[HN Gopher] Waymos crash less than human drivers
___________________________________________________________________
Waymos crash less than human drivers
Author : rbanffy
Score : 100 points
Date : 2025-03-26 20:57 UTC (2 hours ago)
(HTM) web link (www.understandingai.org)
(TXT) w3m dump (www.understandingai.org)
| doctorpangloss wrote:
| Another point of view is, if CVC violations were enforced against
| human drivers as robustly as they are against Waymos; and if
| human drivers were held to the same standards of liability as
| Waymos; human drivers in California would be way safer too. To
| me, the overall safety of all the driverless programs should be
| interpreted as a huge victory for regulators.
| GolfPopper wrote:
| Another point of view is: does the comparison involve similar
| vehicles, with similar maintenance, and under similar
| conditions?
| elmerfud wrote:
| I think this is not accounted for in any of the data. There are
| a lot of bad human drivers that are not removed from the road.
| There are a lot is human drivers that never bother with
| increasing their skills or experience in driving. Those people
| can drag down the numbers.
|
| Because with waymos anything learned by one is immediately
| available to all of them. Therefore you have lots of miles but
| effectively one driver. Now if you take human drivers and pick
| only the best, how do you numbers look then? Many people have
| driven lifetimes and not been in an accident or not in an
| accident that was their fault or reasonable predictable and
| avoidable by them.
|
| Driving needs to be considered a lot more serious of a task
| than it currently is. There are many driving errors that are
| the primary cause of accidents but those people are never
| forced to get extended drivers education. Self driving cars are
| nice but if the goal is to prevent accidents it's a bit like
| requiring everyone to order prepared food to prevent injury
| while cooking.
|
| At some point we need to get back to shaming those who are bad
| at something and requiring that they improve their skills
| rather that restricting everyone else's freedoms.
| doctorpangloss wrote:
| > There are a lot of bad human drivers that are not removed
| from the road.
|
| Well. If you let people get away with murder by saying "Oh,
| I'm a confused little old lady, I thought the gas pedal was
| the brake." That's what I mean by robust enforcement. It is
| 100% not at all about data.
|
| Consider this: A Waymo could be _provably_ confused about a
| situation, we can _directly_ peer into the state of its mind
| and _prove_ that it is confused, despite driving safer than
| the average. Do you know what would happen to the Waymo
| program if one of its vehicles careened into a family of
| pedestrians and killed them all?
|
| > something something about skills
|
| Agency is kind of a myth. For human drivers, experts agree
| that road design has the biggest impact. Experts also think
| we should have better public transport.
|
| The problems are bigger than Department of Unenactable Urban
| Fantasy or American Victim Blaming Institute. Self driving
| cars are exciting because they are like a cheat code around
| all these intractable social issues.
| ikerino wrote:
| > Using human crash data, Waymo estimated that human drivers on
| the same roads would get into 78 crashes serious enough to
| trigger an airbag. By comparison, Waymo's driverless vehicles
| only got into 13 airbag crashes. That represents an 83 percent
| reduction in airbag crashes relative to typical human drivers.
|
| > This is slightly worse than last September, when Waymo
| estimated an 84 percent reduction in airbag crashes over Waymo's
| first 21 million miles.
|
| nitpick: Is it really slightly worse, or is it "effectively
| unchanged" with such sparse numbers? At a glance, the sentence is
| misleading even though it might be correct on paper. Could've
| said: "This improvement holds from last September..."
| 0_____0 wrote:
| I wonder how many of those airbag crashes were an At-fault or
| shared-fault of the Waymo AVs.
| onlyrealcuzzo wrote:
| Considering that there's a >1000:1 ratio of regular cars to
| Waymo AVs - Waymo would have to be EXTREMELY terrible at
| driving to move the numbers for the other group meaningfully
| - which would show up in Waymo's own crash data.
|
| There's also historical data. So if you saw a spike in
| crashes for regular vehicles after Waymo arrives, it would be
| sus. But there is no such spike. Further evidence Waymo isn't
| causing problems for non AVs.
|
| Of course anything is possible. But it's unlikely.
| MostlyStable wrote:
| I'm confused by your comment. We shouldn't expect that
| Waymo accidents should budge overall accidents (which seems
| to be what you are talking about), but it wouldn't be crazy
| for Waymo, even if it was much safer overall, to be
| responsible for some non-trivial amount of all the
| accidents it has.
|
| For example, imagine that Waymo is (somehow) far far _far_
| superhuman in it 's ability to avoid other cars doing
| dumb/bad things. It has a dramatic reduction in overall
| accidents because it magically can completely get rid of
| accidents where the other driver is at fault. But, in some
| very specific circumstances, it can't figure out the proper
| rate to slow down at intersections, and it consistently
| rear ends vehicles in front of it. This specific situation
| is very rare, so overall accidents still are low (much
| lower than human drivers), but, in our made up, constructed
| (and extremely non-sensical) hypothetical, nearly 100% of
| Waymo accidents are Waymos fault.
