Google cars versus public transit: the US’s problem with public goods

 

Hearing my predicament, one friend prescribed a solution: “You need a Google self-driving car!” [...]

 

“I don’t want a Google car,” I tell her. “I want a train.”

 

In much of the world, a train wouldn’t be an unreasonable thing to ask for. New England has a population density comparable to parts of Europe where commuting by train is commonplace.

 

There are too many awesome quotes in this article to excerpt. Go read the whole thing.

This video is largely about computers/robots replacing jobs. Around the 3:30 mark it talks about transportation. The 5:00 mark is the start of driverless cars and the jobs that would be replaced. The whole video is interesting.

 

Engineers Unveil New Driverless Car Capable Of Committing Hit-And-Run http://t.co/gBpb3Pnwzt http://t.co/iHycAtIgKZ

According to one Mountain View resident, Google's self-driving cars "drive like your grandma -- they're never the first off the line at a stop light, they don't accelerate quickly, they don't speed, and they never take any chances with lane changes (cut people off, etc.)."

 

One the one hand, this will make the roads safer... once there are only self-driving cars on the road. But in the mean time, I predict that normal drivers will take advantage of this situation. You can cut off a driverless car and it will let you in. A driverless car on the highway that is leaving a safe gap behind the car in front of it? Plenty of room for you to cut in and make the driverless car slow down.

Interesting tidbit from the article below regarding driverless cars vs. bicycles:

 

Google Will Now Tell You When Its Driverless Cars Crash

 

 

An almost-crash that was a learning experience

 

In a section titled “Interesting Situation of the Month,” Google claims its self-driving car was able to see two cyclists at night better than a human driver: “The cyclist on the left had entered the left turn lane, but veered back into our path to continue straight across the intersection. At the same time, the cyclist on the right entered the intersection, traveling against the flow of traffic. That cyclist then took a sudden left turn, coming directly at us in our lane. Our car was able to predict that cyclist’s path of travel, so we stopped and yielded. This happened at night, when it would have been very difficult for a human driver to see what was unfolding.”

 

I skimmed through the video above but was shocked by a few things the head of Google's driverless car program said. Especially his story about how the a man who is blind and used to have to spend two hours on public transit, but not anymore, because now he has a driverless car! Ya know, instead of just improving our public transit infrastructure.

@grescoe: Planning to retire on a multi-billion $ lawsuit against the company that puts self-driving cars in my 'hood.

(Ouch! You ran over my foot!)

General Motors has thrown its hat into the ring:

 

http://www.manufacturing.net/news/2015/10/gm-plans-55b-in-cuts-to-pursue-self-driving-cars

This is a truly horrifying paragraph.

It's horrifying but it's likely to be rare, if not unheard-of.  Assuming that consumers are still paying for the cost of energy, the father will know that that's a really expensive way of keeping his ride available, even if it saves on parking fees.

First of all, I anticipate when the uniform traffic codes of states are amended for driverless technologies it will probably come down to driverless cars are well and good, but a licensed attentive driver must still be present in the drivers seat, albeit he's not doing much, but he is providing a manual override in the event of system failure.  We kind of almost have this situation now with positive train control, smart locomotives, and 'automatic pilot' manned aircraft.

 

Having said that...let's back off from the engineering and safety bugs for a moment and look at impact on cities.  I really have never heard anybody address some of the larger issues involved.  Somebody needs to look at: 1)what will 'driverless' automation add to the cost of a car; 2) will auto manufacturers pro rate their R+D costs across every product they sell, driverless or not? 3) we can do a great job of testing one driverless car now, but one car is not a system. What traffic engineering changes will cities have to make in things like signalization and what will be the cost? 4) If I want to keep driving an older car without driverless technology, will I be required to retrofit my old car so it can function on a highway full of driverless cars (ie transponder to tell other cars of my location, speed, and turn signal intentions).  This technology is unlikely to make auto ownership any cheaper, and we already have a situation where about a third of households in many of our metro areas can't afford access to a car.  Often the law of unintended consequences takes over and the counter intuitive prevails...so maybe/ just maybe driverless technologies will drive down the proportion auto owners, thereby increasing the demand for public transportation and walkable cities. 

