crosswalks are) and it will slow down if there is a pedestrian entering the roadway (crosswalk or not) since it doesn't want to crash into them.
The explicit signal of a driver noticing you (eye contact) is replaced by the signal above the vehicle. Are you not equally concerned that pedestrians have to get an explicit signal from drivers who are legally required to yield or stop??
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?
Search for the phrase "Vision Zero"
I think this ignores the argument the high speed limit people make which basically boils down to "sure some people will die or get injured but its worth it because driving faster is fun"
I’m absolutely sure people are just interested in shorter commute times rather than higher max speeds. That makes this an easy sell to citizens
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.
I've actually been involved in the commissioning of an FMVSS 202a headrest strength tester.
A lot of science and work goes into the construction of those headrests - if it was less firm, you'd get a concussion from the rotational forces in the whiplash or just break your neck, more firm and you'll get a concussion from the linear impact. It's not at all arbitrary, there's a reason they are exactly as firm as they are.
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.
Can you cite a specific state law that says that?
The last couple laws like that I checked only talked about limiting flow below the speed limit.
https://www.mit.edu/~jfc/right.html
I only know the law in Texas, so I'll cite that.
> (b) An operator of a vehicle on a roadway moving more slowly than the normal speed of other vehicles at the time and place under the existing conditions shall drive in the right-hand lane available for vehicles, or as close as practicable to the right-hand curb or edge of the roadway, unless the operator is:
> (1) passing another vehicle; or
> (2) preparing for a left turn at an intersection or into a private road or driveway.
https://tcss.legis.texas.gov/resources/TN/pdf/TN.545.pdf#545
Note this law specifically mentions "normal speed of other vehicles at the time and place" and doesn't directly mention speed limits. So by the text of this law, if you're driving the speed limit and hanging out in the left lane while the normal speed at that time is like 10 over you're technically breaking this law.
We have specific signage for highways where this is supposed to be the law.
https://www.txdot.gov/manuals/trf/smk/regulatory_signs/left_...
> 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.
> Results: A 5-mph increase in the maximum state speed limit was associated with an 8.5% increase in fatality rates on interstates/freeways and a 2.8% increase on other roads. In total during the 25-year study period, there were an estimated 36,760 more traffic fatalities than would have been expected if maximum speed limits had not increased—13,638 on interstates/freeways and 23,122 on other roads.
https://www.iihs.org/research-areas/bibliography/ref/2188
I do not understand the point you are making here. I did not make an argument in favor of increasing the speed limit.
I believe I replied to the wrong person!
That doesn't make sense to me. If you want to change lanes, and worst case scenario you're right next to someone, go 2mph slower for 20 seconds and they'll be shifted by 60 feet. I'm sure you can plan your lane changes on a freeway 20 seconds in advance.
It is the dynamics. When lanes are pacing each other the gaps all tighten up. So sure, you can slow down a bit to find the gap behind you, except that gap is not big enough to fit in. So you turn on your signal and wait for someone polite enough to let you in, meanwhile the guy behind you is riding you like a pony because you are no longer keeping up with traffic.
When traffic isn't balled up so tight, you can plan for a lane change in advance and accomplish it without having to slow down traffic. Everything flows better.
> 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.
> 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
Most Americans ignore speed limits. This stems from it being socially and legally problematic to permanently revoke our driver’s licenses. We should raise a lot of limits. But many others are fine and still sped through.
Where do you guys have 55 mph where it's not appropriate? A LLM is telling me that 55 mph is your random urban interstate. The equivalent where I live is 80 km/h on motorways inside city/town limits, which is 50 mph, and it feels very appropriate, cars make a ton of noise and you don't want the full motorway speed in the city. And that's basically the fastest you'll ever go in a city, when it's an actual motorway (often elevated, ramps, sound barriers). Other roads (even big and important) are ~43 mph and the general urban speed limit is ~31 mph.
> A LLM is telling me that 55 mph is your random urban interstate.
Yes, this sounds about right. In the metro area, 55 mph on a limited access interstate freeway. Arterial surface streets typically 40-45 mph, lower level surface streets commonly 25 mph and sometimes 20 mph depending on locality.
In the US, in particular out west where I live, 'urban' does not have the same meaning as it does somewhere much more dense, so it amounts to 55 mph in many places you might regard as rural.
Drive 55 stay alive.
On the stretch of motorway that I frequent in Italy, the speed limit is mostly 130 km/h, but the majority of people drive at about 100 to 110 km/h (including me).
But there are also people who drive in the left lane, who will tailgate you at 1 or 2 meters because you're doing 130 km/h. These people are idiots, but you get these sorts of people everywhere.
On American freeways, you don't have a choice, every lane is doing about 10 mph over the limit (or in LA way under) and it is disruptive or dangerous not to. These freeways tend to be running at full capacity so it actually makes sense since it improves capacity.
