I'm rather surprised to see Software Developer as the most common job in any state. I wonder what they consider to be software development. If you are a not an actual programmer but you work at a company who's only product is software, and you contribute to that product in ways other than writing code -- perhaps as a copywriter for e-learning software, or a graphic designer for a web app -- are they counting you as a software developer? Because while the software development industry is certainly very large now, the number of people actually writing code is not large enough in my experience to be the most common job anywhere.
The states which list "software developer" as the most common job are also states which are home to quite a large number of gov/defense contracting companies.
This is perplexing to me as well. I would have thought something that appears across industries would be top. Something like accountant, secretary, sales person. But then again Software Dev crosses a lot of industries. Also professional jobs (desk jobs) have a lot of similar roles with different sounding titles. Biz Dev, consultant, manager, and so on.
The big caveat to this data is that it excludes "Salespeople not categorized elsewhere" and "Managers not categorized elsewhere" since they considered those categorizations too generic and they would be the top jobs in every state.
Whenever you aggregate data, particularly about human behaviors, the grouping is always going to be somewhat arbitrary.
An underlying assumption, which we can make because the consequences of being wrong are so low (god forbid we be entertained by an inaccurate infographic!), is that all occupations will lose some people that should be counted with them, and gained others they don't deserve, and that this will balance out.
It would be interesting to learn if there is some sort of systemic bias, say jobs commonly held by women were more or less likely to be lumped together, or jobs with more itinerant workforces weren't counted well.
If you define an arbitrary worker type 1 as X, and then graph the % of the workforce, average hours, or change in pay you can see meaningful trends.
However, if you create arbitrary category's and then just look for the largest group or average salary etc. then how you define each group becomes extremely important.
IE: Heart Surgeon with GP ends up an average that nobody really fits.
It does appear, if the citation of IPUMS at the bottom (a cool source of data, for those who haven't seen it: https://usa.ipums.org/usa/ ) means anything, that they are at least using standard classifications.
That doesn't mean their distinctions make sense, but at least they're openly available and documented.
I think it may be the reverse, anyone who is holding a CS degree and doing computer-related job, like an IT support guy in non-software, non-tech company, counts as software developer.
For me, it makes the entire chart unbelievable. For example, they separate out teachers into two groups. I find it unlikely that there is a bigger difference between high school and elementary school teaching than there is between installing (programming?) a PoS system than building a video game.
I assume the only reason for the teacher split is tthat they didn't want it to be the top job of every position.
You can get certified by to teach both primary and secondary. It takes very little additional effort. No courses and very little studying if you are already familiar with the material. I know some people who do both and they are more interchangeable than you may think.
Analogy: Front end developer, back-end developer, and full-stack developer. It's really just a focus, they are all developers.
The map clearly shows that software developers are the most common job in at least 3 states, so that is my point. Your attitude about me needing to get out of my bubble is unwarranted. I am referring directly to the map in this article.
As interesting and insightful as this is, I'm very curious about how much the chosen level of categorisation affects the final result. For example, I notice that "registered nurse" and "nursing aide" are considered separate professions. That may well be valid (I assume it's to do with level of training and responsibility? I just picked the example because the jobs sound like they're similar, I'm not picking on nursing in particular.) but if you were to combine them, would they suddenly become the most common job in additional states? And presumably every job could be further subdivided (There are few "secretaries" remaining in part because of technological advances but also because that job has been somewhat rebranded into a bunch of more specific titles.), so where you draw that line might have a big impact on the ranking. Without visualising actual percentages (and/or maybe showing multiple jobs that rank highly), they're really limiting how much this map is telling us.
I would put it more strongly: The statistic is arbitrary because there is no method according to which the bins are chosen.
Sure we can still glean some information from it. But the "most common job" is simply the largest bin for which nobody found it necessary to subdivide it further.
What if we lumped "customer-facing jobs in brick and mortar stores" together. Cashier, waiter, retail salesperson etc. etc. Suddenly that might be largest bin.
