In the wake of an Uber driver arrested after allegedly hitting and injuring a cyclist in Fisherman's Wharf on Sunday, more questions are being raised about driver safety requirements.
Specifically, taxis are regulated by SFMTA and are required to take 28 hours of classes through one of four approved private driving schools, and another day of training through the SFMTA. An hour of the SFMTA training includes instruction on sharing the road with bicyclists and pedestrians, is taught by a member of the San Francisco Bicycle Coalition. At least two hours of testing is required. And as part of the city's Vision Zero plan, which aims to eliminate traffic deaths by 2024, SFMTA is instituting a large vehicle urban driving safety program for all large vehicle drivers who work for the city.
All told, as Central City Extra covered in their latest issue, there are 71 pages of dense regulations for cab drivers to follow. Not so for drivers working for Uber, Lyft, Sidecar or limo companies, which are not regulated by the SFMTA -- they only have 28 simple regulations to follow. [...]
Frisbee confirmed that the San Francisco Bicycle Coalition is in talks with Lyft to include online safety training. What about Uber? "At this point, Uber has decided that our safety training is not worth their investment," Frisbee said. She said Uber told the San Francisco Bicycle Coalition the company "found someone else to do it," but when it asked for details, including who was giving the training, how long it lasted and what was included, "they said 'we're not sharing it'," she said.
Previously, previously, previously, previously, previously, previously.
I have a pile of those those stickers at home. I need to remember to keep them with me while on my bike. Also this is good to know. Consider me boycotting Uber until they adopt the San Francisco Bicycle Coalition training.
Matt O'Donnell where did you get said stickers???????
We don't even yet know what premeditated the biker getting pissed at the Uber driver, but it resulted in the driver running down the bike intentionally. I'm not sure if safety training or some sort of mental competency evaluation, or both, would have been the better investment here.
did they really charge you $60 for shipping?
Uh no? I live in the states. Shipping is $3
I have no idea about the states, but where I live the taxi drivers are one of the most aggressive group against cyclists. All the training doesn't seem to make a difference. (They also tend to speed and skip red lights, but that might be a Spanish thing)
Taxis here are horrible and it was hard to imagine them getting any worse -- and then Uber showed up and we didn't have to imagine it. Welcome to the next level.
Of course, Uber's not a taxi service (yeah right!). Special training would conflict with their position that it's just connecting regular drivers on their way to work with others going the same direction, so I'm not surprised. Still, they're not driving trunks or buses, just regular cars like everyone else. As a cyclist, given that most car drivers don't work for Uber, I'd hope that not hitting people would be a part of everyone's basic driver training and testing. Apparently there are parts of the world where driving tests are much more thorough.
It's like someone's punking the libertarians. Really? ÜBER? REALLY? I'd be happy I moved to Canada, but I used their adjectival service this evening. Dammit.
Cool, SF is getting on board! Anyone have a link to their latest plan?
I like the Vision Zero plan and all, but every plan I've read only addresses roadway design. Some of the best reductions to the severity of collisions in recent years has come from soft bumpers and protected movement ("left on arrow ONLY", pedestrians get their walk signals while all vehicular traffic is stopped, etc.), which is barely discussed in Vision Zero.
We need more improvements to vehicles themselves. After 100 years, cars still have blind spots! How hard is it to slap a couple more side mirrors on cars, or a side radar that lights an indicator lamp on the dashboard? People would get really fucking good at subconsciously knowing when that thing was lit on either side. When GM says they want a pretty good sensor for about $10, companies whip out knives and cut each other for a chance at that contract.
Why do brake lights only have on/off states? The area lit should increase with increasing deceleration. Or change color or something. It's not like these are new ideas, but we all know how good GM is at making decisions and valuing human life.
We've been killing 30-50 thousand people per year on US roads for 70 years! (And in that time, steadily reduced the number of people killed per million miles by a factor of about 15.) We still have a long way to go to reach Vision Zero.
Reducing the city-street speed limit to 20MPH - and enforcing it - would save more than half those lives.
Probably. But what's the data on that? I've never seen a comprehensive highway/city and interstate/rural deaths breakdown. Everyone holds up a few summary numbers, then spends 90% of their column talking about US Highway 6.
I can't decide if 20MPH would also increase or reduce congestion, and increase or reduce travel time; or the conditions of either.
(You might think reducing the speed limit would always increase travel time, but no: it reduces the variance of driver speeds (even considering speeders), which makes the flow of traffic more laminar and more predictable. On the highway it delays the onset and severity of rush hour congestion, and in town it makes traffic signal timing easier to optimize.)
You need to know what the average speed is. 20 mph, in many dense cities, is above the average motor vehicle speed in peak times. The rest of the time the city is so dense that you don't want to encourage more people onto its roads, so build good public transport systems. Essentially my view boils down to lawmakers curating safe spaces for the people who are their responsibility.
The Uber and Lyft drivers need driver's insurance, right? This is a place for internet-herp-derp, so I'll ask if a solution would be that their insurers offer a discounted rate for skilled-up drivers who have taken these training courses.
It's not entirely clear that Uber drivers have the proper insurance. While the gap while driving a passenger has been (supposedly) fixed, at least some insurance companies consider the period with the app running but no passengers to also be commercial time, and that is currently uninsured.
All we need is data showing that taxi training reduces
accidents claimspayouts. I'm sure Yellow Cab's insurers' actuaries have that data and calculated those risk tables long ago.
The set of Uber drivers is a totally different demographic to taxi drivers, though--different age, race, income, marital status, gender, total hours spent driving, etc.
