It was probably in P&R, and probably posted by me. It’s something I’ve hit on a few times.
The buggy whip manufacturer has always been a pretty silly analogy. The more relevant industry was the buggy and carriage manufacturers which was a real industry. In 1890 there were 13,000 companies in the industry employing hundreds of thousand people. In 1900, almost 1 million buggies were made. Only a handful of companies who’s expertise was wood, not metal manufacturing made the transition to automobile part supplier and one buggy whip manufacturer is still around, selling to the dressage fans.
Technology has been destroying jobs since the invention of the wheel. There was far less of a safety net back in 1900 than there is today. Yet by any measure unemployment, or work force participation there are more jobs today than back in 1900. We will have safer roads, with cheaper trucks and more energy efficient trucks (no need to have a driving compartment or sleeping facility). It is true that millions of truck drivers will lose their jobs but they’ll be replaced by something else. I am pretty sure that many of us were born before computer game developer was a job, or web developer, and certainly SEO optimizer. But even outside of the STEM world, personal trainer, personal shopper, and lifestyle coach didn’t really exist before I was born. The fact the people can’t identify what these jobs will be, puts us in exactly the same place as the folks in the 1900s wondering what all the farmers would do.
This is one of those things that sound great like “For every door that closes, a window opens!” that may not be as comforting as the intent. I agree that there will be other jobs and technology has been killing jobs since forever. I don’t agree that those other new as-yet-unknown jobs will be necessarily attainable for the people that are made unemployed. (At least, not without a lot of pain and effort.) For example, I’ve known quite a few personal trainers, lifestyle coaches, and shoppers, but I don’t think any of them transitioned from careers in coal mining. Similarly, when Boeing folks around here get laid off, most of them do not become medical transcriptionists or insurance underwriters.
Holding back technology to try to keep people in their current line of work doesn’t work either, though. The real point is that people need to be steered away from career paths that are threatened. Automated cars have been coming down the pipe for 35 years or more (some experiments with driverless cars date back to the '20s), and a fully-automated trucking industry is still at least 10 years away, right? Plus there will be some lead time after the tech is in place while it gets rolled out. So it’s not like these careers are disappearing overnight - someone who started as a trucker in the 90s will be in their 60s by the time all the jobs are gone, even if they started young. As we got into the 00s, the number of people going into that field should have declined and now it should be very small. Note I say “should be” - the reality may be different because of people downplaying advances in automated cars until it was too late. There are also reasons this wouldn’t play out well simply because some people will jump at a lucrative field and hold onto it until the last gasp, but it’s actually that last group that is the “moral hazard” part of this equation. If we give laid off workers a partial stake in the robots that replaced them (somehow- this isn’t exactly clear-cut when different companies compete), we encourage people to hop into dying industries to get the last bit of salary and then also reap the rewards of the resulting unemployment, which seems a bit unhealthy.
That said, UBI has always been the answer to the worst problems caused by capitalism, and the fact that people resist it so hard is a relic of social institutions designed to exploit the masses and also (probably) of racism. It doesn’t seem at all odd to people when a state sends everyone a check for their share of the oil sold by the state, I don’t know why it should be any different to send everyone a share of the collective productivity of every person and machine in the state. People who work harder still get more money and live more comfortably, but people who don’t work or can’t work get by and still contribute to the economy by helping to select useful products and patronize businesses. People displaced by new technology could use the UBI to train in a new or related field, or to turn a hobby into a business or simply to find an industry that best suits them. If you used it to get rid of minimum wage laws, you might also have an explosion of jobs that are only profitable at wages too low for anyone to live on, which would get rid of the “but a job is more important than just money” argument about welfare and other safety net spending. Best of all, people would be excited by every new technology, because it would increase productivity and therefore increase the per-capita basic income. It’s a far better system, but it’s very difficult to sell people on it, because they would rather have the status quo (even when they know their personal status quo is tenuous).
Is it really true that a long-haul trucker can’t find an industry where his or her skills would be of use? Being able to stay alert and in control of a dangerous machine for long periods of time, being willing to spend much of your time away from home, and perhaps other aspects of the job that I’m not familiar enough to name could easily transfer to a new career. With some online courses and a short effort to sell your skills you could find jobs in quality assurance, drone piloting / traffic control, sales, automated trucker algorithm refinement, and who knows what else? But, do any of the folks who currently do these jobs really want those other jobs? Not many of them, I’d bet - and it’s hard to blame them for that. I love what I do, and while I know there’s lots of other things I’d be capable of and paid reasonably for, I’d rather do this and I’d certainly rather not have to look for and learn a new career.