What if we subdivided urban truck (small truck?), medium, long haul? And split off agricultural use (tractor) too. Maybe it suddenly won't be largest bin anymore. For the purpose of the automated driving discussion urban settings are probably going to be automated much later due to the difficulty of navigating pedestrian zones and things like that.
>urban settings are probably going to be automated much later due to the difficulty of navigating pedestrian zones and things like that.
For those who haven't seen a recent Mobileye presentation[1], they're already well along on this problem with a BoM of only a couple grand in cameras and microchips. At that cost point and safety gain it won't take long for the regulators to just mandate it.
I was trying to say that in the context of discussing job displacement by automation long-haul truck driving might be arriving sooner because its environment is simpler and there are fewer tangential requirements.
And thus there would be a valid reason to split the bin for the purpose of that discussion. Which in turn would alter the results of those statistics.
Just to expand on nursing, those 2 are actually really different jobs so they need their own category. A CNA needs a 3 month program/night school and that's it, for an RN you need to go to school/college for 4 years. There are far more CNAs then RNs but some CNAs do end up becoming RNs so there's that.
RN = registered nurse. To be an RN you have to pass nursing boards, and then you are registered to practice nursing in the state(s) you work in.
There are two year and four year programs to get your RN - the two year programs are an associates degree, 4 years are a bachelor's (BSN - bachelor's of science in nursing).
Most hospitals are trying or would like to try to phase out hiring nurses who only have associate's degrees, but supply and demand (and maybe unions in some states) have prevented that from getting too far.
Of course, on top of this, some areas also have LPNs (licensed practical nurses). These people usually also have an associates degree, but they haven't taken the RN boards. They are closer to RNs than aides or CNAs, but don't have quite the same scope of practice as an RN.
And then you have DNPs (doctors of nursing practice) who are similar but not equivalent to PAs :)
The set of regulations around different healthcare professions is really quite interesting. You have some relatively standard set of professions, but each state does licensure and scope of practice a little differently, so you end up with a huge matrix of what you can and can't do depending on your title and location. Unions will sometimes add their own restrictions on top of this.
Pretty different from the world of software engineering, where there's no real licensure or certification and anything goes.
(1) The disappearance of factory jobs (machine operator, assembler of electrical equipment). The overseas flight of US manufacturing as a consequence of free trade agreements is real; the causes, consequences, and whether it's good or bad is a topic for another post.
(2) How common truck drivers are. I realized some years ago that it's one of the few jobs left that lets people earn a decent income without much education. Also, the job is a prime candidate to be replaced by automation (self driving vehicles) and the economic displacement of all sorts of driving jobs could possibly lead to social/political turmoil.
1) The disappearance of factory jobs has less to do with "overseas flight" and far more to do with automation. In years past, even a small manufacturer would have one machine operator per machine. These days a single CNC machine can outperform a roomful of manual machines and only require a single operator.
2) Living in MN, I was surprised by truck driver being on top since I only know one actual OTR driver. However, this is a very agricultural state, and I suspect that the majority those truck drivers are farmers who make most of their money driving trucks commercially. A lot of farmers need a CDL by necessity.
Living in MN, I was surprised by truck driver being on top since I only know one actual OTR driver.
"Truck driver" covers a lot more than OTR. The guy that drives the Budweiser truck to your local grocery store. The Pepsi truck driver that restocks your local 7-11. I dunno, does "Snap-On truck driver" count, even though he's really a small business person selling tools to professional mechanics? UPS driver falls into that category, I'll bet. Friend of mine used to own his own dump truck, hauling for construction jobs. And so on.
>I suspect that the majority those truck drivers are farmers who make most of their money driving trucks commercially
It's even more simple than that - the long description is "Truck, delivery, and tractor drivers", a category it wouldn't take much imagination to stick a lot of agricultural employees in.