 

But wait...in the old days the horse knew the way home from the saloon even if the rider didn't.  :drunk: So by automating the driver function you are largely reducing the potential for human error...so insurance rates may go down.  So could driving get cheaper?? Again, sometimes the counter-intuitive wins out. So the bottom line is: we just don't have enough information yet to say what the impact on cities will be.

Driverless cars will almost certainly be less lead-footed than most drivers, and from what I'm hearing they will likely be programmed to follow speed limits so they should make real world fuel efficiency go up.  This means they will likely make commute times longer, as well.  At least until all cars are driverless and the speed limit can be raised significantly, if that time ever comes.

https://www.theguardian.com/technology/2016/jun/30/tesla-autopilot-death-self-driving-car-elon-musk

 

The first known death caused by a self-driving car was disclosed by Tesla Motors on Thursday evening, a development that is sure to cause consumers to second-guess the trust they put in the booming autonomous vehicle industry.

 

The 7 May accident occurred in Williston, Florida, after the driver, Joshua Brown, 40, of Ohio put his Model S into Tesla’s autopilot mode, which is able to control the car during highway driving.

 

Against a bright spring sky, the car’s sensors system failed to distinguish a large white 18-wheel truck and trailer crossing the highway, Tesla said. The car attempted to drive full speed under the trailer, “with the bottom of the trailer impacting the windshield of the Model S”

 

That video of autopilot "saving" the Model S appears to have been caused as much by the inattentive Tesla driver as by the guy merging into his blind spot.  I don't understand why the Tesla fanatics are so interested in letting a computer drive their $75,000+ car. 

 

 

Mercedes unveils driverless city bus:

http://mashable.com/2016/07/18/mercedes-benz-future-bus-citypilot/#nq9YBmPbHqqL

 

If driverless buses truly turn out to work (and it seems that their performance could be significantly improved by embedding sensors along their route in a way that really wouldn't be practical for cars) then transit systems could offer much, much more frequent service for the same amount of money they now collect from taxes.  In fact, could municipal bus service turn profitable? 

 

 

 

 

Mercedes unveils driverless city bus:

http://mashable.com/2016/07/18/mercedes-benz-future-bus-citypilot/#nq9YBmPbHqqL

 

If driverless buses truly turn out to work (and it seems that their performance could be significantly improved by embedding sensors along their route in a way that really wouldn't be practical for cars) then transit systems could offer much, much more frequent service for the same amount of money they now collect from taxes.  In fact, could municipal bus service turn profitable? 

 

 

 

 

 

Driverless buses would still face the problem of not creating the kind of economic development that rail does. I suppose there would be slightly more permanence than driven buses due to routes being more profitable and the cost of installing sensors but not that much. I remember watching my old store going from two bus routes stopping in front of it to zero in only 4 years. At one time there were 3 routes stopping in front of it.

Also, it's a shame that someone making 27k a year driving a bus with a capacity of 40+ people makes it tough for transit to pay for itself. Since we decided to stick employers with healthcare costs I suppose we got what we deserve.

A strange look into AI

 

Solve trolley problem scenarios with MIT’s Moral Machine

 

posted by Jason Kottke

 

A group at MIT built an app called the Moral Machine, “a platform for gathering a human perspective on moral decisions made by machine intelligence, such as self-driving cars”. They want to see how humans solve trolley problem scenarios.

 

We show you moral dilemmas, where a driverless car must choose the lesser of two evils, such as killing two passengers or five pedestrians. As an outside observer, you judge which outcome you think is more acceptable. You can then see how your responses compare with those of other people.

