Some of them will definitely be out of work, but not all. Given the difficulties in navigating urban streets with large tractor-trailer rigs, it is likely that the autonomous rigs will be directed to centralized locations just outside cities, at which point human drivers will navigate the last leg of the trip. So, truck driver jobs where one gets to work locally and go home each night.
Even assuming all the rules for all the intersections and roadways are encoded into a database somewhere, and updated regularly to reflect road closures and construction etc, and all cars in the driverless car zone are driverless (a lot of big ifs), there are still complex decisions involving unknowns. Someone steps out in front of the car. Should the car keep going and hit the clueless pedestrian, which would definitely severely injure them, or swerve off to the side and cause unpredictable damage to people and property? For another example, what if there’s equipment failure that will cause a collision, should the driverless car minimize damage to itself and its passengers, or to innocents? There is a lot of soft decision making that computers don’t do well without human assistance.
2 Likes
Yes. The fact that Google can even propose getting rid of the steering wheel shows a lack of imagination. A Google car will only go to places that Google recognizes as places.
A couple of years ago we attended a party thrown by my wife’s cousin at his place in the country. He had arranged for the guests to park in a neighbour’s field, which involved driving off the road across a shallow ditch and finding a spot in a stubble field. A Google car would simply not allow that.
Yesterday I picked up some lumber at the local yard. Even if the car recognized that I could drive onto their property, I have no idea how I would tell it to drive down one lane between the piles of 2x4s, do a 180deg turn, and park in front of the warehouse door. If my experience with Google maps is any indication, I would end up turned 90deg to the laneway, jammed between two piles of pressure-treated posts, with my best option being to walk home and start over.
2 Likes
It would definitely hinder acceptance & be inconvenient to many people…but how many people really cross state lines on a regular basis? I don’t think it would be enough to make self driving cars useless.
I’ve owned six cars in my life and 3 of them have never crossed state lines (or more accurately, not while I’ve owned them). Last time I regularly crossed state lines I lived within sight of the VA/MD border and I still crossed less then once a week.
When I lived close to the MD/PA border I use to go up to PA once a week to grocery shop. I don’t think I actually saved money but I was young, enjoyed the drive, and the scenery was great. If MD had self driving cars and PA made it illegal, I probably wouldn’t have gone, or maybe gone and driven half the trip myself. I would still have bought a self driving model.
When I lived on the MD/VA border (technically the people across the street were on the border not I) I visited friends and family in MD frequently. If my self driving car could only drive itself in VA I would still have visited them, but I would have bought the same car because most of my miles were commuting in VA.
The pain would be more in the designing the self driving cars to physically match the demands of all (or the majority of?) states, and in either doing the same for the software or testing the significant number of combinations of differing regulations. The “pain” is more likely to result in deaths in one state because of regulations in another state…
You might be “forced” to buy a more dangerous car in MD because CA demands that you have driver override and a steering wheel (despite many studies showing how dangerous the first ten seconds after a human takes control back). VA, MD, and PA and every other state might have different turn signal requirements and some table of all of those might be just a little off, and the unit tests might miss just the one case that off (or worse yet, test “to the table”) so it is safe in 49 states, but in SD the car signals 0.2 seconds before lane changes.
Even if all that works out muddling through 50 sets of regulations will delay getting self driving cars on the road. It is easy to imagine it taking an extra year or two because of that, at 30000+ deaths per year (in the USA) assuming self driving cars could cut that to a tenth that is 58,000 lives we are talking about (note: I’m not assuming we save 27,000 lives the first year we start selling self driving cars but that is the eventual steady state we ramp up to, start that ramp two years later and we won’t hit 27,000 until two years later – my bet it it takes about a decade maybe two between wide spread availability and most cars on the road having it…but who knows I could be pessimistic!)
So yeah, I would really prefer one set of regulations (assuming they are sane) to 50 sets, but one way or another self driving cars will still be viable.
Addendum: the state laws for trailers varies. For example in many states you can tow a trailer with another trailer. In CA you can’t unless you have a class A license. It doesn’t stop me from owning a trailer. It doesn’t even stop out of state tourists from towing their RV with a boat (illegally) behind it. CA cops don’t seem to enforce it for out of state drivers.
But the real attraction of a self-driving car is the ability to automate long-haul drives. I drive across two state lines 5-6 times a year to visit my folks in Cleveland, and it would be wonderful to be able to make the drive in a self-driving car.
2 Likes
Interesting, most people seem to think making the daily commute easier is the big deal. For me I spend about two hours commuting a day. So that is a huge amount of time to get back (however I’m working at a bay area tech company so 45 min of that is already on a nice bus with wifi, so already I have that time back…really I would only get another half an hour back…or I could stop working to the bus schedule and work to my own schedule again)
Some people I’ve talked to thought it would be great to get a RV, wake up, have a shower and breakfast on the road. (personally I think RVs get crappy gas milage, and I don’t want to take a shower when the driver might be changing lanes or doing hard breaking to avoid running into asshats ahead of us…but who knows if most traffic is self driving maybe the ride will be gentle enough for that)
I don’t deny that it would be nice to have for long trips, I’m just surprised that “long trips” is anyone’s driving force here as opposed to what they do with the car most days (my commutes in MD and VA also sucked, no company bus, and a self driving car would have been awesome)
1 Like
Point, very much point. I’m a freelancer, so I don’t have a commute so much as I have the occasional mad dash across town.