> The disappearance of factory jobs has less to do with "overseas flight" and far more to do with automation.
Do you have any data on this? Personally I would lean more towards poor trade policies and imbalances but would be curious to see some numbers on automation.
Best to look up the GDP figures for the manufacturing sector. Essentially, the US is #2 in manufacturing by a huge margin, while employing a fraction of the workers of China, #1.
Correct. China makes a lot of stuff, to be sure, but most of it's pretty low value, like toys and raw materials. The iPhones are actually the exception, statistically. The US still makes just as much stuff as it ever did, just with far, far fewer people.
China makes the aluminium case for the iPhone and the glass. The key chips and sensors come from Korea. The cameras, radios, and RAM from Japan. Some bits from Taipei. The software is written in California.
On hand data, no. My "data" comes from having seen this topic beaten to death and from spending a lot of time on manufacturing forums populated by actual manufacturing business owners and from knowing people who have owned manufacturing businesses.
Relevant quote "I don't want to leave you with a facile explanation, but for the purposes of space, I think it's acceptable to say that as manufacturing and agriculture got more efficient, they required fewer American workers, "
>Also, the job is a prime candidate to be replaced by automation (self driving vehicles)
I'd be willing to bet my soul to the devil that this is where the long con is in the automation deal. The first trillionaire in commerce is going to be the entity that owns this.
"It puts the lotion in the basket, and it drives itself to the appropriate shop"
The most common jobs in the Northeastern states seem to shift toward people jobs: nurse, nursing aide, primary school teacher. These involves complex tasks requiring tight coordination between motor, natural language, and emotional skills. Thus they are unlikely to be replaceable by technologies in the near future. (From my knowledge of AI, these sets of skills are probably even harder to automate than those of software developers.)
Nursing and healthcare jobs in particular will clearly be in higher demand with aging population in all developed countries (and several developing ones).
Many of these jobs require moderate formal education and average cognitive skills, but also a certain set of aptitude (e.g., empathy, communication, cooperation) best inculcated at home and in one's culture. If more analysis shows that they are indeed the most available jobs in the future, perhaps we should start preparing the youths now (especially since these skills are themselves valuable in life and in most other careers).
I had a professor pull this visualization up on one half of the screen, and then a news article about Mercede-Benz' self driving semi on the other half of the screen.
He left them up for a about a minute, looked at all of us, said "All of your actions have consequences, some of them unintended." and moved on with the lecture for the day.
A friend of mine working as a transport economist produced used modeling and projections to estimate things like traffic congestion, road usage, miles driven and so forth over the next however-many years (this was in the late 90s and the effects were striking -- as in bad). His boss was startled by them and said "what about the effects of the internet and virtualization?" "That's what you're looking at."
A whole lot more stuff is being delivered to people's doors today than twenty years ago.
I remember seeing a documentary about why, even though rail is more efficient, it has not replace long haul trucks.
The main reason had to do with how our rails are setup in an inefficient way and that all freight currently has to stop route through Chicago (iirc). This is a manual process where cars are manually moved, counted and routed.
Many former railways I know have actually been abandoned as I have an uncle who worked for Rails for Trails converting them into bike paths.
Rail is still incredibly cheap and efficient in the USA. In fact, the USA has some of the most cost effective freight rail in the world. [1]
Rail freight is also quite a bit larger than truck freight in the USA. A lot of things are railed across in an intermodal and then hitched onto a truck for regional delivery. [2]
In short, rail freight is massive in the USA and our primary mode of goods transportation.
What is the connection between those two? Or conversely, how does the presence of human drivers prevent moving cargo onto freight trains? Isn't that what standardized containers are for?