|
| So I don't think it's ridiculous to ask how many of the
| accidents Waymo has been involved in are the fault of the
| Waymo vehicle. It turns out that (assuming Waymo's side of
| the story is to be trusted), almost none of them are their
| fault, but it didn't _have_ to be that way, even in the
| case where Waymo accidents were more rare than human
| accidents.
| MostlyStable wrote:
| Assuming you trust Waymo's account, the article details them,
| saying the following:
|
| >So that's a total of 34 crashes. I don't want to make
| categorical statements about these crashes because in most
| cases I only have Waymo's side of the story. But it doesn't
| seem like Waymo was at fault in any of them.
| QuercusMax wrote:
| Back when I used to pay attention to this stuff, the vast
| majority of crashes occurred when the Waymo vehicle was
| rear-ended while stopped at a traffic light.
| barbegal wrote:
| Of course it's not worse, these numbers have huge error bars.
| Statistically the two statistics are not significantly
| different. But trying to explain that to most people with no
| knowledge of statistics is tough.
| nukem222 wrote:
| 83 vs 84 percent doesn't seem too difficult to present as
| "essentially the same". I also don't think this matters much
| --the result is impressive regardless of alleged rate of
| change.
| unreal6 wrote:
| One would hope that a writer on this subject would have at
| least a cursory knowledge of statistics
| awongh wrote:
| The number of miles driven seems large, but Gemini says there
| are thousands of crashes per day in the US- so 78 or 13 crashes
| is a really small sample size....
| shadowgovt wrote:
| One of the more interesting things Waymo discovered early in the
| project is that the actual incidents of vehicle collision were
| under-counted by about a factor of 3. This is because NHTSA was
| using accident reports and insurance data for their tracking
| state, but only 1/3 of collisions were bad enough for either
| first responders or insurance to get involved; the rest were
| "Well, that'll buff out and I don't want my rates to go up,
| so..." fender-taps.
|
| But Waymo vehicles were recording and tracking all the traffic
| around them, so they ended up out-of-the-starting-gate with more
| accurate collision numbers by running a panopticon on drivers on
| the road.
| nukem222 wrote:
| I would hope so! Are these rides actually cheaper? Assumably this
| is orders of magnitude less expensive than hiring a human and the
| driver is what you pay for. I don't see myself ever getting in a
| car if my safety is actually a concern.
|
| I don't see myself using any of these any time soon, I tend to
| drive and walk everywhere and don't see much point to paying
| someone else to drive barring extenuating circumstances. But
| _assuming actual cost benefits are delivered to customers_ this
| might be pretty exciting.
| jsight wrote:
| Right now, they are pricing them similarly to Uber and Lyft,
| though they do seem to offer more frequent discounts and there
| is no tipping yet. The experience is better too, IMO.
| theLiminator wrote:
| > But assuming actual cost benefits are delivered to customers
| this might be pretty exciting.
|
| I imagine that will require a viable alternative. Right now
| they're in some senses superior to ride-sharing/taxis as they
| are (potentially) safer, and cleaner. So there's no real market
| pressure forcing them to lower prices beyond that of uber et-
| al.
| mjburgess wrote:
| Waymos choose the routes, right?
|
| The issue with self-driving is (1) how it generalises across
| novel environments without "highly-available route data" and
| provider-chosen routes; (2) how failures are correlated across
| machines.
|
| In safe driving failures are uncorrelated and safety procedures
| generalise. We do not yet know if, say, using self-driving very
| widely will lead to conditions in which "in a few incidents" more
| people are killed in those incidents than were ever
| hypothetically saved.
|
| Here, without any confidence intervals, we're told we've saved
| ~70 airbag incidents in 20 mil miles. A bad update to the fleet
| will easily eclipse that impact.
| thaumasiotes wrote:
| > The issue with self-driving is (1) how it generalises across
| novel environments without "highly-available route data" and
| provider-chosen routes; (2) how failures are correlated across
| machines.
|
| Why is (1) an issue? Route data never gets worse.
| bluefirebrand wrote:
| Route data gets worse all the time
|
| Any time there is a detour, or a construction zone, or a
| traffic accident, or a road flooded, or whatever else your
| route data is not just "worse" it is completely wrong
| zmmmmm wrote:
| it's true, but if the scenarios are rare you can treat them
| as outliers and have the vehicle switch to a "safety mode"
| or similar. Even human drivers effectively do that (or
| should!).
| ceejayoz wrote:
| > Route data never gets worse.
|
| Construction? Parade? Giant tire-crunching pothole in the
| middle of the freeway?
| virtue3 wrote:
| Oddly enough Waymo does a solid job of avoiding those.
| mjburgess wrote:
| Consider London: a series of randomly moving construction
| sites connected by patches of city.
