A study last month done by Swiss Re for Waymo suggests that self-driving Waymo cars are far, far safer than human drivers:
I have long recommended that if you want to evaluate the state of autonomous cars, look at Waymo. Forget about Tesla or Cruise or any of the others. They may do certain things well or poorly, but Waymo is the state of the art.
Note that this chart is based solely on miles driven with no human behind the steering wheel. It is 100% autonomous driving. And as many others have already pointed out, autonomous cars will get better every year. Humans won't.
But this does make me wonder: When are we all going to be able to buy driverless cars using Waymo technology? At this point, Waymo:
- Works on city streets.
- Works on highways.
- Has been tested in San Francisco, Austin, Phoenix, and Los Angeles.
- Is able to drive, park, let you off at the curb, and reliably take instructions about where to go.
- Is extremely safe.
So what's left? I suppose there's still work to be done to make the radar stuff smaller and cheaper, but is that all? It sure seems like Waymo is very, very close to being able to sell this stuff on the consumer market.
I want one.
Some points:
- In Phoenix, Waymo does not take highways.
- The Waymo I was in took a right turn going around a car on its right also trying to take a right turn.
- It seems to have extreme difficulty with heavy traffic. One parked itself in the middle of traffic in the middle lane. Ours took a few minutes navigating past it on the right, practically blocking traffic in two of three lanes.
- It has some difficulty with pedestrians in the middle of heavy traffic. It scared pedestrians crossing in front while it took its time coming to a full stop on a red light.
- It does bizarre things like avoiding flashing police cars by diverting around the block.
- When it comes to stops, it has this odd habit of slowing down late then hitting the pedal twice to cause the vehicle to jerk forward.
but no reported bodily injury or property damage.
😉
Apparently that’s all that matters according to the chart above. If the waymo goes 10 miles out of its way to avoid all other cars and pedestrians and comes to jerky stops at every intersection… well, it’s safer than a human driver!
That WAS the question that was asked!
Want a different answer, ask a different question. But this definition of safety is what obsesses the public...
"- It does bizarre things like avoiding flashing police cars by diverting around the block."
Sounds reasonable to me. Just recently, I turned into a street and saw a fire engine, lights flashing at the other end of the block. So I turned around mid-block and picked another route. Google Maps adjusted to the changed circumstances on its own.
We are looking for unemployed or part-time boys and girls for data entry jobs. The shift lasts from 8:00 a.m. to 1:00 p.m.and from 1:00 p.m. to 6:00 p.m. – public holidays from Saturday to Sunday. You will work from home. Training will be provided. ee data is entered into online portals. The salary for this position is $30,000.
Join the team here……. http://datawork881.blogspot.com
It's amazing that cities allow these things to operate. What is in it for the city?
Publicity and improved Brand. Cheaper fares for there citizens. Free public transit for the handicapped. Possibly fewer traffic accidents. Jobs working for the companies.
It's almost like not EVERY locality in America wants to strangle itself to death with bubble wrap...
Maybe in the Southwest, but there are too many other kinds of topography and weather it's never been tried in.
What kind of topography are you thinking of? As far as I know, there are basically flat roads that are sometimes straight and sometimes curvy, and then there is San Francisco.
Snow, ice. Not actually topography but I don't think waymo has ever dealt with this issue. They've avoided place with lots or rain as well.
The upper Mississippi valley has a very crabbed topography of small, but very steep, irregular valleys with no obvious pattern to them, with old narrow and eccentric streets that wind around the hillsides, often with hairpin turns, perhaps multiple hairpin turns, essentially no grid layout of the streets at all; the possibilities of tornadoes, flash floods, derechos, blizzards, trees fallen across the streets, garbage cans and children's toys blowing around up and down the hills, whether there's wind or not, children playing, ice storms, a declining but still prevalent level of alcoholism, and wandering dingbats.
I think this will be the last place Waymo will get their cars working.
And in the surrounding countryside it can be even more eccentric with all of the above plus gravel roads, blacktops with no markings and very poor maintenance and driveways, some very long and effectively dirt tracks with weeds growing in them and moo-cows roaming around.