ISO 20/40 containers are not optimal for long haul over the road trucking -- you generally want something bigger (53') and lighter. You can rail those on their own flatbeds, but that's also not optimal. Intermodal ISO containers make a lot more sense for sea and (sort of) rail; some sense for trucking, especially final delivery, but not OTR; very little sense for air freight (although I've seen a bunch of specially configured equipment shelter 20' containers shipped in cargo aircraft, it's not done for general cargo because the tare weight is in the tons when you're paying $1-20+/pound; they also don't work well with aircraft cargo handling, and can't fit as standardized ULD or loose cargo in passenger aircraft.)
Repacking your cargo from one container to another sucks for a lot of reasons: cost, time, theft, damage, etc.
There's also a latency and performance improvement if you can go door to door with a truck, vs. do intermodal with a train in the middle. Unless you're moving a bulk commodity to a high-volume location, it's unlikely they have rail directly to the destination. (Coal power plants do; lots of modern light industrial manufacturing sites do not.)
Optimizing air vs. truck vs. rail vs. ship across your whole supply chain is an interesting problem.
Another fun thing: with robo drivers, you could go essentially non-stop (except for fueling) door to door. Otherwise, you need a team driver or relay concept (usually you carry 2-3 drivers and swap them enroute; you could also do something with depots where you swap out drivers while retaining the vehicle, which I think is done for some high security convoy operations.) That makes the truck vs. rail advantage even greater.
One unintended side effect of automated trucks may be pushing more long-haul freight to rail. Fuel and driver's salary are the biggest expenses in operating a truck -- remove the driver's salary and fuel becomes the dominant expense. Right now, rail is around 4 times more efficient than trucks, so if fuel is your dominant expense, moving to rail makes more sense.
Of course, self-driving trucks could help with truck fuel efficiency by driving at a more fuel-efficient speed, and they may be deemed safe enough to haul more trailers behind a single truck. Though whether or not human drivers (even if their cars drive themselves) will put up with a 4 trailer 40mph truck on the freeway is a different question.
I would expect that anyone who's worked on enterprise software has been to client meetings where ROI is expressed in terms of human resources.
It's part of advancement that certain jobs will no longer be viable for people. People will train for other work instead.
It's not like suddenly, with no warning, 100s of thousands of drivers around the world will need to find new jobs. But in the very long term, fewer drivers will be needed. People who would have, at one time, become drivers, will do other work.
Most people who are taking classes from a professor haven't built enterprise software or been in client meetings. Most of them don't realize that their work while it probably makes life better for young rich white yuppies like them may destroy a way of life for huge swaths of other people for better or for worse. And generally college students haven't been trained to be cold or callous enough to express ROI in terms of human resources quite yet -- I'm sure the slide was gentle preparation for what is to come once they enter the capitalist workforce.
Heh, for auto-driving trucks it very likely will mean suddenly 10's of thousands out of work.
And folks out of work are finding it harder and harder to find word. Automation's ultimate goal is putting us all out of work. We're in a transition, and admitting that will help guide policy from here on.
Its not the case because its unlikely that the technology will be adopted everywhere, simultaneously, overnight; it will be adopted gradually over time; it'll probably first manifest in the job market as a decline in growth in trucking jobs, then a decline in trucking jobs, rather than going from the existing pattern one day to no humans working as truckers the next.
I certainly won't be overnight. But you are assuming current truck drivers will have time to migrate to new jobs as those appear - I really doubt two decades is "gradual" by that measure. Yet, I can not see how it could take that long.
> But you are assuming current truck drivers will have time to migrate to new jobs as those appear
No, and I'm not even assuming that new jobs will appear.
That there will be warning and that there will be much that can be done with the warning are very different ideas; OTOH...
> I really doubt two decades is "gradual" by that measure.
Two decades is enough time for people who are mid- (and often even early, though not quite at the beginning) of a career to retire and not need to worry about a new job, so its potentially gradual enough.