|
| Waymo, as far as I recall, relies on pretty active route
| mapping and data sharing -- ie., the cars arent "driving
| themselves" in the sense of discovering the environment as a
| self-driving system would.
| bsder wrote:
| Huh? All the Waymo cars in Austin have lots of LIDAR spinny
| bits and are driving around all the active construction
| zones downtown and near campus.
|
| I'm pretty sure that counts as "discovering the
| environment".
| mjburgess wrote:
| Sure, my understanding is that they collectively share
| data, and this is combined with central mapping.
|
| On net, yes, they are sensitive to features of the
| environment and via central coordination maintain a safe
| map of it.
|
| The mechanism there heavily relies on this background of
| sharing, mapping, and route planning (and the like) --
| which impacts on the ability of these cars to operate
| across all driving environments.
| AlotOfReading wrote:
| Waymo's map data is a prior, not an authoritative reference
| to the world. The cars report update when the map data is
| wrong, and a lot of what they use it for (e.g. traffic
| light identification, fine localization) degrades
| gracefully when there's new information in the environment.
| jrussino wrote:
| I wonder if you can decrease the impact of (2) with a policy of
| phased rollout for updates. I.E. you never update the whole
| fleet simultaneously; you update a small percentage first and
| confirm no significant anomalies are observed before
| distributing the update more widely.
| timschmidt wrote:
| Ideally you'd selectively enable the updated policy on
| unoccupied trips on the way to pick someone up, or returning
| after a drop-off, such that errors (and resultant crashes)
| can be caught when the car is not occupied.
| nukem222 wrote:
| Presumably management would also be highly regional.
| Functionality in san francisco doesn't imply anything about
| functionality in oakland, etc.
| mjburgess wrote:
| One measure of robustness could be something like: the
| ability to resist correlation of failure states under
| environmental/internal shift. Danger: that under relevant
| time horizons the integral of injury-to-things-we-care-about
| is low. And then "safety", a combination: that the system
| resists correlating failure states in order to preserve a low
| expected value of injury.
|
| The problem with machines-following-rules is that they're
| trivially susceptible to violations of this kind of safety.
| No doubt there are mitigations and strategies for minimising
| risk, but its not avoidable.
|
| The danger in our risk assessment of machine systems is that
| we test them under non-adversarial conditions, and observe
| safety --- because they can quickly cause more injury than
| they have ever helped.
|
| This is why we worry, of course, about "fluoride in the
| water" (, vaccines, etc.) and other such population-wide
| systems... this is the same sitation. A mass public health
| programme has the same risk profile.
| seper8 wrote:
| Does waymo also choose the times of driving, and conditions? Or
| do they always drive, even at night and in heavy rain?
| maxerickson wrote:
| Correctly estimating capability is pretty safety positive.
|
| Meaning, humans choosing to drive in more difficult
| conditions probably means they sometimes drive in conditions
| that they shouldn't.
| a2128 wrote:
| I was waiting for a Waymo in Austin during the weekend storm
| and the Waymo suddenly cancelled on us right after a power
| outage that lasted a second or two. According to local news
| the vehicles had stopped and were blocking traffic.
|
| https://www.msn.com/en-us/technology/tech-companies/waymo-
| ve...
| qgin wrote:
| This is good to know, but we should note that humans choose
| also.
| tonyhart7 wrote:
| Yeah but that's the point no???
|
| machine don't make mistake when they are get perfected in
| certain route, sure human drive would be better in dynamic
| areas but you dnt need machine to be perfect either just want
| (80% scenario)
| arghwhat wrote:
| > The issue with self-driving is (1) how it generalises across
| novel environments
|
| That's also an issue with humans though. I'd argue that traffic
| usually appears to flow because most of the drivers have taken
| a specific route daily for ages - i.e., they are _not_ in a
| novel environment.
|
| When someone drives a route for the first time, they'll be
| confused, do last-minute lane changes, slow down to try to make
| a turn, slow down more than others because because they're not
| 100% clear where they're supposed to go, might line up for and
| almost do illegal turns, might try to park in impossible
| places, etc.
|
| Even when someone has driven a route a handful of times they
| won't know and be ready for the problem spots and where people
| might surprise they, they'll just know the overall direction.
|
| (And when it is finally carved in their bones to the point
| where they're placing themselves perfectly in traffic according
| to the traffic flow and anticipating all the usual choke points
| and hazards, they'll get lenient.)
| mjburgess wrote:
| People have eyes, ears, a voice, hands, etc.
|
| You've a very narrow definition of novel, which is based
| soley on incidental features of the environment.
|
| For animals, a novel situation is one in which their learnt
| skills to adapt to the environment fail, and have to acquire
| new skills. In this sense, drivers are rarely in novel
| environments.
|
| For statistical systems, novelty can be much more narrowly
| defined as simply the case where sensory data fails a
| similar-distribution test with historical data --- this is
| vastly more common, since the "statistical profile of
| historical cases, as measured, in data" is narrow.. whilst
| the "situations skills apply to" is wide.
