How often do driverless cars crash into each other? Seems like a better proxy for the future, i.e. if there were no human drivers would the crash rate be zero?
It's always how vehicle safety data is calculated.
With human drivers it matters--high mileage drivers are more likely to be driving tired or in poor conditions, low mileage drivers often are losing their abilities. Thus you see elevated risk on the ends of the curve. That doesn't apply to machines, though.
This study compared national averages of Waymo to hyper-localized NYC cab drivers. NYC, where the maximum speed limit on streets in 25mph, should rightfully be viewed as an outlier. “ During the period analyzed, New York City's for-hire fleet was involved in 16 crashes that resulted in serious injuries or fatalities, while Waymo recorded three nationwide “
Their argument is that comparisons against the driving record of cab drivers better represents what they should expect from waymo vs comparing them against every kind of driver everywhere. Considering the driving skills of some cabbies I've ridden with I'm surprised that cab drivers come out ahead of the average driver, but I'd agree that a lot of that is probably thanks to fewer highway miles.
Waymo compares its safety record with that of the average human driver. The study says New York City's for-hire vehicles are involved in serious or fatal crashes at a rate that is 14 times lower than the average vehicle, making them a more appropriate benchmark for robotaxis.
...
The researchers compared Waymo's nationwide crash data with New York City's for-hire fleet and found the robotaxis had a serious injury or fatal crash rate about 1.5 times higher per mile driven.
I looked through the report. It’s a bad report but it couldn’t have been done well because of not having apples to apples data. They compare data from very different environments (all CA Waymo vs NYC taxi) and just kind of shrug aside that difference.
"safety" is usually the first point out of the mouth of a self-driving advocate. If safety = preservation of life, I wonder how truly "safe" full Waymo supremacy would be when there are millions of Americans who drive for a living [1], in a country where working age adults need a full-time job to get basic healthcare.
Through that lens, I honestly don't know if Waymo's future would lead to preservation of life. I actually think it might increase physical harm to Americans, only that harm is laundered out of our roadways and into our failing healthcare system.
This is an important aspect of any tech that threatens whole sectors of employment.. "safety" is always thrown into question when you cut & consolidate economic lifelines without a societal safety net. If it happens to truckers it can happen to you, too.
You can say the same for most of the technology progress. For example, self-propelled combine harvesters caused mass migration of Americans in 1940s from rural areas to cities because they were out of job. With your logic, it should have not happened.
The better approach is to ask government to come up with a solutions that does not involve limiting science and tech.
Should we have also not cheered on electric computers replacing all the expensive, skilled mathematician/computing jobs that used to exist? Keypunch operators? Telephone switchboard operators? Advanced printers replacing the vast majority of the typesetting work and tons of maintenance involved with early printers? Bowling Pinspotters? Tollbooth collectors?
Should we smash the machines in the factories and go back to everything being handmade? Get rid of the pick-and-place machines for our electronics and have everything hand-soldered on hand-drawn boards? Think of the jobs!
I hear these new-fangled steam locomotives can move as much freight as hundreds of horse-drawn carts. That's hundreds of jobs being replaced by just a few on the train. Think of the jobs for all those horse drivers, all the barn workers, and more!
There are tons of jobs that used to exist that don't anymore. Many were very common. We used to have to pay people to go light all the street lights. Can you imagine having to pay an army of people to walk all the city streets every night to go light all the lights and every morning to extinguish them?
To be fair, I am a bit of a skeptic of driverless cars taking over 100% of the driving industry. I think as long as roads are open to the public there will be some humans behind the wheel. There's probably just going to be enough "odd" kinds of transportation needs that don't map well to the automation that cover 90% of people and trips just fine.
Wow, heckin' epic comeback good sir, upvoted and take my Reddit gold. Never heard it before. I'm just saying it's awfully convenient to the owner class that silicon valley wants to speedrun "full autonomous takeover" while politicians saw away at individuals' economic security.
I would LOVE an autonomous utopia but if Waymo wants do the "autonomous" without the social safety net, they can spare me the BS about this saving lives and admit that their automation is about the cold, hard savings of eliminating human workers
I fundamentally agree with you, but at the same time I don't want the answer to be preserving labor sinks when the answer could instead be more fairly distributing the value produced by labor so that we just don't need to work as much in the first place.
Waymo certainly can't as currently designed. Their system must have a very detailed mapping of the situation and it has a fallback of showing a human operator somewhere the situation and asking for the right course of action. (This does not violate level 4. A level 4 car must be able to safely resolve anything a human could, but it is allowed to stop and ask how to proceed.)
How often do driverless cars crash into each other? Seems like a better proxy for the future, i.e. if there were no human drivers would the crash rate be zero?
Do they have more accidents in any sorts of scenarios where humans are less likely to?
They are not without their faults. But it's hard to deny that self driving cars enable a much safer space for us to live.
It would be great to see an open framework for the cars to become aware of each other, self driving or not.
> Per vehicle mile traveled (VMT),
does scale matter at all though?
do crashes stay constant per miles driven
Its a common way of analyzing these kinds of statistics especially on large scales.
What other divisor makes sense to you?
same divisor as human comparision.
let me ask you the opposite, what divisor(if any ) would make you question it.
2 replies →
Why in the world wouldn't they?
It's always how vehicle safety data is calculated.
With human drivers it matters--high mileage drivers are more likely to be driving tired or in poor conditions, low mileage drivers often are losing their abilities. Thus you see elevated risk on the ends of the curve. That doesn't apply to machines, though.
wouldnt we mass approve these everywhere if that was true then?
2 replies →
[dead]
Unless those people are also driving taxis maybe.
Waymo Robotaxis Record Higher Serious Injury and Fatal Crash Rate Than NYC's For-Hire Fleet (https://www.yahoo.com/news/science/articles/waymo-robotaxis-...)