The kids in college don't make the big decisions on how the economy works, so these types of liberal guilt speeches always seemed hollow to me. Sure, maybe someday, but at that point market forces will be calling the shots, not personal preference or pet political activism. If it wasn't for automation, modernity, etc would Mr Professor even have a job? Before various cost savings techniques of the past 40 years, the amount of people who went to college were a much smaller percentage than the amount today. You can't have these enrollment numbers with tiny classrooms and each professor having his own secretary. It just doesn't scale.
Its amusing to think that some people think of themselves as above the economic fray. Sorry, but we're all in this together.
To me, they aren't "liberal guilt speeches", first off, just a recognition of cause and effect.
But the real problem, I think, lies somewhere in your interpretation of of the the goals of pointing such things out. I'm sure "Mr. Professor" is more than capable of recognizing that kids in schools aren't political leaders. "Mr. Professor" is pointing this out, not because they fancy themselves above the economic fray, but precisely because we're all in this together.
That, in fact, is the actual message being conveyed by what you perceive as "liberal guilting".
If you're looking for people who act above the economic fray, a far better source of those sorts of airy pronouncements would be, for instance, David Brooks or Tom Friedman, for whom massive economic dislocation tends to be a "to be sure" or "perhaps" tacked on to the end of a column about how great disruptive innovation is.
I looked at the graph and saw the statistics, but your short and simple comment gives me pause. If in fact we achieve autonomous trucks it would mean we might potentially displace the most common job in 27-odd states.
We (the USA) as a country should begin thinking about the ramifications of this and how we might need to change to preserve not the economy, but society.
I'd guess we see tractors as the first step. They typically drive so slow the controller only needs to have stop button.
If a farmer could simultaneously monitor just 4 tractors, you could make tractors half the size while having twice the work done. This would double the number of sold tractors, which gives economy of scale. While also making assembly lines lot smaller and simplifying logistics.
Trucks don't scale in similar fashion. Inter-modal containers are lot bigger than ploughs.
I suspect more likely would be some kind of security team to prevent the truck's cargo from being jacked? Unless, of course, the driverless savings assume some loss of contents.
Probably neither. I doubt the presence of a driver today is much of a deterrent to thieves. Most theft from trucks is by people working in the warehouse.
There's not much opportunity to steal stuff from a truck except when it's being loaded/unloaded. That doesn't change whether or not it's driverless. If you're planning on hijacking it enroute, then whether or not there's a driver makes little difference. cf, carjackers.
You base this on what? Someone wanting to go to the effort of stealing the contents of a very large vehicle that is already being tracked by GPS and is for the most part in public view is not going to be deterred much by a human driver.
Scenario today: truck stops at rest stop, driver gets out, truck is stolen.
Alternate: truck stops at rest stop or for traffic light, driver does not get out. Robber gets in passenger side, points gun at driver's head and tells him to GTFO. Drives off about as fast as a loaded semi can, which is not very.
Scenario with automated truck: truck never needs to stop at rest stop except for fuel and will need a human attendant (unless we also automated those away). Odds are, there are no windows or steering wheel/human controls anyway since it's automated. Now the robber needs to unload a trailer onto another, waiting truck which is a time consuming operation. And take off with that one. All the while the automated truck is reporting its location, and that it's being tampered with.
So yeah, I don't think crime will change much. It will still be far, far easier to steal stuff from the truck when it's being unloaded. And that has nothing to do with a driver being present or not.
Not everyone has what it takes to point a gun at a driver's head (actually, not everyone has a gun in the first place). If the truck is driverless, then it removes the courage factor, the inherent risk of having to deal with a human, and now it only takes cunning and intelligence. Since it is driverless, you could even use a RPG rocket against it, without risking a human life. Maybe, I played too much GTA. That being said, I find it hard to predict whether crime or security will prevail.
I'm fairly sure if you just park your car in front of the truck on a remote highway the drive safety systems will kick in and stop the truck. Then you jump out and crowbar/bolt-cutter the container.
Of-course it will probably detect you doing that and call for help complete with its GPS position.