|
| An example definition of narrow/wide, here: the amount of
| situations needed to acquire safety in the class of similar
| environments is exponential for narrow systems, and sublinear
| for wide ones. ie., A person can adapt a skill in a single
| scenario, whereas a statistical system will require
| exponentially more data in the measures of that class of
| novel scenarios.
| npunt wrote:
| Generalizing across novel environments is optimal, but I'm not
| sure the bar needs to be that high to unlock a huge amount of
| value.
|
| We're probably well past the point where removing all human-
| driven vehicles (besides bikes) from city streets and replacing
| them with self-driving vehicles would be a net benefit for
| safety, congestion, vehicle utilization, road space, and hours
| saved commuting, such that we could probably rip up a bunch of
| streets and turn them into parks or housing and still have
| everyone get to their destinations faster and safer.
|
| The future's here, even if it still has room for improvement.
| floxy wrote:
| >congestion
|
| I'd think congestion would go up as AVs become more popular,
| with average occupancy rates per vehicle going down. Since
| some of the time the vehicle will be driving without any
| passengers inside. Especially with personally owned AVs.
| Think of sending a no-human-passenger car to pick up the dog
| at the vets office. Or a car circling the neighborhood when
| it is inconvenient to park (parking lot full, expensive,
| whatever).
| npunt wrote:
| Up to 30% of cars on city streets at any given time are
| looking for parking [1].
|
| Cars are also the least utilized asset class, being parked
| 95% of the time [2].
|
| AVs, by virtue of being able to coordinate fleet-wide and
| ability to park anywhere rather than only one's home or
| destination, would be able to gain incredible efficiencies
| relative to status quo.
|
| Atop those efficiencies, removing both the constraint of
| having a driver and the constraint of excessive safety
| systems to make up for human inattentiveness means AVs can
| get drastically smaller as vehicles, further improving road
| utilization (imagine lots of 1- and 2-seaters zipping by).
| And roads themselves can become narrower because there is
| less room for error with AVs instead of humans.
|
| Finally, traffic lights coordinating with fleets would
| further reduce time to destination (hurry up and finish).
|
| Self-driving vehicles give us the opportunity to rethink
| almost all of our physical infrastructure and create way
| more human-friendly cities.
|
| [1]
| http://shoup.bol.ucla.edu/PrefaceHighCostFreeParking.pdf
|
| [2] https://senseable.mit.edu/unparking/
| xboxnolifes wrote:
| > AVs, by virtue of being able to coordinate fleet-wide
| and ability to park anywhere rather than only one's home
| or destination, would be able to gain incredible
| efficiencies relative to status quo.
|
| > Atop those efficiencies, removing both the constraint
| of having a driver and the constraint of excessive safety
| systems to make up for human inattentiveness means AVs
| can get drastically smaller as vehicles, further
| improving road utilization (imagine lots of 1- and
| 2-seaters zipping by). And roads themselves can become
| narrower because there is less room for error with AVs
| instead of humans.
|
| The first part is mostly describing taxis, so the
| incredible efficiencies relative to the status quo can be
| loosely observed through them. Just subtract out wage and
| a slight "technological scale" bonus, and you can
| estimate what it would be.
|
| The second part, I wonder. Cars already average under 2
| occupants, with most just being the driver. If this is
| what is was needed for significantly smaller cars, we
| would already have them. Lack of smaller cars is _mostly_
| a cultural issue, not a technical one.
| kccqzy wrote:
| I don't agree with this novel environment argument about
| routes. As a human, there are a limited number of roads that I
| have driven on. A taxi driver drives better than me because
| none of the routes are considered novel: the taxi driver has
| likely driven on every road in a city in his/her career. The
| self-driving machine has most definitely driven on every single
| road in the city, perhaps first as testing with human backup,
| then testing with no passengers, and finally passenger revenue
| miles.
| KoolKat23 wrote:
| I think you underestimate how many novelties the car will
| encounter on existing routes and how adept these cars are at
| navigating novel routes.
|
| I imagine this route data is an extra extra safeguard which
| allows them to quantify/measure the risk to an extent and also
| speed up journey's/reduce level of interventions.
| ogogmad wrote:
| > In safe driving failures are uncorrelated and safety
| procedures generalise. We do not yet know if, say, using self-
| driving very widely will lead to conditions in which "in a few
| incidents" more people are killed in those incidents than were
| ever hypothetically saved.
|
| Can you provide some examples of what you mean?
| ozim wrote:
| I can imagine whole city areas well known closed to manual
| drivers.
|
| Sure I would love to read a book while car is driving me to
| visit family in the countryside but practically I need city
| transportation to work and back, to supermarkets and back where
| I don't have to align to a bus schedule and have 2-3 step overs
| but plan my trip 30 min in advance and have direct pick up and
| drop off.
|
| If that would be possible then I see value in not owning a car.