And there are deer everywhere.
Oh, yes, now we see the true meaning of Trust the Science when it gives results people don't like...
Waymo STILL struggles under some circumstances. For example,
- under difficult weather (snow or rain) where the technology can't clearly see lane lines.
- responding to unique situations such as road work that requires one to break traditional driving rules or, for example, a policeman giving hand signals to direct traffic.
- the technology seems to struggle under heavy traffic.
Similar to many newish technologies, under some circumstances Waymo fails
Thankfully, most American cities resemble Phoenix, LA or Austin in terms of weather, topography and density.
Heh. Commenting from St. Paul.
Waymo was tested in San Francisco. If you want to say waymo wasn't tested in the Snow and rain, say it. But don't gaslight us about hills and density. No American city looks like Bangkok or Hong Kong.
"If you want to say waymo wasn't tested in the Snow and rain, say it." I said that the technology struggles in rain and snow (versus not tested).
I never claimed Waymo was not tested on hills.
You can find article talking about Waymo stopping (not breaking down) under heavy traffic.
A good start--but beware of statistics of low numbers....and the amount of time on the road for Waymo is still low.
these stats are with new correctly maintained hardware; in kevin's shiny new future we'll all be sharing the road with really cheap and poorly maintained versions of these systems
imagine a world filled with Lada FSD or diy installers putting god knows what in their vehicle
my car puts up phantom error codes up on the dash every so often; i suspect related to a glitchy wheel speed sensor
the sensor is basically a magnet; no moving parts, so should be pretty reliable, and yet...
now look at an engineering diagram for these self-driving systems and think about all of the possible failure modes
imagine driving a ten year old waymo and a sensor/connector/voltage problem pops up when you're rolling up on a school crossing or you're on the 405 with people riding six inches from your bumper
The cars will take themselves to the mechanic. They will be better maintained than the current fleet.
that would be great!
hopefully waymo and melon musk are also working hard on a 'full self-financing' feature that automatically pays for those repairs
I wrote a blog piece back in 2015 with various questions I wished tech journalists would ask about driverless cars. One of them touched on this. I have yet to see the question asked, much less answered.
asking impertinent questions like that tends to piss off future employers/advertisers
Is this accurate? SF Chronicle reports ("Here’s how many Waymo and Cruise vehicles have been in crashes in past 2 years") 150 Waymo crashes, and the study you linked says they've done 35 million miles = 4.3 accidents / million miles, property damage. Very different figures.
Why would you expect the safety record of driverless cars to improve? Sure the best version of AI tech will improve, but today the companies are in test phase, with lots of extra highly trained staff, continuous expert data analysis of incidents, etc.
Will companies make a safer driverless car system? Certainly it would be more profitable to run a low cost system--fewer, cheaper sensors, cheaper staffing, keeping the accident record a trade secret, etc. Certainly 'as good as human' is good enough, right? Or 'plausibly as good'?
The Pinto was cheap to make, too. Until it wasn’t.
The Pinto wasn’t as uniquely dangerous as it was made out to be.
If true, that actually strengthens kkseattle's point.
It weaken's seattle's point. The pinto wasn't all that dangerous, but Ford stopped manufacturing it anyway because the public perception of being unsafe made it unprofitable.
Ah! The poor Corvrair!!!! On the many trips down 101 to the inlaws and back to the Bay Area it was common to see a VW on its head. I never saw a Corvrair on its head.
I never saw a VW on its head...
We'd need to build out the infrastructure for driverless cars...something to help designate lanes in bad weather, etc.
I’m sure all jurisdictions will have the spare money to do this, especially the ones who currently have trouble repairing potholes, repainting yellow lines, replacing street lights etc.
The study Kevin links to looks at insurance claims, so it only includes accidents where Waymo's insurance carrier had to pay out. If Waymo driverless cars really are safe, it makes sense that in most accidents involving a Waymo driverless car, the other vehicle will be at fault.