I also expect some enterprising think-outside-the-box robbers will temporarily come up with some creative way round this. (Will phone jammers stop the truck calling for help?)
Well, you'd need to force the cargo hold/container open; there's no driver to threaten. Assuming it's locked, you'd need some serious power to do that. A tractor or explosives, perhaps. I don't think many will bother.
Right now most trailers are secured with either a simple padlock or a metal ribbon with a unique number stamped on it (so if it is cut, the person signing off will know).
Is theft from trucks on the move a real issue? Especially if you consider thousands of identical self-driving trucks on the road? Who would know which contain valuable cargo and which don't?
Companies used to have a single secretary for each executive, manager, and other mid level worker. Mad Men isn't exaggerating.
For the very high end employees, sectaries morphed into admin assistants. Executives and other other very high level people still get a single assistant.
But the junior executives and middle management don't have a designated assistants. They may share one with many other employees. But most of them just do a lot of the work themselves or have underlings do it in addition to their other work.
I work in the legal field, one of the few that still has secretaries. Our office space is set up with a secretary desk right outside each attorney office. But now I share my secretary with NINE other attorneys.
My mom is a secretary. She's gone from working for one executive to essentially being the secretary for the entire staff.
So yes, the title and role changed. But there still has been a massive reduction.
No, but the jobs they use to do have been replaced to varying degrees. So, a company needs far fewer of them (and more likely just munge their tasks into something else).
Some secretarial tasks much less common these days and/or mostly automated so primary worker does them themselves.
Typing, retyping, dictating, managing calendars, meeting scheduling, appointments, taking voice messages, dealing with mail(physical opening/organizing), internal memos, filing, searching for stuff in those files, copying, faxing.
I work in a 100 person company and we have no dedicated Admin assistant, just a couple employees who act as receptionist, HR, office manager, and part-time admin assistant to the CEO. so 2% of our workforce falls under the "Admin assistant" role. If we didn't have extended office hours that needed two people to cover, we'd probably have only one person in that role.
So while Secretary/Admin Assistants haven't disappeared completely, there aren't as many as there used to be.
The tasks are different. A secretary back in the 70's and earlier was primarily a typist/dictation taker. An Admin Assistant is closer to a small office manager than a traditional secretary.
This info graphic shows more the specialization of the workforce, or at least of titles, between 1978 and 2014. I'm ready to bet there's almost the same percentage of truck drivers in 1978 compared to today, it's just that people who would have been 'just' secretaries have split into marketing assistants, HR workers, accounts receivable, sales people, etc, and the total number of white collar workers hasn't changed that much. However, a truck driver is a truck driver is a truck driver, let it be a dump truck, a big rig, a van or whatever.
Farms have consolidated - 300 acres used to be a big farm. Now you have to have 10,000 and a workforce of 100 people to farm profitably. So the numbers have reduced.
The road I live on had 13 farms when I was a kid. My brother was the last. Now he works in town and rents the land to a corporation.
Sad point you bring up. Agriculture has morphed into agribusiness which consists of large corporations with no personal connection to the land growing Monsanto clones fertilized by ancient guano deposits and ammonium nitrate synthesized from methane. Strange new world we now live in.
Subsidizing farmers so heavily that even our socialist welfare pales in comparison. Here 50% of farms make 80% of produce. The rest just keep the land from accommodating wildlife.
Although idyllic farmland communities are a thing of the past (if they ever were) the primary intended reference wasn't to sociology or even to business model. A question arises as to what will be the long term impact on general health of a lifelong consumption of cloned plant material raised on a steady diet of junk food consisting of mined potash and reconfigured natural gas. Not to mention the vulnerability of that homogeneous plant stock to a single new pathogen.
Amusing way to put it. Not really the case, as plants are not 'made' of the elements they consume biologically, and plant stocks are designed to be resistant to pathogens better than the 'wild' varieties.