| grakasja wrote:
| This statistic could be misleading, because not all miles are
| equally dangerous. Google is very careful about selecting where
| it deploys and tests Waymo, preferring flat, safe, well-designed
| areas. Routing is also closely monitored and I would imagine that
| problematic roadways are avoided. The article says they compared
| it to human accident rates "on the same roads" but doesn't
| clarify their methodology for "same"ness. It also doesn't factor
| in driver experience. A taxi driver who has memorized a
| particular route is likely going to drive safer than a tourist
| who has never gone on that same road before. Waymo may be safer
| than the average driver on X road but that doesn't mean it will
| have the same comparative performance if you drop it onto a
| random road it has never driven before with no assistance from
| human support staff.
| bryanlarsen wrote:
| > preferring flat, safe, well-designed areas.
|
| Like downtown San Francisco?
| grakasja wrote:
| It took them a long time to deploy there and required
| extensive (human) preparatory work. I'm guessing that their
| algorithm was fine-tuned for SF specifically, possibly being
| taught to avoid certain roads and areas where possible, thus
| behaving like a veteran local driver. Not clear how well that
| scales globally, or if it would still drive much better than
| an average human in a random new environment that it hasn't
| been pre-trained for.
| jrussino wrote:
| > The article says they compared it to human accident rates "on
| the same roads" but doesn't clarify their methodology for
| "same"ness.
|
| Reading your comment before the article, my first thought was
| that "on the same roads" must mean literally the same roads -
| right?
|
| But the article actually says:
|
| > Using human crash data, Waymo estimated that human drivers on
| the same roads would get into 78 crashes
|
| I agree that this is unclear. What data did they use, and why
| did they have to _estimate_ at all? Shouldn 't they be able to
| get the actual data for how many human drivers got into such
| accidents on this same exact set of roads over this same exact
| time period?
| ra7 wrote:
| See https://waymo.com/safety/impact/#methodology
|
| They compile human crash data from various sources (NHTSA,
| state data). But they are at a city level, not specific
| streets or areas. So they adjust the human benchmarks to be
| more representative of the areas where Waymo operates.
| floxy wrote:
| Robotaxis might also make the human piloted accident rate go
| down, if enough inebriated people take AVs instead of driving.
| jsight wrote:
| If Waymo is systematically choosing safer routes, that sounds
| like a good thing.
| jyounker wrote:
| They're going wherever customers in SF are ordering them to go.
| I was in SF last month, and everywhere I went there Waymos
| skulking around, even on little side-streets on top of hills.
| bmitc wrote:
| False. Humans are driving in a much wider range range of weather,
| road conditions, car conditions, passenger conditions, routes,
| unknown destinations, etc.
| murderfs wrote:
| Something like a third of traffic fatalities are due to one of
| the drivers being drunk.
| floxy wrote:
| OT, but does anyone know what the shape of the curve for the
| number of automobile crashes per human driver? Is it uniform
| distribution, where everyone is more or less likely to get into,
| say, 1.2 fender-benders per lifetime? Or there is a cluster of
| people who are much more likely to be involved in crashes? I
| suppose automobile insurance companies would have this type of
| information.
| paxys wrote:
| I'm willing to bet it's the latter. There's a reason your
| premiums go up after you get into an accident. Insurance
| companies are betting that you are going to have more of them.
| AlotOfReading wrote:
| It's more complicated than just "good drivers" and "bad
| drivers". Your driving performance is very different on
| minute 10 of a trip vs minute 600, or 11am vs 11pm, or happy
| vs having a bad day. You're also very likely to be in very
| different environments at all of these times, and have
| different familiarities.
|
| Insurance risk estimates are practical simplifications
| instead of trying to model how risky an individual driver is
| on a particular day at a particular time, in a particular
| area. Waymo's results are trying to compare to a
| statistically average driver with all other variables
| controlled.
| jeffbee wrote:
| It is very much the latter. It is like the statistics of
| average alcohol consumption: half the people consume none, and
| that one guy drinks non-stop. In fact, the drinking guy and the
| crashing guy might be the same guy.
| nickvec wrote:
| Very excited for the future of AVs. So many lives will be saved
| using this technology.
| JKCalhoun wrote:
| The idea that everyone has their own "public transportation" (a
| computer chauffeur) seems lacking in foresight.
|
| I suspect, but don't know, that buses are safer still? (Not aware
| of any airbags triggering, ha ha.)
| doctorpangloss wrote:
| > The idea that everyone has their own "public transportation"
| (a computer chauffeur) seems lacking in foresight.
|
| How so? As you increase the density of stops in a bus network
| and increase the rate of arrivals, there will be fewer
| passengers per bus, going on journeys that approach the fastest
| they could be. Why look at the one thing that has a really good
| chance of "fast, cheap and good" and say, "lack foresight?" My
| dude, it's the only game in town!