1 Like
I’ll second this, and add that it seems likely to me that autonomous trucking may be a primarily nocturnal activity initially on the basis of clearer roads :: more efficient use of fuel.
The reality is that there will never be driverless cars, only self-driving cars. But there will be areas like major highways that will likely mandate that self-driving be enabled. And because the most ideal, speediest system will likely rely on multiple automation, I’m okay with that.
Why? What prevents the same car from operating in both environments?
Automation will, over time, put many millions of people out of work by eliminating the need for them, just like it did for farm workers a century ago. Truck drivers are one near-term casualty there. Some of them will find new jobs in other fields. Some will become permanently unemployable, unable to quickly or perhaps repeatedly acquire new not-yet-automated skills. A century ago there was plenty of demand for unskilled labor in factories to make up the losses. Then we massively subsidized education, built a social safety net, and unionized to keep up for a few generations.
Eventually, I think the only way to deal with automation is to socialize enough of the gains to eliminate the concept that everyone should need (or could get) a job to support themselves - something like a universal basic income that rises over time. But I think it’s going to take a massive disaster to generate enough political will to make that happen.
3 Likes
Don’t let David Weber hear that, or he’ll make a strawman out of the idea again. 
2 Likes
Weber took a specific class of UBI proposals (provide enough to live, and funded by phases out existing programs) and showed they don’t work today. He is right. This is irrelevant to the claim that reasonable assumptions of future development over the course of decades can radically change the situation, or that other classes of UBI proposals might be economically, if not politically, feasible soon.
I would say you’re showing a lack of imagination. Open google maps/EARTH, drop a pin instead of typing an address at rear loading dock or where ever you want to go “off road” and let the car figure out the mechanics of getting there avoiding obstacles along the way. The car is not simply following a path on a road. It has sensors. It’s “thinking” how to get to the destination; combining data from many sources. It’s likely paved, possibly dirt. What’s the big deal? You forget self-driving cars got their start off-road! It’s how the tech began as a DARPA experiment and contest. It’s actually a huge improvement over typing address. Sure, tell the car to go to Ted’s Lumber Warehouse, but when close zoom in on the map, drop a pin and let the car finish the job.
2 Likes
That wasn’t the impression I got when I read his Honorverse stuff (my impression was that it was more of a general, all-encompassing strawman that “any form of government-supported welfare is bad and people are innately lazy”). However, it was a few years back when I read it, so I will most certainly accept that I might have been missing context that he was writing against.
Without any sarcasm, could you point me towards something reasonably detailed on the “specific class of UBI proposals” so I can soak it in? (Heck, I still have another 20 books to go in my Goodreads challenge for the year; this could work out; it’d be nice to revisit the stuff before he got Protection From Editors and started deliberately bloating out his plotlines for the money)
1 Like
I’m very excited about self-driving cars and I am happy to hear that this regulatory hurdle is leaped.
Things that I like about self-driving cars:
-
Old people able to get out and socialize more because they have access to reliable, safe transportation
-
Old people not on the roads driving dirt slow because they now have reliable, safe transportation
-
People with disabilities able to live more independently
-
Being able to relax during a daily commute
-
Long drive to new place? Don’t have the stress of navigating a new route; drive overnight and sleep, wake up there.
-
integration with Solar Frickin’ Roadways.
2 Likes
You’re making the assumption that “truck drivers” is just the big-rig folks, but that’s not the case. From the article:
The prominence of truck drivers is partly due to the way the government categorizes jobs. It lumps together all truck drivers and delivery people, creating a very large category. Other jobs are split more finely; for example, primary school teachers and secondary school teachers are in separate categories.
So again, once the drones take the local deliveries, and the auto-driving trucks take the long-haul stuff, my original question stands. What are those people going to do for a living?
Also consider “employment programs” and note that a program creating a million jobs is a BIG deal–what programs will we use when those auto-drivers start?
1 Like
Well, since we’re fairly firmly in the cyberpunk timeline at this point, probably security guards for the automated trucks so that anarchist hackers don’t reroute them to steal the groceries for their communes. 
1 Like
From the first part of your paragraph you sound as if the technology is not currently in use. It is. And has been for quite some time. There are driverless cars that manage to cope with complicated situations. Driverless cars are system and incorporate many different data sets that you seem to be worried about being updated, but ultimately the car directly is observing its surroundings and making decisions; not checking a db to see if this street is 45 or 25 mph. It knows what a stop sign is and doesn’t need a database to tell it there is a 4-way or a 2-way intersection. Since we’re using scenarios here: Imagine a driverless car approaches a 4-way stop intersection. An unscrupulous hooligan vandal has removed one of the stop signs for a frat party creating an obviously dangerous situation. Two cars approach at perpendicular angles, one will stop the other will not. Oh, and the guy with the stop sign is distracted and won’t notice the car without the stop sign and will pull out in front of him. If you’re the other guy without the stop sign approaching the intersection from an seemingly uncontrolled intersection at 35mph would you rather have the distracted driver at the wheel and broadside you or a driverless car that easily saw the other car and calculated it’s speed and distance and avoided the ensuing crash? As for the ethics, yeah. that’s way beyond me, but it is a solvable problem. People make ethical decisions every day–mostly incorrectly I might add.