Really? But these plants are not being grown for aesthetics but as a food source. Nutrition value (read vitamin, mineral content, et al) of contemporary produce is considerably less than it was 50 years ago. If interested see a sample low-tech Scientific American article on the subject.
And for the sociology: what percentage of the population are responsible enough to take the now necessary dietary supplements to maintain a healthy, productive physiology?
Amazing. GMO tech is probably the only methodology with any hope of combating the impact of that single new killer pathogen when it arises. Your grandfather must know about fungal pathogen Ug99 devastating the world wheat crop and that only about 10 percent of other wheat varieties are resistant. Question is: why isn't a mixed variety regimen a general rule? Given that it takes a long time to introduce pathogen-resistant genes into a single plant variety, dependence on single or a mostly clone stock (a general rule based on economics) could prove fatal at some point.
Although, a real problem with the current GMO tech is the diminished nutritional value of the foodstock produced relative to historic produce. But that, of course, is due in part to depleted soil.
It's a very blue collar job concentrated in mid-low to low income areas. There is a heavy separation here.I went to visit family of my girfriend's in El Paso and was surprised to find out a good portion of her uncles and cousins were all truck drivers. If you don't generally hang out with blue collar types, you aren't likely to meet any truck drivers, but once you go into an area where income is on the lower end of middle class you will likely find that they are fairly ubiquitous.
Don't take away my pinch-zoom, and especially not when you've sized the page so that I can't see the date and the mal at the same time.
Honestly, I wish there was a way to tell Android Chrome to ignore the html tag that blocks zooming - using it is a moronic act of hubris and it frustrates me every time.
I'd counter with: complaining something doesn't exist, without making any effort to see if in fact it does exist, is the real "moronic act of hubris" here
Thanks for pointing this out. It also covers all the objections discussed elsewhere here (e.g. that some professions are fine-grained, others lumped together, and that two categories were omitted for being too huge and vague).
What is by far the most interesting is that software developer is becomming the most common job. So if in 30 years secretaries was replaced with smart phones and word processing software, soon truck drivers will be replaced by self driving cars, then in 30 years (probably less) what will software developers (then maybe the most common job) be replaced with?
It depends on where you think we are on the technology curve. If you go by projections based on the rapid advancement of the aerospace field from 1940-1970, we're about 20-30 years late on having regular hypersonic flight between DC and London. Same story for things like fusion.
AI as a field plateaued somewhere in the 1970's. Much of the recent innovation has come from applying existing, understood techniques to ever more amounts of data (enabled by faster processors, and data sources like GPS and digitized map data). But it's possible that AI that can code programs will require fundamental leaps in the field. In fact it's quite probable--we don't even really understand what such an AI would look like. At least with things like fusion or hypersonic commuter travel, which are still far away, we understand roughly what the solution would look like, we just face massive engineering challenges.
So, my takeaway is that if the transportation is automated, our country will have a "Few problems" with regards to unemployment, taxation, and human welfare.
If? Make that when.
So, the next question, considering all the companies entering in this landscape: How do we change our country to handle a mass forced exodus into unemployment? What do we do with 5 million unemployed?
Worse yet, what do we do when we are looking at 50% unemployable? (Not unemployed, unemployable. As in work done at free != absolute lowest cost of food/shelter) And when that number climbs, where do we go? What do we do?
The change won't happen over night, it will be gradual. Fewer people will pursue truck driving as demand dwindles, instead training for a more in-demand field.
Software development is taught in elementary school now. So I suspect a lot of the next generation will pursue that.
The underlying issue is machine learning is making more and more jobs unemployable. Those jobs would not be economical for humans to do at any price (ie: $0).
And saying that software development is going to be it is simply fallacious. It's obvious to me that most software development is also going to be eaten up by lower levels of machine learning. And also, software dev is very easy to export and 'grow' in the lowest wage countries you can find. And import is easy, unlike any physical good.
And it still leaves the big question: What are we going to do with massive unemployment?