| labrador wrote:
| I was initially skeptical about self-driving cars but I've been
| won over by Waymo's careful and thoughtful approach using visual
| cues, lidar, safety drivers and geo-fencing. That said I will
| never trust my life to a Tesla robotaxi that uses visual cues
| only and will drive into a wall painted to look like the road
| ahead like Wile E. Coyote. Beep beep.
|
| Man Tests If Tesla Autopilot Will Crash Into Wall Painted to Look
| Like Road https://futurism.com/tesla-wall-autopilot
| KoolKat23 wrote:
| To be fair, I'm sure there's a few humans that would crash into
| a giant painted road in the middle of a straight road in the
| middle of nowhere. Humans crash due to less.
| labrador wrote:
| Well, I wouldn't and neither would a Waymo, but a Tesla did,
| which means it's no better than a bad human driver
| ggreer wrote:
| Mark Rober's video is misleading. First, he used autopilot, not
| FSD. Second, he sped up to 42mph and turned on autopilot a few
| seconds before impact[1], but he edited the Youtube video to
| make it look like he started autopilot a from a standstill far
| away from the barrier. Third, there is an alert message on his
| screen. It's too small to read in the video, but it could be
| the "autopilot will not brake" alert that happens when you put
| your foot on the gas.
|
| In the water test, Rober has the Tesla driving down the center
| of the road, straddling the double yellow line. Autopilot will
| not do this, and the internal shots of the car crop out the
| screen. He almost certainly manually drove the car through the
| water and into the dummy.
|
| One person tried to reproduce Rober's Wile E. Coyote test using
| FSD. FSD v12 failed to stop, but FSD v13 detected the barrier
| and stopped in time.[2]
|
| Lidar would probably improve safety, but Rober's video doesn't
| prove anything. He decided on an outcome before he made the
| video.
|
| 1. https://x.com/MarkRober/status/1901449395327094898
|
| 2. https://x.com/alsetcenter/status/1902816452773810409
| labrador wrote:
| Tweets on X are not a reliable source of information. In
| fact, Grok says when asked that Elon Musk is the biggest
| spreader of misinformation on X
|
| "Elon Musk's own AI chatbot has turned on him and says he's
| spreading misinformation"
|
| https://metro.co.uk/2025/03/18/elon-musks-ai-tool-says-
| sprea...
| ggreer wrote:
| The first tweet I linked to is Mark Rober's unedited video
| of the crash. The second tweet I linked to is a video of
| someone trying to reproduce the Wile E. Coyote test. Unless
| you think the videos are faked (one of which was posted
| _by_ Mark Rober), I 'm not sure what objection you're
| making.
| labrador wrote:
| My objection is Elon Musk and Tesla superfans will go to
| great lengths to spin events in Musk and Tesla's favor
| and X is their mouth piece. I looked at the replies under
| Mark Rober's video and it's the typical flood of Musk and
| Tesla super fans raging at him. Someone needs to explain
| why Mark Rober would post a misleading test. He seems
| like a solid guy. People I respect follow him, such as
| Palmer Luckey, Leopold Aschenbrenner and Andrej Karpathy.
|
| Let's get back to my main point, that Tesla's not having
| Lidar is stupid and I don't trust a self-driving car that
| can't adequately detect solid objects in it's environment
| amelius wrote:
| Yeah, humans would crash less too if they had a Lidar that
| stopped the car if a crash appeared imminent.
| floxy wrote:
| Looks like 52% of the time:
|
| https://www.mitre.org/news-insights/news-release/parts-annou...
| wnissen wrote:
| Serious crash rates are a hockey stick pattern. 20% of the
| drivers cause 80% of the crashes, to a rough approximation. For
| the worst 20% of drivers, the Waymo is almost certainly better
| already.
|
| Honestly, at this point I am more interested in whether they can
| operate their service profitably and affordably, because they are
| clearly nailing the technical side.
|
| For example data from a 100 driver study, see table 2.11, p. 29.
| https://rosap.ntl.bts.gov/view/dot/37370 Roughly the same number
| of drivers had 0 or 1 near-crashes as had 13-50+. One of the
| drivers had 56 near crashes and 4 actual crashes in less than 20K
| miles! So the average isn't that helpful here.
| dangus wrote:
| I saw a transit enthusiast YouTube video try out Waymo from the
| most distant part of the network to fisherman's wharf in SF and
| it cost twice as much as an Uber while having a longer wait
| time for a car.
|
| It also couldn't operate on the highway so the transit time was
| nearly double.
|
| One shouldn't underestimate how economical real human operators
| are. It's not like Uber drivers make a ton of money. Uber
| drivers often have zero capital expense since they are driving
| vehicles they already own. Waymo can't share the business
| expense of their vehicles with their employees and have them
| drive them home and to the grocery store.
|
| I'm sure it'll improve but this tells me that Waymo's price per
| vehicle including all the R&D expenses must be astronomical.