Why would you expect the safety record of driverless cars to improve?
Why would you expect driverless systems to be the exception to the "technology usually improves as time goes by" rule?
I was watching one of my streaming services and noticed how bad the video was. As an automobile moved against a background of trees, I could see the shimmering outline of the vehicle as the compression technology struggled with the areas of the picture where the pixels were not static.
I had the same experience watching figure skating over-the-air and noticed that the advertisements on the walls around the rink only came into focus when the camera stopped panning to follow the skater.
Is the technology improving? Probably. Are the customers experiencing that improvement? I say, "No."
I realize that we are squeezing more content in the the same transmission channel but the end result is a product that is unpleasant to watch. Similarly, don't trust the corporations to funnel their technological improvements into better self-driving cars rather than more and cheaper cars that line their pockets.
In the Mel Gibson version of _Mutiny on the Bounty_, there is a scene where a flickering candle creates artifacts. That's very high resolution film where the only compression technology is periodic sampling. Plus wheels that look like they are spinning backwards.
Elon dropped LIDAR to save money and complexity, making the Muskian argument that humans get by with passive optical sensors, computers should to. Meanwhile Tesla’s keep running into things.
A large selling point of driverless cars is/will-be the improved safety.
The Japanese car makers could have made cheaper cars, but they wanted to capture market share, and figured out the way to do that was through increased reliability.
The NTSA sets safety standards and otherwise prevents accident records from being trade secrets. And manufacturers respond by touting their safety rating in their ads.
Every time there is a plane crash, a team descends on the crash to figure out root causes and improvements. Driverless cars will be similar.
When Japanese cars arrived in NA they were dirt cheap death traps that rusted out in months.
Saying that humans won’t get better is a bit silly. Human-driven cars will also continue to be aided by crash-avoidance technology.
Also, who would buy a Waymo? What purpose would be served by having a Waymo sit parked for 23 hours a day?
What purpose does any car serve when it's parked?
And the first market is to serve as taxis anyway.
you don’t want a driverless taxi, unless you enjoy sitting in other people’s bodily fluids.
My daughter who lives in San Francisco recently took a Waymo ride. She says she really liked it. The roads where she lives are kind of complicated but the Waymo handled it no problem. The only oddity was the car approached her apartment on the opposite side of the road. A taxi would have just crossed the (undivided) road and let her off. The Waymo passed her destination and continued until it could make a u-turn and returned to her address.
She intends to use Waymo for rides in the future.
I’m guessing waymo actually follows all traffic laws and human drivers don’t.
Sure. And in her case it was no big deal. But in a busy city, I can see times when slavishly adhering to all laws would make dropping a person off at a destination nearly impossible. I wonder how it works in such situations.
My daily commute is largely on rural roads with one lane in each direction and for the most part a double yellow line down the middle. It also frequently garbage trucks making their rounds. The way it works in practice is that if you are directly behind the truck you inch over to the left and if there is no oncoming traffic, you make a mad dash around the truck. The garbage men often will signal if there is oncoming traffic. The mad dash is illegal, but actually quite safe. The road markings assume something like highway speed. Getting around a stationary truck is much quicker. But still illegal.
Presumably a driverless car will trail behind the garbage truck, stop and go and never above five or ten miles an hour. Adding to the fun, this will make it harder for a human-driven car behind it to pass. Get a line of three or four driverless cars behind the truck and traffic will be backed up for miles.
If you have ever driven in Naples, the rule is there are no rules. If you try to follow the rules you just screw it up for everyone else who expects you not to follow the rules, and they are not shy about letting you know it.
Human drivers sometimes follow laws that they make up. It pisses me off when people can't figure out how to make a right turn on red in California when there is absolutely no cross traffic.
And all those Waymo cars are really well maintained, coddled daily by a team of caregivers.
Use them as weapons. That’s all they’re good for. Point one at trump!
Sop far nobody seems to talk about selling to consumers as an alternative to cars. The technology is pricey what with all the sensors and radar and giant computers.