So "computer analyst" replaces Truck Driver in 2002 in Colorado, and then goes back to Truck Driver in 2004? Presumably this is census data, I see is the https://cps.ipums.org/cps/ IPUMS data set.
I'd be interested in figuring out ways to cross check that set. On the surface it looks like self driving trucks would throw most of the country out of work.
Truck drivers, huh? Well, I'm sure there's a convenient free-market solution for all of these people losing their jobs in the next 10 years to self-driving vehicles. Maybe 'automated truck refueling technician'?
On a meta level, this is probably a good example to study for the impact of grouping data into classes (most likely the truck driver is just a job that is more easily groupable than the different types of office jobs etc).
Also, secretary was a category that covered job descriptions ranging from phone operator to business analyst. Factory worker probably covers as much variation as office worker, if not more. And there's little inherent reason to separate "primary school teacher" from teachers at other levels
I don't think some jobs are inherently more groupable, I think the choice of grouping is purely arbitrary. We could just as easily have separated long-haul truckers from local delivery people and these charts would completely change.
If there are 100 people and 2 are truck drivers, but the other 98 all have different jobs, then truck driver is most common, but there's still not a lot of truck drivers.
This image is interesting but lacks any context (including what the asterisk is for which is what initially bugged me) so I searched and found the actual story which has the graphic and explains what it means including that the asterisks indicates: "We used data from the Census Bureau, which has two catch-all categories: "managers not elsewhere classified" and "salespersons not elsewhere classified." Because those categories are broad and vague to the point of meaninglessness, we excluded them from our map." http://www.npr.org/sections/money/2015/02/05/382664837/map-t...
The article that the submitted interactive map is part of [1] explains some of the things that people are speculating about here.
An important note: they are using Census Bureau data and categories. The CB has two broad categories, "managers not elsewhere classified" and "salespersons not elsewhere classified". They map makers left those out because they consider them broad and vague to the point of meaninglessness.
• There are so many truck drivers for a few reasons. One is that it is much less affected by globalization and automation (so far) than most other jobs, and another is that the "Truck Driver" is a very broad category in the CB data. It includes delivery drivers, for instance. Some other large jobs are split across more than one category. Teacher, for example, has separate CB categories for primary school teacher and secondary school teacher.
• Secretary rose in the '80s as the economy shifted away from factory work and toward office work. Then the personal computer took away more and more of the work that secretaries did, so Secretary fell.
According to Marx, everything of value is produced by human. Some humans (like engineers and programmers) can just produce more valuable stuff than others.
If we believe Marx and that jobs are going to disappear, that requires either rich people not using money or money disappearing from the system. Currently inflation encourages rich people to use money in some way or another. And we have no reason to expect deflation.
Statistics based on classification systems are entirely ontological artefacts -- how many of X you have (or how many X you have) depends entirely on how you subdivide space.
In an earlier look at this story, I went through a set of occupational categorisation and census data dating back to the 19th century. As might be expected, there's been a tendency for the total number of occupational classifications to increase over time.
As might not be expected, the high-water mark of classifications isn't the present, but the classification scheme used for 1910 - 1920 census data.
If truck driver is a most common job then the future looks pretty bleak. There are billions of dollars behind making driving any and all vehicles obsolete.
I wonder if people know that Google, Amazon, Tesla and even Apple are hard at work at killing their jobs!
Don't get me wrong, driver-less future is a nice future! But a lot of people may get caught off guard.
This is quite likely an example, from a data modelling point of view, of an extremely slowly changing dimension driven by historical regulations, laws, and taxonomies.
It may still be meaningful for a few industries, but not in general.
I live in Utah. Its now coined "silicon slopes". There are a lot of tech companies out here. Driving down the highway, most of the billboards are "I'm an awesome tech company and we are hiring".
Makes sense when you realize Novell and word perfect were headquartered in Utah, and both are essentually gone now leaving all their former employees scattered around the state starting companies.
