As robots take more jobs from humans we can expect a massive unemployment problem. Let's start planning for that now. I'll limit this discussion to the United States just to keep things simple, but the concepts should be applicable everywhere.

I'm about to describe a huge government program. Please don't leave comments saying huge government programs are never a good idea unless you have a better idea that doesn't involve the government.

So let's say the plan I'm about to propose has about a 20% chance of working. That still beats the do-nothing option of massive unemployment leading to certain doom.

My plan is to turn the United States into more of a tourist attraction than it already is by building vast networks of interconnected canals across the nation. These canals would accommodate hotel and residential houseboats. And let's say the houseboats are all computerized so you just plug in your destination and go have a drink on the top deck while the boat does the rest, including making its own scheduled maintenance and refueling stops.

I'm stealing the Google self-driving car concept and applying it to a canal network with houseboats. The boats would be aware of other boats and obstacles and avoid them.

Now let's say that by law the only boats allowed on the canals are the self-driving and highly "green" types that don't pose much risk of polluting the canals. Then the canals become a solution to water shortages across the country as well, so long as they are fed by Canadian water sources. So this project is also a partial solution to climate change and the water shortages. And because these canal routes will crisscross the country, perhaps it makes sense to build out the next generation of our energy grid along the same rights-of-ways.

I imagine all of the house boats on the canal being built with a common docking standard and individual identity beacons. If you want to dock with a friend's boat for the day, just enter the identity numbers for your two boats, wait for the confirmation, and the boats do the rest. The two boats might be miles apart when you program them, and they adjust speeds accordingly to meet. You'd use the same method to dock with restaurant boats, gift shop boats, and other service crafts.

I can imagine that a portion of the house boats are rented to tourists and another portion are full-time homes for retirees and people who just prefer continual travel. Each boat will have full Internet access so folks can work and travel at the same time.

I would think that at some point the cost of a houseboat would be far below the cost of a home because there is no land involved and the houseboats would be energy-efficient. So this could also be an answer to affordable housing.

Obviously robots would be a big part of the labor force for a massive project of this size. But you'd also need huge numbers of humans for planning and implementing. And if the human workers for the canal project are the first occupants of the houseboats, their cost of living might be so low that they become competitive with robot labor. A big reason that human labor costs so much is that our lifestyles are relatively expensive. The boating lifestyle could be designed from the bottom up to be inexpensive (yet awesome). When humans can live inexpensively, they can charge less for their labor and compete with robots for a bit longer.

This one massive project would modernize the energy grid, solve the water crisis, expand tourism, create affordable housing, stimulate the economy, solve unemployment, reduce shipping costs, and make travel affordable.

And it would touch every part of the country. Every town along the canal system would want to create a commercial center and docking area for boaters.

What do you think?

[Update: It occurred to me that you could also design the canals with power generators at the bottom that are powered by the current. And you might as well lay fiber along the route while you're at it, with wifi towers all along the route. Oh, and let's add lots of edible fish to the canals so boaters can catch as they go.]


Scott Adams

Co-founder of CalendarTree.com (Scheduling made simple)

Time.com says my latest book is not crazy.



Rank Up Rank Down Votes:  -47
  • Print
  • Share


Sort By:
Mar 25, 2014

"[...The powers that be have presented many a decreasing percent unemployment number as a wonderful thing that seems to imply job growth, but when you look behind the curtain, the dominant change that is driving the number down is that many unemployed people are giving up and dropping out of the workforce, so they don't get counted...]

That may account for some of it but according to http://data.bls.gov/pdq/SurveyOutputServlet?request_action=wh&graph_name=CE_cesbref1 there is indeed real job growth the past three years. "

Of course there has been some real job growth. No conflict with my comment. Job growth just hasn't come anywhere near tracking growth of corporate profits during this recovery, which it did in previous recoveries. Plus, the labor participation rate has fallen, meaning that job growth has not kept up with the 125k of new workers added every month. Many of the long term unemployed have gone on disability when their unemployment benefits ran out. Makes it look like unemployment improved much more than it actually did. The Capital side had a robust recovery in 2009, but the Labor side is recovering much more slowly. My point was that we are currently in the middle of a major labor dislocation caused by technical innovation, both hardware and software. The difference this time is that the incremental incorporation of new technology did not precipitate the expected labor dislocation until a sudden change in the economy (the housing crash) overwhelmed corporate inertia that had held staff levels up. The new staffing ratios are permanent, so the recovery will not put everybody back to work. Full employment will have to wait for business expansion, which is occurring at a reasonable clip, at least in the USA. By that logic, this is one of those times that we should expand the safety net and expect high unemployment for a while.

So far, so bad. Nothing really new here. The media's various ideological blinders are firmly in place and have prevented them from telling us that we just went through a major labor dislocation caused by technological innovation - one that is unrelated to politics of any stripe. I suspect that a lot of people would be less apprehensive and act more appropriately if they understood what is going on - growing pains.

The scary thing (for me) is that the rate of innovation has been increasing for a very long time, so at some point I expect it to overwhelm the ability of the economy to expand enough between disruptive innovations to put everybody back to work. I don't understand why there exists huge resistance to this idea among the technical community! Are there no more apolitical pragmatists among us? The new reality will (I think) saddle us with continuous high unemployment and underemployment, really sucky economic times in general, unless we figure out a way to get a proper chunk of business earnings back into the pockets of the population - who will not be employees - in a way that rewards actual value added by each individual's activities. In essence, business will need to fund the activities of humans who are doing what business needs to have been done so the economy can thrive in the future and technological progress can continue. This cannot be part of a general safety net program.
Mar 25, 2014

I think I agree with most of what you're saying. But...