In the absence of a UBI, I think the best hope for controlling the effects of automation is to train each new generation to love learning, while reducing the cost of training. A generation that expects to change jobs every few years will not have this problem of their job disappearing. They will always be moving on to new companies doing things that are currently viable.
Just to be clear, I’m not advocating a halt. I just think we may be barreling into this without really thinking about the consequences. I know that’s sky-is-falling territory, but I’m old. I’ve earned a little caution.
As for the rest of your post, sure, I’d love UBI here and everywhere. I think we’ll get flying cars and household robo-maids for everyone before that happens though.
Sure, but they had factory work, which actually takes less skills than farming.
So we need to romanticize a modern low-skill job. The noble Turker, toiling away at an honest living supplying the robot overlords their training data!
Actually, they had a lot more than factory jobs. Manufacturing never represented the majority of the jobs in the country. I think people just forget all the other jobs out there.
Take a look at the impact the automobile had on the country. It devasted the carriage and buggy industry, severely hurt the ranch business, drove a further nail in the demand for farm hands. It also could have wiped out sanitation worker jobs, but since sanitation workers were doing such a crappy (sorry) job keep the streets cleared, in all major cities, of horse shit and horse carcasses, it simply made cities less smell, more sanitary, and far more livable.
But look it all the nonfactory jobs it created
Cabbies, Truck Drivers, Chauffers
Gas station owners, auto mechanics, and service station attendants
Auto dealer owners, car salesman, tractor salesman, car wholesalers
Auto insurance salesman, insurance adjusters, and auto loan bankers
It causes a boom in tourism, creating a demand for everything from more coffee shops, and motels, to road map creators.
Practically nobody in 1900 would have for seen these new jobs and not many people even in 1920. But the negative impact was seen far earlier than positive impact.
We seen an explosion in delivery service in the last decade, but most failed for a simple reason. It’s bloody expensive to pay somebody to drive to your house to give you something. I had a craving for a Pizza last week and getting it delivered would have more than doubled the cost after tip. Since it was faster to drive 10 minutes to get the pizza, I skipped delivery. If it was cheaper and/or faster I would have had it delivered. Yet pizza delivery has been a staple for decades.
But what are the impacts of eliminating the major cost of door to door delivery? There is a big demand for fresh non-process food, which includes everything from produce, to fish, and even raw milk, to healthier meals. If the delivery cost plummets, that opens up more jobs in cooking, and organic farming.
But that’s just the start. As more or delivery gets automated, there will be a need to develop an infrastructure to support these systems. I see a market for high-tech mailboxes, that are secure but also provide climate control to keep hot food hot, and fish and produce cold. They’ll need to interface with a wide variety of delivery vehicles from automated pizza trucks, to farmer produces trucks and Amazon drones. There are high skill jobs needed to design these and medium and low skill jobs to sell, install and service these.
The previous generations managed to adjust to these changes. Most of the evidence points to the pace innovation slowing down this century, so I find hard to understand why would have a more difficult time adjusting than in the past.
I wouldn’t use the haphazardly hurried system used by Uber here as a metric stick to measure how the car automation is progressing and the riks that will present in the future. Why not use Google, who started years ago, and it has an incredibly rock solid performance?
I’m a bit of a skeptic in terms of how fast this will roll out. On a recent tech podcast the well-respected journalists were debating it and one of them said 50% of miles would be fully driverless in 5 years. Sorry but this is an area where tech geeks can get it wrong with all the enthusiasm. The average vehicle on the road is just under 10 years old so even if AVs were 100% of sales starting today then it would take a decade to reach that threshold. There are zero AVs on sale today at any budget. In fact all the hype about the Uber cars in Pittsburgh and it turns out “one of the cars was spotted in public” and “it’s unclear if they are picking up passengers yet”. Since Uber is getting nearly unlimited media exposure for it, and their cars still have a driver, it makes me think AVs are still coming more from marketing budgets than from staff payroll budgets.
Wired magazine has taken a much more measured (and well researched) angle on AVs through several articles. Basically AVs are certainly coming and they will disrupt the driving economy. But the economic disruption doesn’t happen when there’s always a backup driver. It requires full level 5. That will take a decade before it is a 10-thousand dollar option package on a car, with an expensive monthly subscription (level 5 AVs need accurate 3d survey data to compare against their senaor data) and only available in certain urban areas and on major highways. Even then there will be all sorts of issues with parking lots, construction zones, and poor weather (hidden pavement markings under snow). For that 10 grand you’ll get a car that is so hyper conservative it will be noticeably slower, and subject to bullying by other drivers. So would you sign up? I for one would probably skip this 10 thousand dollar option a decade from now given these limitations. Taxis, sure, and they’ll feed the hype so we all keep talking about it. Delivery trucks? As stated above the driver is still needed as he is the delivery person. So instead of destroying the world eceomy we’re now more talking about letting UPS drivers play games on their phones between destinations. Legally, asthey already do that :).