| They are burning $2 billion a year at the current rate even
| though they have revenue service.
|
| Plus, they actually have a lot of human operators to correct
| issues and talk to police and things like that. Last number I
| found on that was over one person per vehicle but I'm not sure
| if anyone knows for sure.
| fossuser wrote:
| The wait times have gotten better, they're getting freeway
| approval shortly which will be nice, the price is still at a
| premium (but worth it imo). I only take Waymo in SF now.
|
| The only time I take Uber in the bay area is to the airport
| (and when they approve Waymo for SFO I won't take Uber then
| either).
| BurritoAlPastor wrote:
| I generally find that Waymos are _cheaper_ than Uber /Lyft
| _including tip_.
|
| I've also seen that, although Uber and Lyft peak times seem
| correlated to each other, they seem uncorrelated to Waymo
| peak activity. But this might be stabilizing as Waymo
| ridership increases.
| fossuser wrote:
| Ah I guess I almost never tip uber (unless something
| exceptional happens) - holdover from the Travis era.
| orangecat wrote:
| _I saw a transit enthusiast YouTube video try out Waymo from
| the most distant part of the network to fisherman's wharf in
| SF and it cost twice as much as an Uber, had a longer wait
| time for a car, and cost about double._
|
| That's literally an edge case. For shorter trips, I've found
| it to be slightly cheaper (especially factoring in the lack
| of tips) with maybe a slightly longer wait.
| agildehaus wrote:
| Both the longer wait time and the double price can likely be
| explained by the lack of highway.
|
| Highway is coming.
|
| And scale will make it cheaper. It's only cheaper than Uber
| sometimes currently. That will change.
| VirusNewbie wrote:
| In LA, wait times were the same as Uber and the price was the
| same as well (for a nicer car some of the time).
| whyenot wrote:
| My experience using Waymos in SF is that they are a little
| less expensive than an Uber. The other advantage is that you
| aren't stuck with a driver who hits on you or wants to share
| his opinions on the best way to slaughter goats.
| jonathantf2 wrote:
| Miles?
| radpanda wrote:
| Waymo rides are also potentially slower because they strictly
| follow speed limits. Not really problematic in downtown SF
| but it'll be interesting to see how it'll be received by
| riders when they expand to highway driving where most people
| generally expect to drive over the speed limit.
| thot_experiment wrote:
| On most trips people do speeding saves an irrelevant amount
| of time. If somehow you encounter zero traffic from Palo
| Alto to SF and you go 15mph over the limit the whole way it
| makes the trip about 5 minutes shorter.
|
| You have about 50% more KE at 80mph as you do at 65mph btw,
| if you find yourself needing to dissipate that energy
| rapidly.
| radpanda wrote:
| Sure, there's the math, but there's also the human nature
| part of it. If you're sitting in the right lane doing the
| speed limit, watching dozens of cars consistently zip
| past, it _feels_ like you're "falling behind" all of that
| traffic. I wonder how that will be received by the riding
| public.
| michaelmrose wrote:
| What about the benefit to the 80% if the 20% were obligated to
| use software instead of their own wetware in a hypothetical
| world where this was feasible in all respects. Imagine if you
| transitioned to most new drivers for instance being issued only
| permits to use self driving vehicles and older drivers being
| obligated to switch at 65.
| nickvec wrote:
| Interesting. Pareto principle in action.
| https://en.wikipedia.org/wiki/Pareto_principle
| ra7 wrote:
| This article neatly summarizes Waymo's latest safety numbers, but
| Waymo's site provides much more detail, including a full
| breakdown of their comparison methodology:
| https://waymo.com/safety/impact/
| jedbrooke wrote:
| purely anecdotal: one thing I've noticed is that Waymo's ALWAYS
| use their turn signal. that's already gotta put them above a
| large portion of human drivers in terms of safety ;)
| dangus wrote:
| You know what has even better safety statistics? Public transit
| including buses and trains.
|
| Cycling also has better statistics when the infrastructure is
| given real attention, and it leads to better health outcomes.
| AlotOfReading wrote:
| Those are all options that local governments should consider
| investing in. Private companies that don't take public money
| aren't the ones that should be building public transit.
|
| Given that Waymo is testing in Tokyo now (and formerly Cruise),
| autonomous vehicles are clearly not mutually exclusive with
| well funded public transit.
| paxys wrote:
| Worth repeating the same comment I've left on every variant of
| this article for the last 10 years.
|
| Being better than "average" is a laughably low bar for self-
| driving cars. Average drivers include people who drive while
| drunk and on drugs. It includes teenagers and those who otherwise
| have very little experience on the road. It includes people who
| are too old to be driving safely. It includes people who are
| habitually speed and are reckless. It includes cars that are
| mechanically faulty or otherwise cannot be driven safely. If you
| compile accident statistics the vast majority will fall into one
| of these categories.