To be economically feasible they need to be driving far more hours / miles than a privately owned car does. Taxis and similar public transit use will be their domain for some time to come.
I still believe that this is a waste of money and working time at a time when far more urgent problems in connections with climate change remain unsolved. I understand the fascination of the field but autonomous cars contribute zero to the solution of the climate crisis (except for the fact that they are electric but electric cars are available for human drivers).
The human population is large and humans can engage in a wide variety of activities from completely frivolous to world-saving. You can't expect everybody to focus on world-saving.
If a problem is everyone's problem, it is no one's problem.
I still believe that this is a waste of money and working time at a time when far more urgent problems in connections with climate change remain unsolved.
Won't happen tomorrow, or even in a few years. But eventually, when this technology truly matures and driverless technology is both far safer than human driving and highly affordable, it will be an absolute game-changer. I think it may be as revolutionary as the original arrival of affordable cars/trucks 100+ years ago. That development didn't just improve transportation—it revolutionized modern life in countless ways (suburbs, housing, the retail sector, work, demographics, whole new industries, shipping, goods transport, farming, land use, etc).
Yes. Too bad we can't walk and chew gum at the same time. We should always devote all resources in the entire world to solving one problem at a time.
I disagree that autonomous cars contribute zero to the solution of the climate crisis.
If (and it's a really big if) Waymo driverless taxis take off, and reduce the cost of taxi rides, more of the urban people will forego their own car and just take a taxi, or rent a regular car for the trips out of town. That would reduce the resources spent building cars that (as a commenter above points out) sit idle 23 hours of the day. Cars are big and take up space, so reducing the need for cars, and parking spaces, can make new urban development more compact and more efficient, reducing the carbon footprint of those who live there and travel by robotaxi.
The following article by Timothy Lee seems good, and it wasn't paid for by Waymo. Lee tentatively concludes that Waymo is safer than a human driver.
https://arstechnica.com/cars/2023/09/are-self-driving-cars-already-safer-than-human-drivers/
thanks for the link!
The article did mention that a minority of people are responsible for a majority of accidents. So under the right conditions, Waymo driving is better than a teenager.
😉
Honestly, nobody really cares about this chart. Given the other safety improvements that almost nobody cares about, this chart isnt going to sell anything:
- slower speeds
- lighter cars
- narrower roads
- buses
- roundabouts
- etc....
Thank you. Most of this is also the best solution to reducing the environmental impact of cars. We'd probably be better off building cars that get 150 miles to the gallon that investing so much into high performance electric cars that weigh 5 tons.
"high performance" is a direct consequence of "electric".
If a car that gets 150 miles to the gallon is viable, then the electric version of that car doesn't weight 5 tons and is also viable.
Other reporting has said that Waymo cars actually have 1.5 dedicated personel assigned to every car at all times?
Did i read that right? These self driving cars are not only not self driving but actually requure more people to operate. Oddly Kev rarely mentions this.
Waymo needs to start calling their cars AI powered to ensure Kev stays captured by the storyline.
The linked study also indicates that person driven Waymo cars with no self-driving assist are much safer than regular humans.
The study also appears to say that 100% autonomous driving Waymos are not much safer than human driven Waymos.
Kevin appears to be using the 'human in the drivers seat to take over in case of an emergency' safety stats for Waymo....but hes calling this '100% autonomous'?
Hmmm, curious. When youre committed to a narrative.....
Waymo doesn't have a human in the diving seat. The 1.5 people you are talking about ae programmers working off line to get the maps the vehicles have updated and to fix the code when problems are found. If you increase the number of cars by a factor of 10, and keep them driving in the same area, the number of programmers won't change.
You are wrong.
The insurance study that Kevin links to says that Waymo uses (or has used during thr study) 3 basic setups:
1. Human drivers behind the wheel and making all decsions in the Waymo cars
2. Human drivers in the car assisting the autonomous or self driving car
3. Fully Self driving cars
In all of these cases, additional staff is back at Waymo HQ monitoring the cars as they drive around.