Road capacity does not increase with speed above 50 km/h on urban roads or 70 km/h on highways. Following distance scales with speed, so more speed can actually mean fewer cars per unit of time.
In theory, braking distance scales quadratically with speed. In practice, people leave less room on highways, because they rely on others driving predictably, but spacing still increases faster than linear.
speed limits in the US are barely suggestions, let alone rules
You drive an ambulance? Or a fire truck?
Being forced to drive the speed limit isn't that big of a deal
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.
> 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.
Do you have a solution for preventing humans from driving while under the influence of alcohol or drugs? Or even preventing humans from driving while texting?
The insurance industry will solve that on their own as technology advances. It simply won't be worth insuring a human that uses alcohol/drugs.
Most accidents are caused by drunk drivers, distracted drivers, jackass drivers and others, who together still constitute a minority, and there are plenty single-vehicle crashes. 13x improvement over average may still constitute diminished safety for a diligent sober driver.
13x better than an average that includes drunk, fatigued and distracted drivers would not be very impressive on a closed course.
On a real world course where the only way to achieve those kind of numbers is to avoid getting hit by those drunk, fatigued and distracted drivers? Very impressive.
Yes, but Waymo also has to drive on the road with those drivers, and these stats include crashes that are their fault. Diligent drivers get hit by drunk/distracted drivers all the time.
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.
Yes, but I don't actually trust most folks to be in a "good condition". Even the best drivers have their off days.
Well, if we were talking about forcing people to stop driving and transition to current waymos it's plausible that diligent sober drivers would be facing greater risk. Would that be acceptable to improve average statistics?
> if we were talking about forcing people to stop driving and transition to current waymos
We're not
If you're arguing to the regulator that "look these cars are not as bad as humans overall" I guess that's fine, but if you're trying to sell me on using self-driving by comparing to something that's completely unrelated to my use-cases that's just BS.
Unless their message is "if you're drunk turn on self driving" which I could get behind, I sincerely doubt current self driving is better than humans - simply because of the data they chose to compare to. If they were better than professional taxi drivers I'm sure they would tout that data widely.
This perspective makes the statistic even more impressive. Every day driving they interact with probably thousands of people who are impaired in some way and manage to drive safely far more often
No not really - because it doesn't compare to other comparable drivers. You're assuming that human drivers can't drive defensively.
Like a realistic comparison would be comparing it to taxi drivers/ride share drivers - then you'd see the risk vs using a cab/uber/whatever. Not a drunk 18 year old with a clunker with no brakes.
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.
This is very dependent on the environment. I would not be so confident saying this about the USA in particular.
If you can stick to at least half-decent dedicated cycling infrastructure separated from cars, sure. If riding alongside traffic, I wouldn't be surprised that, depending on the exact route, the gains wouldn't outweigh the risks.
Generally speaking, operating a bike safely is considerably more difficult than a car and the margins of error are tiny.
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.
Sorry, yes, I didn't necessarily mean full grade separation, I was mostly thinking about just separated bike lanes in general. Having some kind of real buffer between the cyclists and the murder monkeys.
https://www.txdot.gov/manuals/des/rdw/chapter-18-bicycle-fac...
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.
People won't collectively realize squat.
Insurance companies will simply price people out of driving. And I welcome it SO very much.
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.
Self driving doesn't mean forced self driving, there will always be a manual driving override. But your insurance may be more expensive.
"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.
And they're in Pittsburgh. Seen them driving while snowing.
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.
Ooh, that's a worthy challenge. Of course, I can imagine getting enough data on all of those cities and deciding to launch everywhere else but not Boston "because your roads are garbage and you all drive like you're impaired 24/7" :-)
They’re in Denver now (not available to the public yet) for this exact reason. Unfortunately we’ve had our least snowy winter in recorded history this year - so they only got a few days of real snow.
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
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.
OK right, so they're at 170 million now, maybe ~1 life saved. I see it might be a bit early to be making claims about that.
Also in comparison the cases of disruption/blocking intersections/emergency-services do seem significant - seems plausible in the right circumstances you could lose lives.
Seems they need to 10x their miles before they can start making confident claims about lives saved.
Wow. >40% of their accidents have a speed delta-V of less than 1mph. They should just sell this thing as a kit to pop onto your existing car, with a major discount on car insurance. I’d buy it in a heartbeat if they can get it under $30k.
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)
https://people.com/waymo-exec-reveals-company-uses-operators... might be relevant content as well
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.
Is this an independent study?
They make the data available on that page for you to do your own research.
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.
This page is old, but they just refreshed the data shows Waymo is 13x safer than human drivers (in the cities it operates in).
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.