 

http://kottke.org/16/10/solve-trolley-problem-scenarios-with-mits-moral-machine

 

 

In Colorado a beer truck made 100+ mile delivery autonomously.

 

http://www.recode.net/2016/10/25/13392326/uber-otto-self-driving-truck-first-commercial-delivery

 

So what happen when the most common job in most states are gone...

 

We're going to end up with a lot of unemployable people.

It's a bit odd that we don't have any video of this truck, and very little of any of the driverless cars actually driving without a driver.  Instead we have a lot of bullshitting from Uber and others making statements that all of this stuff is suddenly going to appear, by the millions, in 2025, when not even one currently exists. 

It's a bit odd that we don't have any video of this truck

 

Did you even try to look for it?

 

I think it is realistic to expect to see convoys of driverless trucks on the ohio turnpike in the next couple of years.  The stipulation is the lead truck has a driver and i believe the back of the convoy also has a driver.  Testing on this was to begin soon.

So what happen when the most common job in most states are gone...

 

 

Ironically many of the people employed in the truck driving profession that hate "government handouts" are going to be asking for them once the free market eliminates their jobs.

It's a bit odd that we don't have any video of this truck

 

Did you even try to look for it?

 

Yes, however I am on my ancient work computer which doesn't always show videos or has ads in place of videos. 

 

People who are all rah-rah about this don't understand what actually happens in the trucking process and that driving a truck on the highway is the most conspicuous but fastest and easiest part of moving stuff around.  There are several types of trucking, several of which are more or less impervious to an autonomous takeover because the driver performs another task in the process aside from driving the truck:

 

1. local delivery box truck from local business -- spends very little time on interstate.  Driver typically drives just 1-5 miles between stops on a 10-20 stop route.  Might leave at 7am, come back at 1pm, and head out on another run.  Comprises a huge percentage of trucks seen on the road, and driverless is basically impossible because the delivery driver also unloads the freight or consumer product at each stop.

2. local delivery from terminal -- similar to above, but moves LTL freight that is consolidated at one terminal then broken up at another

3. terminal-to-terminal run -- trucking company's dedicated daily connections between its regional terminals.  On the surface it appears to be vulnerable to a robot takeover, but many trucking companies have their terminal-to-terminal drivers work in the remote terminal for a few hours, then drive another load back.  For example, a driver might take a load from Cincinnati to Indianapolis every morning, then drive a fork lift for 4-5 hours, then take another load back. 

4. Owner operator -- freelance truckers who find loads through brokerages like the slimy Cincinnati-based TQL.  This is a realm that is ripe for the robot takeover, both the drivers and the brokers.

5. Internal corporate trucking like Wal-Mart -- definitely vulnerable to a robot takeover, since a big company like Wal-Mart could also standardize its dock procedures so as to introduce automation in the receiving and warehousing processes that do not currently exist. 

 

There is already a rise of Uber-type brokerages that are connecting owner-operators directly with load.  This is completely circumventing TQL, which is Cincinnati's largest privately-owned business.  The owner is Cincinnati's newest billionaire, and by all accounts the whole operation is scum.  One of my brothers worked out of one of their offices outside Ohio (I'm not saying where) and the sales team made a game of selling photocopied stub hub tickets to tourists outside their building. 

 

 

 

 

 

 

 

 

 

 

^ Yeah there are definitely some areas of the trucking industry that are ripe for robot takeovers, and others that don't seem to be very feasible any time soon. A lot of warehouses and other large shipping locations that have dozens or even hundreds of truck docks would definitely see improved efficiency because in my experience their biggest concerns are truckers parking their trailers in the wrong spot, not having paperwork ready, etc. - all human error. They could even have driverless yard jockeys and really automate the entire process.

 

However I think even the finest software might have some trouble working with smaller loading docks and urban environments. I once had a knockoff Roomba type thing that rolled around and 'swifftered' my floors. It would routinely get stuck in tight areas and just wiggle back and forth. I imagine that would happen with driverless trucks in any type of non-interstate situation all the time.