There are certainly programmers and various admin staff that wont grow if Waymo scales.....but there is also very large team of people currently working on/monitoring/assisting the Waymo cars as they move around.
Which other reporting?
This site permits links. Definitely avail yourself of the opportunity.
I have seen this reporting, but I think it was Cruise, not Waymo.
I think youre correct. Kevin has asked us to try to forget about Cruise which we have the most data and reporting on.
The focus of Kevins post is a Waymo funded study and related promotional material. Which may be completely accurate and a good representation of results.
I think we can add waymo to this list. Why does Mr. Drum want to lose elections so much?
Today, there are a multitude of factors that have driven working-class voters out of the Democratic Party. They include:
• Democrats’ support for trade deals that led to factory closings in many small towns and midsize cities in states that were once Democratic strongholds.
• Democrats’ support for spending bills that the working and middle classes paid for but that were primarily of benefit to poor Americans, many of whom were minorities.
• Democrats’ enthusiasm for immigration of unskilled workers and the party’s opposition to measures that might reduce illegal immigration.
• Democrats’ support for strict gun control.
• Democrats’ insistence on eliminating fossil fuels.
• Democrats’ use of the courts and regulations to enforce their moral and cultural agenda, whether on the sale of wedding cakes or the use of public men’s and women’s bathrooms.
https://www.liberalpatriot.com/p/where-have-all-the-democrats-gone-989
Gun control is objectively good. Eliminating fossil fuels is necessary to prevent societal collapse. And it is right and proper to use the courts to enforce a moral and cultural agenda, just as it was in the 60s.
And imagine saying "Democrats helping the poor is bad."
So would it be right and proper for republicans to use the courts to enforce their moral and cultural agenda?
What are the substantive merits of that moral and cultural agenda, and are the courts an appropriate venue for it based on those merits?
The law enters into it, of course; the courts sometimes aren't an appropriate venue even if the merits are strong. But many issues regarding the culture wars are about people substantive civil rights and liberties, and the courts are absolutely a proper place for that.
If their agenda was moral, it would be.
"Democrats’ support for trade deals that led to factory closings in many small towns and midsize cities in states that were once Democratic strongholds." Factories closed because of automation, Democrats had nothing to do with that. The claim that trade deals hurt employment was always deliberate and intentional lie. And we're presently near and all time high in jobs relative to working age population. (Only higher during Clinton's 2nd term in office.).
It would appear:
https://www.hks.harvard.edu/sites/default/files/Taubman/AVPI/Testing%20and%20Policy%20Basic.pdf
that most of the WAYMO 3 million RO miles are in Arizona.
Indeed this appears to be the only place where RO operation commercially its allowed.
So:
- WAYMO seems to be the leader
- there is a reasonable question on whether the RO statistics are as explanatory as suggested
That set of slides appears to be from 2018. Waymo is also operating commercially in San Francisco.
You won’t be able to buy one. Waymo cars are loaded with over $100,000 of gadgetry that is bolted on (not integral) to the car. They are geo-fenced to specific urban areas, and there is no way to make them function outside those areas. I don’t think they can go on highways. There are 1.5 people per car remote-supervising them at all times, and they still have lots of issues. Even if they worked great and didn’t need remote supervision, there is no way to scale them for sale to consumers. The best you can hope for are more Waymo taxi fleets in cities.
The only viable approach to self driving cars so far is Tesla’s approach. But even Tesla, with zillions of miles of driving data from millions of Teslas on the road, and (by the end of next year) with more compute power than any other company, may not get it to the point where it can handle ALL edge cases better than humans. I have one, and the current FSD Beta is astounding. On the highway it is flawless. It really is, but it is not perfect yet on other roads (and probably not on highways if something really weird happens.) Given all the data and compute they have, it will be interesting to see if their machine learning approach will get them to where the company can assume responsibility for whatever happens when FSD is engaged. It is already better than a typical human driver most of the time, but it will have to be a lot better than human ALL the time. And who knows how many people will actually pay an extra several thousand dollars for FSD? If the tech works well enough maybe the end result will be robot-taxi networks in reasonably populated areas providing reliable transportation at a lower cost per mile than owning a car. This will get drunk or very elderly drivers out from behind the wheel, and perhaps eliminate the need for, say, empty nest couples to buy a second car. We will see.