It may also be worth noting that waymo doesn't report all crashes to the DMV and they went to court to keep crash data secret (https://techcrunch.com/2022/02/22/waymo-to-keep-robotaxi-saf...)
This study compared national averages of Waymo to hyper-localized NYC cab drivers. NYC, where the maximum speed limit on streets in 25mph, should rightfully be viewed as an outlier. “ During the period analyzed, New York City's for-hire fleet was involved in 16 crashes that resulted in serious injuries or fatalities, while Waymo recorded three nationwide “
That "study" is even more misleading than that, the Waymo-involved incidents were in no way Waymo's fault.
Their argument is that comparisons against the driving record of cab drivers better represents what they should expect from waymo vs comparing them against every kind of driver everywhere. Considering the driving skills of some cabbies I've ridden with I'm surprised that cab drivers come out ahead of the average driver, but I'd agree that a lot of that is probably thanks to fewer highway miles.
Waymo compares its safety record with that of the average human driver. The study says New York City's for-hire vehicles are involved in serious or fatal crashes at a rate that is 14 times lower than the average vehicle, making them a more appropriate benchmark for robotaxis. ... The researchers compared Waymo's nationwide crash data with New York City's for-hire fleet and found the robotaxis had a serious injury or fatal crash rate about 1.5 times higher per mile driven.
1 reply →
I looked through the report. It’s a bad report but it couldn’t have been done well because of not having apples to apples data. They compare data from very different environments (all CA Waymo vs NYC taxi) and just kind of shrug aside that difference.
If they weren't motivated to be anti-Waymo, an invalid "study" like that wouldn't be published at all.
"safety" is usually the first point out of the mouth of a self-driving advocate. If safety = preservation of life, I wonder how truly "safe" full Waymo supremacy would be when there are millions of Americans who drive for a living [1], in a country where working age adults need a full-time job to get basic healthcare.
Through that lens, I honestly don't know if Waymo's future would lead to preservation of life. I actually think it might increase physical harm to Americans, only that harm is laundered out of our roadways and into our failing healthcare system.
This is an important aspect of any tech that threatens whole sectors of employment.. "safety" is always thrown into question when you cut & consolidate economic lifelines without a societal safety net. If it happens to truckers it can happen to you, too.
[1] https://www.census.gov/library/stories/2019/06/america-keeps... This article lists 3.5 MILLION truckers, not even counting rideshare drivers and other CDL holders.
You can say the same for most of the technology progress. For example, self-propelled combine harvesters caused mass migration of Americans in 1940s from rural areas to cities because they were out of job. With your logic, it should have not happened.
The better approach is to ask government to come up with a solutions that does not involve limiting science and tech.
Should we have also not cheered on electric computers replacing all the expensive, skilled mathematician/computing jobs that used to exist? Keypunch operators? Telephone switchboard operators? Advanced printers replacing the vast majority of the typesetting work and tons of maintenance involved with early printers? Bowling Pinspotters? Tollbooth collectors?
Should we smash the machines in the factories and go back to everything being handmade? Get rid of the pick-and-place machines for our electronics and have everything hand-soldered on hand-drawn boards? Think of the jobs!
I hear these new-fangled steam locomotives can move as much freight as hundreds of horse-drawn carts. That's hundreds of jobs being replaced by just a few on the train. Think of the jobs for all those horse drivers, all the barn workers, and more!
There are tons of jobs that used to exist that don't anymore. Many were very common. We used to have to pay people to go light all the street lights. Can you imagine having to pay an army of people to walk all the city streets every night to go light all the lights and every morning to extinguish them?
To be fair, I am a bit of a skeptic of driverless cars taking over 100% of the driving industry. I think as long as roads are open to the public there will be some humans behind the wheel. There's probably just going to be enough "odd" kinds of transportation needs that don't map well to the automation that cover 90% of people and trips just fine.
Wow, heckin' epic comeback good sir, upvoted and take my Reddit gold. Never heard it before. I'm just saying it's awfully convenient to the owner class that silicon valley wants to speedrun "full autonomous takeover" while politicians saw away at individuals' economic security.
I would LOVE an autonomous utopia but if Waymo wants do the "autonomous" without the social safety net, they can spare me the BS about this saving lives and admit that their automation is about the cold, hard savings of eliminating human workers
3 replies →
> I think as long as roads are open to the public there will be some humans behind the wheel.
Pretty much every car will have a human driver. Autonomous vehicles don't work.
Waymo's cars are driven remotely by humans.
1 reply →
I fundamentally agree with you, but at the same time I don't want the answer to be preserving labor sinks when the answer could instead be more fairly distributing the value produced by labor so that we just don't need to work as much in the first place.
If health care is the only problem, then Medicare for all is the solution. We can have better health care and possibly fewer automobile deaths.
It is yet to be seen whether waymo can really drive everywhere that humans drive, though.
Waymo certainly can't as currently designed. Their system must have a very detailed mapping of the situation and it has a fallback of showing a human operator somewhere the situation and asking for the right course of action. (This does not violate level 4. A level 4 car must be able to safely resolve anything a human could, but it is allowed to stop and ask how to proceed.)
It never ceases to amaze just how many logical backflips Americans will do to avoid saying “our healthcare system is shit and must be fixed”.
If you work at it, I’m sure you’ll find a justification that Americans should all stop breathing because it’s bad for health insurance profits !
To misquote Milton Friedman: "Oh, I thought you were trying to have a taxi service. If it's jobs you want, we should have 2 drivers per car."
Prop up the existing wage-slavery system because anything else would cause people to not have healthcare coverage? It’s shockingly unimaginative.
Anyways, my last Uber driver was playing games on his iPad while driving. Did a good job of hiding it at first.