That's possible, but I could also imagine that if the trailers were outfitted with an array of cameras and sensors, a computerized semi could accurately back a trailer into a loading dock in situations that a human driver never could.

When you move from one apartment to another, it is easy to assume that you're "halfway done" after you get all of the large furniture moved.  But usually that only takes about 2 hours but you're still there 8 hours later screwing around with little stuff.  The world of logistics is similar.  The big, bulk stuff is easiest to move, but there are innumerable custom shipments of small items that consume most of the manpower.  It takes one mailman 6-8 hours to deliver a few bags of letters to a neighborhood but that same mailman, if put in control of a freight train, could have moved exponentially more stuff.

Google Scales Back Self-Driving Car Ambitions

 

Google is reported to have shelved plans to develop its own self-driving vehicle in favor of nurturing partnerships with existing car makers to continue its work in the automotive space.

 

Both Apple and Google have backed away from their plans to build their own self-driving cars and are now wanting to partner with existing car makers to license their technology. Is this a sign that self-driving car technology is not coming along as these companies would have liked?

I think it tells the most about the complexities and barriers to entry inherent to building automobiles.

Sure, but new companies don't seem to have any trouble getting into the business of building "regular" cars. Tesla was founded in 2003 and had their first car on the road by 2008. Which makes me think, either Apple and Google didn't realize how complicated it would be to build cars on a large scale and recently decided it wasn't worth the effort; or they're having problems with the self-driving technology.

Most of the 100 year-old car companies now have luxury electric cars in development so Tesla might be in ruins and the Musk-Man might be broke by 2025.  I haven't ridden in a Tesla but I'm sure once Mercedes, BMW, etc., get their cars on the market, it'll be obvious that they can make better cars.  Not only luxury cars and sports cars, but ordinary cars in the $35k sales range as well.   

^ GM just beat Tesla to the punch with the ordinary electric car:

 

https://www.theguardian.com/environment/2016/dec/14/gm-delivers-first-chevrolet-bolts-sparking-electric-car-price-race

 

It will be interesting to see how this compares to Tesla's sub $40k car.

Even 70 years after Tucker in the 1940s, the best that a new car manufacturer might be able to do is knock the dust off the legacy manufacturers and get them to be more innovative.

Tesla has had much more cash thrown at them than other modern upstart car companies such as DeLorean, Koenigsegg, Qvale and Vector. Tesla burns cash just like those others did but has been given much more to work with. The belief in the company stretches much further than the iconoclasts that founded the other upstarts and their few heavily-monied investors, but the old car companies haven't had to run on beliefs for an extended period of time for almost 100 years.

 

I do know that the old car companies definitely don't want Tesla around and will build money-losing models to rival Tesla's and can afford to lose money on them much longer than Tesla can before investors turn on them. Chevy is losing 8-10K on each Bolt and can easily do it since they're making so much off of full-size trucks and SUVs that most people don't even need.

^ GM just beat Tesla to the punch with the ordinary electric car:

 

https://www.theguardian.com/environment/2016/dec/14/gm-delivers-first-chevrolet-bolts-sparking-electric-car-price-race

 

It will be interesting to see how this compares to Tesla's sub $40k car.

 

The profit margin is much lower on compact cars than on sedans, luxury cars, sports cars, and SUV's.  The Big 3 are lucky to pocket $1k off of an $18k car like the Ford Fiesta or Chevy Cruise whereas they're getting upwards of $5k per unit on all the stuff selling above $40k. 

 

Allegedly the Model 3 is going to be like a BMW-3 whereas the Chevy Bolt is a subcompact.  But the Bolt is going to beat the Model 3 onto America's roadways by a full year, so Chevy will be far ahead in assessing the strengths and weaknesses of their new car and my guess that they will be able to develop a better car faster than Tesla can improve on the Model 3.   