You won’t be able to buy one. Waymo cars are loaded with over $100,000 of gadgetry that is bolted on (not integral) to the car.
Not in the 2020s, probably. Eventually, though, they'll either A) be cheap enough for ordinary folks to buy or B) they'll be so cheap most people won't want to own one, but will just subscribe to a service—like having a robot Uber on call 24/7 that almost always arrives at your residence in under 10 minutes.
The number of incidents is not the point. The point is that when something does go wrong, it is impossible to assign the blame. And that is also the purpose.
Wow, are you clueless. When something goes wrong it is far, far easier to assign blame with self driving cars.
What fraction of traffic accidents are caused by inattention or bad decision making?
A quick survey of the one's I've been involved in gives me 100 percent, at a count of maybe 30. (I'm counting parking lot scrapes, mind you.) Computers really will never fail because of inattention.
Stacked against that are the two times which I have done something unusual on the highway, which prevented me from being involved in an accident. There's no guarantee a Waymo wouldn't have avoided it too. However, both of them happened on the freeway, so as yet, a Waymo wouldn't have been there anyway.
Living in Mountain View, I encountered Waymo vehicles on the streets all the time. They drive like someone who is alert, but super cautious. They never speed or do anything risky. Safety is not going to be a reason to avoid them. But "gawd they are slow" might be. Human beings really like having control of the vehicle, and it's going to be a slow process acclimatizing them to doing something else.
> Human beings really like having control of the vehicle, and it's going to be a slow process acclimatizing them to doing something else.
Baloney. See, I can assert stuff too without proof. Humans like having someone else paying attention to the driving so that the human can entertain themselves with something fun or profitable.
How can Kevin compare Waymo running in cities that are perpetually sunny, perfectly-mapped, well-marked versus to human drivers where it rains and snows, where the roads are covered with fallen leaves, where the traffic markings have been removed by crumbling roads or where the roads are continually under construction thanks to snowplows, sand and snow-chains?
I know he loves figures and charts but he keeps falling for lies and damn lies even when he mentions the issues behind the numbers.
Liability.
Waymo might or might not be very safe, and it might or might not get safer. But it WILL fuck up. There will be accidents. Damage will happen. People will get hurt or die.
Who will be responsible for this? The answer cannot be "nobody."
Now, my answer is be "Waymo is. Their system, they're advertising it as useful and safe, people are using it under that assumption. The law should thus be 'if a Waymo car hops a curb and flattens a pedestrian, full liability falls upon Waymo, the same way it currently would fall upon the driver in a human-driven car."
Whatever is driving the car should be liable. Always. I can't speak for others, but I will absolutely not use a driverless car if the deal is "you have the same liability you would if you were a driver, but something else is driving it."
"Who will be responsible for this? The answer cannot be "nobody."" Insurance companies. And they are experts at costing and assigning the right prices for liability.
That just moves the question to "who is responsible for carrying the insurance, and what penalties will they suffer if it needs to be used?" Which is basically the same question.
My answer again would be "Waymo should need to carry the insurance, and if their cars kill people or damage things, their premiums should go up." It shouldn't need to be on the part of the human who is using their service to assume that risk.
I guess this is a remedial question, but why are driverless cars desirable? Is it just so taxi and rideshare companies can cut costs by not paying human drivers? If so, why are some people and municipalities so eager to roll out this technology?
So you can reclaim the time and energy spent driving?
If I didn't have to be, you know, driving to work, I would take a nap or read something.
That is not a universal opinion. I very much enjoy driving and, even if I didn't enjoy it, would never allow myself to be driven with this technology. (At least not any time soon when it is still in it's infancy.)
But is your answer that municipalities are eager to roll this out so that their citizens who don't enjoy driving can take a nap or read? Other than cost cutting for taxi companies, I still don't see what the benefit is.