Trucks are certainly the interesting case. But most goods movement above a certain distance (600 miles) is multi-modal. Short distance will often require a driver. Distribution center to distribution center, that is where AV could take over. But truck movements in a yard and reversing manoeuvres to attach trailers are very complex. Truck simulator is hard and it doesn’t do it much justice. Can an AV computer fill in the gaps as to what’s invisible on the other side of a trailer blocking the view as well as an experienced driver? How about a double trailer? Not sure. Personally with all the negativity about truck driving careers i think there’s a chance there could be a shortage.
I’ll get off my soap-box now, good discussion…
I have a little more faith in technology to solve these issues. Reversing maneuvers and trailer linking are hard for humans to do, but it doesn’t follow that it would be equally difficult for computers, especially if they were given additional (now very cheap) sensors to assist them. The actual physics involved in backing up and positioning a trailer are really simple - every Freshman engineer has to solve them in their “dynamics” class. What makes it tough for a human to do is the counter-intuitive nature of the movement coupled with (usually) viewing the action through a mirror, and the added complication of the jerky acceleration of the vehicle.
I agree with your depiction of the technology in 10 years. But if the automation requires special maps or infrastructure, and if it is limited to certain areas, than that is SAE level 4, not level 5, and will have very limited economic impact. If a driverless car can’t maneuver to the curb at the airport, it’s not really economically viable for Uber. They’ll be used in controlled, predictable situations, but not as general human driver replacements.
People talk about infrastructure improvements to support automated vehicles, but there’s no consensus on what those should be, much less actual planning for them. The industry hyping itself as if fully autonomous vehicles are around the corner is not building any sort of political impulse for those improvements.
A real level 5 vehicle would have a net present value of about $2 million, each. They just aren’t going to happen in the near future. I am confident that in 2030 there will still be millions of people employed in the US as drivers.
I just don’t see the public buying into the concept. People are willing to deal with untrustworthy other humans a lot more than a program.
That’s why those automated teller machines never caught on.
Oh people don’t care about that sort of thing. Automate McDonalds? People are probably fine with that.
Automate a large object moving at high speeds? Not so much.
They don’t need to catch on with the car-buying public, they just need to catch on with the car-renting public. Once Uber and taxis are all AI-driven and cheaper than owning and relevent costs, no one in the city will need a car anymore, further fuelling demand. Once all business travellers and vacationers start renting cheaper AI rentals (you won’t need license + insurance anymore!), it’ll fuel demand. Soon, large swathes of the population who don’t need to regularly commute a long distance by freeway permanently switch to the the AI-car-on-demand model, all that will be left are the suburbanites on commutes, and country folk… which admittedly is still substantial… but 50% of the market is not a small thing, especially when they will control 90% of the urban one.*
*pulling numbers out of my butt
As an aside, here’s an article from Bloomberg talking about how Tesla has 1.3 billion miles of data since they install their sensors in every car - the sensors collect data even when they’re not being used for Autopilot. They also have over 200 million miles of Autopilot data.
https://www.bloomberg.com/news/articles/2016-12-20/the-tesla-advantage-1-3-billion-miles-of-data
This stuff is coming, and I still can’t wait.
Evidence suggests otherwise. These were everywhere a few years ago, I can think of a few now and they’re rarely in use. They were yanked out pretty quickly, and I can’t imagine they were cheap to put up either. I prefer to do my banking remote, but if I am physically in a bank, there better be a human to talk to. I’m there for a reason.
I am not opposed to auto-driving cars. I just think we could start smaller, like self-parking. And in the long-run there are things the computer can do better than we can with the right sensors. We were almost in an accident yesterday. one car taking a left, another a right, and with ongoing traffic for a bout two seconds no one was looking in the direction. It was avoided but I assume a computer would just sense when the cars came close to all stop or adjust. The computer can have more than two eyes looking in more than one direction. My concerns about other limitations.
I phrased that poorly, I meant as a concept.
If there is a glitch and you get the wrong sandwich, it’s not a big deal. If there is a glitch and you’re hit by 40 ton truck doing 70 mph, not so much.
To be fair, this is pretty much why I’m against the ideas of teleporters :)