I wasn't trying to bash on Vision Zero for an aggressive timetable, or for focusing on different things than I do.
Rather, if lawmakers want to guide government use of taxpayer money to make roads safer, they can guide the mostly-government-owned GM to make vehicles safer.
The gubmint doesn't own General Motors anymore. And, even if they did, GM has a bit less than 18% market share in the US. Any gubmint mandate to GM would therefore affect only 18% of the new cars sold in the US. And, if the price of a GM car were increased compared to all other cars, GM market share would reduce further.
Whereas, legislation would affect 100% of new cars sold--leading to a much faster adoption rate.
Blind spots have got worse in some ways. Older cars have pillars just strong enough to hold the roof on, which are consequently quite thin.
Newer cars actually have some roll over protection, but this means thicker pillars which are big enough to block the driver's view of a cyclist/pedestrian/motorbike even when they're quite close.
I'm mainly talking about the A pillar (ie the ones at the front on either side of the windscreen) and I'm not sure how a mirror could be positioned to help here.
Some companies are looking at radar/sonar systems to detect objects around the car, but even if they worked perfectly and were introduced tomorrow it'll still take years for the technology to trickle down to every, or even most cars.
Yeah the A pillar problem isn't going away.
I've heard this for decades. They can't be seriously looking.
Making the sensor isn't hard. The issue, aside from convincing auto manufacturers to care, is demonstrating that drivers will observe, comprehend, and act on the added information. GM is the kind of company that can complete that task.
I wonder if it's not a liability issue. Adding a new safety feature to a car means exposing themselves to claims when people don't use it and wreck.
Well, Volvo will sell you such a car today, and Subaru sell them in Japan.
The issue isn't making the sensor, it's integrating that and weeding out false positives or negatives. Say you are designing a sensor to prevent a car from turning into the path of a bike coming up from behind. Sticking a couple of sensors on the car is the easy part, then you have to work out what pattern of inputs unambiguously indicate a cyclist (is your sensor looking for a metal frame? oops, you just killed the guy riding a carbon fibre bike) and so on and so on.
I suspect it might be easier to implement in a self driving car, because at least then you know the intent of the driver. But it's not a trivial technical problem, and at the end of the day, it'll add cost to cars that people buying cars might not want to pay for ("if I hit a cyclist, I'll still be ok in my metal box").
I hate the "added cost" argument. If people want to pay less for a car, they should ask a friend to negotiate for them at the dealership.
We currently kill about 1 person per million miles driven in the US. Let's suppose that 1% of those deaths can be prevented by our proposed sensor. That's 1 person per 100 million miles driven.
Let's suppose cars on average last 100,000 miles. That's a low estimate, but ok. So we kill 1 person per 1,000 cars.
When valuing human life for auto safety purposes, I think the amount is $5 million per life. Maybe it's $7 million by now, I don't know. So one life per one thousand cars means manufacturers should be willing to spend up to about $5,000 per car on parts and research for our sensor.
At 10-20 million cars sold per year, that's an obscene amount of money.
Fuck the added cost. "Safety features will cost more!" means "I don't care about other people's lives."
It's not like we're talking about the added cost of a new gizmo on a toy. This is the cost of improving a system that kills more people per year than ten 9/11s. How many billions did we waste on that again?
Your math is technically OK, but your assumptions are wrong.
First off, we kill about 1 person per 100 million miles traveled--not one per million.
1 per million doesn't even pass the sniff test--if everyone you know drives 10000 miles per year (maybe high for SF, but low where I live in metro Detroit), then 1% (10,000 miles * 1 death/1,000,000 miles) of your friends would die every year in car accidents.
So using your 1% effectiveness assumption, you've got a budget for it of $50/car--not $5000. Or, put another way, the budget for all automotive safety is $5k/car. That $5k is basically already spent in today's cars--because automakers already made the same analysis when the cars were designed.
By the way, it's not the manufacturers that spend the money. Well, OK, they do spend it, but they want to get it back. Meaning that it doesn't come out of their profit margin...it goes into the sale price of the car. And, human nature being what it is, the average consumer will absolutely say "fuck that" and buy a different car--one without your sensor.
That's not to say there's no way to improve the safety of cars for all road users (drivers, passengers, pedestrians, cyclists, motorcyclists, ....). But, saying "increase the cost of cars by $5k and find a friend who is a killer negotiator" isn't it.
Oh, yeah, it is a hundred million. I still hate the cost argument.
Manufacturers had to be mandated by the government to put seatbelts in new cars. I don't think anyone begrudges them now. The government should do more to fund research and development of safety features in vehicles and should continue to mandate the installation of ones that are proven effective.
Datapoint to lack of training of sharing economy drivers, at least here in San Diego.
My wife was in a Lyft the other day and the driver pulled out of our busy street driveway on a left turn, clipped a car that swerved to miss, and kept driving down the road. My wife pointed out that According-to-Hoyle she needed to pull over. So she did, 100 yards down the road. The other car came up behind and the Lyft driver related that she'd drive to an ATM and get 'em $50 for their trouble. The other-car occupants rightly insisted, actually, no, we're exchanging insurance information right now or cops will be called for hit-and-run. That's all the story I have as my wife flagged another Lyft to skedaddle out of there.
I've had such universally awesome experiences with Lyft this bummed me out. Upside, when my wife reported this to Lyft the customer service was quite responsive.
All in all the collaborative sharing economic utopia is a sidegrade at best.
It's almost as if a deregulated market is worse... but that can't be possible, because all humans are fully-informed, fully-logical, fully-spherical atoms of pure, frictionless intellect! I read it in that S&M erotic fiction Ayn Rand wrote!