Or literally anything else they would prefer to do rather than driving.
"I get an hour or more of my day back, albeit in a limited form" is absolutely a benefit to an enormous number of people.
"So what's left?" The ability to work in cities other than Phoenix, Austin and San Francisco. Not clear how scalable Waymo is. i.e. it's one thing to be able to work intimately with the transportation departments of three cities, it's another thing to do it with every jurisdiction in the nation. Not clear if it can work in cold weather. (Note all three are dry climates. with no cold temperatures.)
Does Waymo depend on finely detailed mapping? I used to read discussions of this, but I have not seen it lately. Is this because they have progressed to better tech, or because they don't want to talk about it? I am guessing that if the former, they would be eager to talk about it. Because a moment's thought will reveal that so long as detailed mapping is required, this will be restricted to areas that are densely populated and wealthy, i.e. White.
Waymo routinely talks about how their initial approach to new markets is to put out cars with supervisory drivers to completely map the city.
How is "densely populated" "white"?
You elided the word "White." I'm sure this was a completely honest mistake made in good faith.
Pingback: Driverless cars - Angry Bear
As the AI Taxi...ehr, Auto Car buzz wears off and companies pull out, struggle, get banned, the cheerleaders are now telling us that WAYMO is the true North Star!, the only True Light! and suddenly always has been.
Maybe so!
Any day now these MetaVerse futures are going to hit and I'll buy myself a Waymo!
Discussions of accident rates are certainly part of the discussion, but not the biggest part when it comes to the question is can this be scaled up. The discussion I would like to see is how much remote human intervention is required? How often does the car find itself in a situation it can't figure out? So long as the answer is above zero, this means the technology is restricted to locations with good telephone coverage. But even this is not the problem in scaling up. This is driving by call center. We all know how well those work. Right now there is a strong incentive to have a full staff of expensive personnel. But does anything think this would be the model with these cars in general use? Or will a car sit in traffic until its ticket works its way to the front the queue, only to turn out to be a problem the low-level flunky isn't trained for, so it goes into the next higher up queue. Rinse, lather, repeat.
This is one of the many questions the enthusiasts discreetly overlook.
Pingback: Driverless cars – Angry Bear
Working online brings in $285 dollars an hour for me. My best buddy shows me how to do this and makes $29,000 a month doing it, but I never realized it was real, visit the sa05 following link to have.
A look at it------------------------------------>>> https://careershome48.blogspot.com/
Most cars owned by individuals sit in garages or parking lots unused most of the time. Driverless technology means never having to own a car or deal with chauffeurs, reducing the extraction and the energy required to make an automobile for every adult individual. Call Waymo when you need a ride and save the money invested in owning and maintaining a car while helping the economy degrow.
But we already have taxis, car co-ownership companies and public transportation that allow you to go almost anywhere at any time of day in a relatively safe and fast manner.
Its not clear how cutting out the low paid driver really changes the taxi or bus business in a way that causes people to dramatically change their behavior.
If Uber didnt change the world, its not clear that slightly cheaper, slightly safer Uber will change the world. And its also not even clear that we will get to a better version of Uber.
You can take a guy out of the tech industry, but you can't take tech brain out of the guy.
Waymo is pursuing a blind alley
Their self driving relies on incredibly accurate mapping - this is useful in a limited number of places
The Tesla system uses vision and drives the way a human does - its a much more difficult method but when its complete it will work everywhere
Tesla is pursuing a blind alley.
Constructing a 3D picture of the world around you from a 2D camera image is a mathematically unsolvable problem. Applying computer vision to a complex and changing environment whose contents cannot be predicted is, if not an insoluble problem, then at least one that that cannot be currently solved with any conceivable hardware. There's a strong reason why Tesla continues to mow down innocent people.
In fact, detailed mapping is considerably easier than that. You do it the same way you create regular maps and street views -- drive a car down a road and capture data using the same sensors. Waymo is mapping New York and Los Angeles. Tesla still hasn't worked out how to not kill people.
BTW, I don't think you understand how a human processes vision.