GM is expecting to produce 30-40k Bolts per year. Tesla has already had 373k Model 3 reservations made and the first one won't be delivered until the second half of 2017. First in this case might not mean a lot if they can't afford to ramp up production and sell them in large quantities.

It might be better to sell fewer of them if each unit loses money. You can't only build cool cars and expect to be a long-lasting car company without a moneyed parent company. That's why Jaguar, Lamborghini and Land Rover have changed hands so many times and why Porsche and Ferrari had to stay family companies for so long then finally being absorbed by the giants rather than being publicly traded. When you're publicly traded, you can be dumped for some biotech company in mere seconds.

^Right, but the big difference is that Tesla is expected to make a decent profit on the Model 3, whereas Chevy is going to be losing money on every car they sell.

This article sums up the steep hill Tesla must climb to profitability...it needs to sell upwards of 10X as many cars as it is currently to lower R&D costs to the percentage the other manufacturers operate at, and that's unlikely since all of them will have similar electric cars on the road by 2020:

http://fortune.com/2016/09/09/tesla-profits-musk

jmecklenborg[/member] Tesla is way ahead of the game since they're going to make a product that they can actually sell for a profit (excluding R&D costs). Most tech companies only have the business plan of: (1) get investors to pump in money (2) get a bunch of users (3) figure out how to make money later.

Tesla is disingenuously advertising the cost of the Model 3 as $35k.  It's really going to be more like $45k.  Also, many will not qualify for the $7500 federal tax credit since that credit is capped at 200,000 models.  Meanwhile, buyers of competitors' electric cars will enjoy that credit for some time after Tesla exhausts their, unless some change is made to the law. 

 

And AV Uber's are coming to San Francisco:

 

Uber Expands Its Self-Driving Car Service to San Francisco

 

By MIKE ISAAC

DEC. 14, 2016

 

SAN FRANCISCO — Uber has always had a special relationship with this city. The ride-hailing company was founded and headquartered here. In its early days, one of the towns where Uber grew fastest was its hometown.

 

On Wednesday, Uber again highlighted its special relationship with San Francisco. The company has started offering its self-driving car service to passengers here, making it the second place in the world where Uber offers autonomous vehicles for public use.

 

That's awesome. The more our culture shifts to driver-free cars, the more lives are saved

That's awesome. The more our culture shifts to driver-free cars, the more lives are saved

 

Maybe.  It's very interesting that Uber is giving riders the option to opt-out of a "driverless" car (one with an "engineer" aboard).  The fact is that a few dozen people have died in traditional Uber cars, but it will just take one or two in a driverless car before nobody dares ride in one.

 

Commercial airliners fly mostly with autopilot but there are two pilots aboard.  If a commercial flight cost $25 less if it were flown entirely by computer, would you take it?  Same dilemma with driverless taxis -- slightly cheaper but no driver.  The ownership keeps the surplus. 

 

 

I can't wait until the thing tries to kill me.

About 1.3 million people die a year in car accidents globally; in the US it's roughly 100 a day.

 

Globally, an additional 20-50 million are injured or disabled due to car crashes every year.

 

 

If automated technology can dramatically improve those numbers (didn't the google car only get in one accident, and that was the fault of another driver?), then I'm all for it.

The Google car has a dude in it all of the time.  That dude is correcting the damn thing when it wants to do something weird.  One of the first reporters to ride the "driverless" Uber car in Pittsburg noted that the "engineer" had to intervene pretty regularly to keep the thing on "track". 

About 1.3 million people die a year in car accidents globally; in the US it's roughly 100 a day.

 

Globally, an additional 20-50 million are injured or disabled due to car crashes every year.

 

 

If automated technology can dramatically improve those numbers (didn't the google car only get in one accident, and that was the fault of another driver?), then I'm all for it.

 

I don't care about your numbers when I am flying through the air as the car tries to drive over a bridge that was removed 60 years ago as a part of a road realignment that its GPS magically resurrected. Or it Blue Screen of Deaths me into the back of a semi.