|
| For self driving to be widely adopted the bare minimum bar needs
| to be - is it better than the average _sensible_ and
| _experienced_ driver?
|
| Otherwise if you replace all 80% of the good drivers with waymos
| and the remaining 20% stay behind the wheel, accident rates are
| going to go _up_ not down.
| kccqzy wrote:
| What kind of dataset do we have to determine the subset of
| accidents caused by _sensible_ and _experienced_ drivers?
|
| I personally have doubts as to whether this dataset exists.
| Whenever there's an accident, and one party is determined to be
| at fault, would that party be automatically considered not to
| be a sensible driver?
|
| If we don't have such a dataset, perhaps it would be impossible
| to measure self-driving vehicles against this benchmark?
| floxy wrote:
| Why wouldn't alcoholics and the elderly be early adopters of
| self-driving vehicles. Or what can we do to encourage them to
| be early adopters? You get a DUI, and you are forced to pay for
| FSD? Get a reduce rate on booze taxes if you "drive" an AV?
| Have to take a driving test every 2 years after you turn 75,
| unless you have an AV?
| paxys wrote:
| "XYZ demographic should be forced to use self driving cars"
| is a fantasy that the tech crowd continues to believe but
| will never happen. Everyone is able to drive and will
| continue to be able to drive. In fact you should assume that
| the worse someone is at driving the more likely they are to
| want to drive for themselves, because that's how the world
| usually works.
| aithrowawaycomm wrote:
| Even in self-driving, Telsa's behavior proves there is a
| market for cars that are programmed to speed and roll
| through stop signs. Waymos are safer than the average
| human, but the average human also intentionally chooses a
| strategy that trades risk for speed. Indeed, Waymo trips on
| average take about 2x as long as Ubers:
| https://futurism.com/the-byte/waymo-expensive-slower-taxis
|
| What happens if an upstart self-driving competitor promises
| human-level ETAs? Is a speeding Waymo safer than a speeding
| human?
| senordevnyc wrote:
| Over the next decade or two, insurance will solve a chunk
| of this problem (it'll be way more expensive to drive
| yourself), regulations will solve another chunk, but the
| biggest thing that will solve it: we're lazy.
|
| We might drive every now and then, but come on, do you
| really think once this is ubiquitous and you can get in a
| (or your) car and then play a video game, take a nap, text
| on your phone, or doomscroll, that we're still going to
| want to drive all the time?
|
| Nah.
| Retric wrote:
| There's no statistics for how much a sensible and experienced
| driver crashes.
|
| Sorting people by past behavior runs into survivorship bias
| when looking back and people who stop being sensible going
| forward. I'm personally a poor driver, but I don't drive much
| so my statistics still look good.
| IneffablePigeon wrote:
| There's _no_ statistics? Surely those statistics are
| precisely what all car insurance premiums are based upon.
| They might be proprietary but I am certain such statistics
| exist.
| senordevnyc wrote:
| This is already much better than average, enough that it's
| going to take over. No one cares about car accidents in this
| country enough to stop this, even if it was only slightly
| better than average. For proof: Tesla.
|
| Self-driving being the dominant form of driving is now a done
| deal, thanks to Waymo (and probably Tesla, though that's a
| policy failure imo), it's just a question of how long it takes.
|
| My money is on a decade.
| D-Coder wrote:
| > Being better than "average" is a laughably low bar for self-
| driving cars. Average drivers include people who drive while
| drunk and on drugs. It includes teenagers and those who
| otherwise have very little experience on the road. It includes
| people who are too old to be driving safely. It includes people
| who are habitually speed and are reckless... (etc)
|
| But... that's the reality. If we replace human drivers with
| self-driving cars at random, or specifically the bad drivers
| above, then we've improved things.
|
| We are not going to easily improve the average human driver.
| ein0p wrote:
| This is very cool if it's not the case of "lying with
| statistics". I hope we will at some point see independent
| research into these safety claims. I don't really think too
| highly of my own driving skills, so I also hope that I live to
| see the day when I can just let my own car drive itself on most
| or all trips. And no, Tesla FSD ain't that, at least not yet - it
| does help with long trips, but in difficult situations it
| disengages at the worst possible moment. I want to tell the car
| where to go, and then never touch the wheel until I'm there, day
| or night, rain or sun.
| xlbuttplug2 wrote:
| Off topic, but I've always felt that the ideal way to address
| safety is by limiting the amount of freedom a vehicle has, i.e.,
| force it to follow a predefined path. That is, replace road lanes
| with something like tram tracks. (I'm imagining vehicles being
| two/four person pods zipping around.)
|
| This would vastly reduce the number of accident scenarios, be
| more efficient, and be much easier to automate. And would
| probably be good enough for 99% of use cases (i.e., work
| commute).
|
| Obviously I don't seriously see anyone splurging on the
| infrastructure and bespoke vehicles for that. But I can dream.
___________________________________________________________________
(page generated 2025-03-26 23:01 UTC)