You seem to be thinking this is compared to median driver, and not average one: if most accidents are caused by small number of bad drivers, your average will be bad, but even median might still be good.
And this does not even compare the drivers, but simply miles driven.
So I think that 80% of human drivers would likely be safer than Waymo unless they are driving under the influence or extremely tired or distracted.
Note that "13x safer" already implies being in the top 10%, though.
From a purely technical perspective, I would expect a Waymo to react to unexpected external stimuli much, much faster than a human. It gives the Waymo options that are unavailable to a human's slower reaction time.
there was a startup I saw once here on HN selling driverless kits that could be added to general cars.
if Waymo can do the same - then partner with insurance companies - they can easily be a $1 trillion dollar company even if they just reduce accidents by 5X.
The insurance market or cost thereof is their market.
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.
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.
anecdotally from Mesa:
DoorDash has these little cute robots doing delivery. I often seen them followed by a person on a e-bike. This has been going on for more than a year. My recent Lyft driver said one reason is because the Waymo's ignore the other robot and kill them and the bike ensures they don't.
I don’t use Waymo’s because they are really slow. At least for me. I get they’re safer. What’s new is everyone I know is complaining how expensive they are. Why are they so expensive? Is it because they are trying to make a profit on the total cost of ownership of a vehicle? They’re electric so it’s not gas. Uber and Lyft definitely outsource the cost odd maintenance onto the contractor.
Pretty cool to see. But the UI of the visual animations has some weird re-re-rerender bug, at least on mobile safari.
I love Waymo. Using the word "impact" is unfortunate.
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".
I wish the executives at Waymo would go to prison until their cars stopped breaking traffic laws left and right. I get they're safer then humans. I've enjoyed my rides. But they're doing one of this "fuck the laws while we cash in things" because they know they can get away with it.
I guess more realistically, I wish the government would step in demand footage because the cars have records of every law they broke.
So how long until this is enshittified and how will it go? Will people ask for cars to be faster so they will take riskier decisions? Will a cheaper company with a worse training model capture the most market share? Pilot products are usually great, but that rarely lasts into their profitable period.
Waymo's funniest home videos:
https://x.com/xdNiBoR/status/2027296859803336869
https://x.com/TexasTSLA/status/2030400998611788226
https://x.com/barstoolsports/status/1995629394308821184
https://x.com/niccruzpatane/status/1950676793335152693
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.
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.
No, I think those are solidly less concerning than crash and injury rates.
Like, could you imagine yourself seeing a self driving car that's perfect in all the ways like that, no dumb quirks, but it crashes twice as often as a human, and saying "well it got the important stuff right"
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.
You seem to confirm the point that you adapt your behavior in presence of Waymos, even if you believe it is in the other direction.
The argument was for _how_ people react to Waymos. The OP said folks are more careful. The respondent said, no it's the opposite.
Neither argued that people do not adapt their behavior in the presence of Waymos?
The poster who claimed they are super careful also followed that with:
The fact that someone adapts their behaviour (regardless in what direction) still supports that claim."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.
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.
Median and average driver are close to identical. This isn't income.
(Edit: I now think my comment above is wrong)
Driver risk is a Pareto distribution.
I mean, if some subset of those 15% used Waymo, we'd be better off, no?
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
Maybe at your country. Not at mine. Driving irresponsibly might have consequences. If you race, advance red light on don't stop for pedestrians, if there is not enough evidence you couldn't do anything about it, you might go to jail.
I want whatever causes fewer deaths and injuries total.
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.
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.
> 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.
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”
My understandig is that in private settings, third-party consent has a very difficult legal bar to meet.
“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.
... no? That children, animals etc. acts differently everyone knows. But what about a self-driving car that looks the same as every other car?
Anyway, it was Waymos own findings when they started out. They got into more accidents, none of which where their own fault, than expected and realized that they had to make it behave more like a human to not confuse human drivers.
> ... no?
Which part are you confused about?
> what about a self-driving car that looks the same as every other car?
My Aussie friends, when they visit America, don't put up a sign in their window saying they're Aussie and will occasionally try to turn left at a red.
> They got into more accidents, none of which where their own fault, than expected and realized that they had to make it behave more like a human to not confuse human drivers
Sure. I'm still filing this in the nothingburger file. If anything, it screams that we have a lot of people on the road who should not be.
Nothing.
Yeah? I thought that was a given...
I think the other way to think about it is that Waymo is probably in the 99th percentile for not being distracted and 99th percentile for reaction time, always, just from a pure sensor and computation standpoint.
Even the best human can't say that, at all times?
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.
I, and I expect most human drivers, also probably drive measurably worse if I'm in a city I'm not familiar with.
> 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.
London cabbys can be hangry or have broken up with their partners that day. What makes you so sure they're fully attentive on any given day?
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
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"
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.