What does a human being -- even at minimum wage -- cost per year? It's not a small amount. $15,000 at the lowest, I should imagine, for what amounts to 40 hours of work performed per week (out of 168 total hours in a week). And for round-the-clock workers, that's a $60k one-year ROI for a single robot. If you figure that 3 years is a reasonable ROI for a business to invest in automation, that's $180k very-reasonably justified for buying a robot. That's just labor; what of other benefits? Of that, we can probably attribute less than $20k to hardware.

If you take a specific task -- like washing dishes -- there's no reason to believe that a general purpose android couldn't be programmed to handle everything that it's likely to encounter. Will it take a lot of programming? Yes. Is it an impossible amount? Well... With a US market of tens of thousands of dish-washers, with a roughly $150k per-unit value in that software, I don't see where there's an economic hurdle. There are literally hundreds of millions, if not billions of dollars available.

Rinse and repeat for dozens, if not hundreds, of jobs.

And since it's just software, and you may not need round-the-clock dish washing, why not load "DishWasher v2.5, and CleanTheGrill v1.3, and TakeOutTheTrash v1.1" all on the same robot? Put a cron job on it to take out the trash every 2 hours, clean the grill (round-robin) every 4 hours. And maybe have it run "MakeFrenchFries 3.7" every time the fries gauge triggers a low state alarm.

I'm not even suggesting that robots are going to do everything that people do, or even need to make intelligent decisions. They could just reduce the amount of labor necessary, and replace many low-skill people with one supervisory person with a toolset to intervene.

Secretaries are mostly a thing of the past because of desktop computers. They're not as useful as a secretary in many cases, but do enough that we all learned to type and use simple tools to collaborate and coordinate. The machine does enough of a secretary's work that it was worth our time to "fill the gaps" with some new skills.
Mar 25, 2014
Hi, it's me again, the "robo-doubter" (nice one). Let me clarify a bit.

It's true that automation did kill jobs in the past and continues to do so. But this happened, and still happens, in fields best suited for automation: mass production, repetitive tasks, tasks requiring particular precision, speed, power, mass data processing and the like.

The common denominator of this kind of automation is programming. Robots and machines of today perform pre-programmed activities and nothing else. They do not decide (except pre-programmed decisions). They do not take responsibility for what they do. That means they are still generations away from even the lowliest janitor with an IQ of 80 who can still do hundreds of different tasks without being programmed, because even he has a lifetime of learning in himself, identifying millions of objects in nature or household or office as what they are and how they are applied and how not to break them, including living things.

You might argue that it might be possible to pre-program a machine to do the same, and technically this is true, but it would be prohibitively costly due to the complexity of real life around us, and why should anyone pay for this effort if there are billions of comparably cheap such machines available, called humans?

The alternative to pre-programming would be some sort of self-learning system, but such a thing is nowhere at the horizon. There is no such thing as artificial intelligence or learning. Not even the Turing test (http://en.wikipedia.org/wiki/Turing_test) has ever been passed yet by a machine, and this test does only test for a successful play-acting of AI, let alone real AI.

All what passes for AI today are simply pre-programmed devices of a certain complexity. And the results are as you would expect from that. Look at software translations of natural language - what you get is mostly gibberish, because the machine behind the translation does only perform algorithms, it does not understand language. It cannot react to new and unfamiliar information.

Or look at the much-trumpeted self-driving car. It will not happen (in this or the next decade). For a simple reason: No one will cover the risk. No insurance company will cover the risk of a self-driving car running over a child on the street without the chance to blame it all on the human driver. No manufacturer will run the risk of being sued for such a kind of accident. No, the smart thing to do is present more and more automobile assistance systems to the driver but keep the human driver at the steering wheel and in responsibility.

And that's where the job market is heading too. Jobs will be still done by humans for quite a time, but they will be assisted more and more by machines and systems. Decisions and responsibility will continue to be the human part. Of course this trend will cut into certain human jobs but as others pointed out, new jobs and new opportunities can come out of it too.

But robots as in "self-dependent, self-controlled, responsible, multi-purpose working units"? Not for the next 20 years, and that's optimistic.

[Your comment has the feel of "nothing heavier than air will ever fly." Self-driving cars are already safer than human drivers. And robots that can make decisions better than a janitor are nearly guaranteed by the normal pace of technological advancement, although no one can predict when the crossover happens. My guess is 10-15 years. -- Scott]
+3 Rank Up Rank Down
Mar 25, 2014
I wonder if the same kinds of reactions occurred when projects like the Railroads or the Interstate Highway System were introduced.

>>The math doesn't work. Simply put, there isn't enough water for your plan.

If there was only some way we can raise the levels of the oceans somehow.

[I laughed out loud at your ocean comment. Something tells me that running out of water on a planet whose surface is mostly water and ice isn't the biggest problem with my plan. And the people who say evaporation will be a problem are hilarious. -- Scott]
+4 Rank Up Rank Down
Mar 25, 2014
I think you've underestimated the effort involved with your canal idea. Given the amount of soil that would have to be removed, problems with evaporation, the taking of private land, environmental concerns and the many other things, your idea could never happen.

You forgot to say "don't take infrastructure advice from cartoonists!"

Anyway, I'm from the Great Lakes area and folks in the desert Southwest aren't getting our water.
+5 Rank Up Rank Down
Mar 25, 2014

I like the tourist attraction idea. But rather than canals, lets build a massive tourist attraction based on Roman Gladiator Games. Kills two birds with one stone, literally.

Solves the unemployment issue and the economy at the same time. Massive influx of tourists, no internet streaming forcing people to visit, and take a cut of all the gambling...

In fact, since I hate all the people anyway, forget the tourists. Just do the streaming pay per view, and take the percentage of the gambling. (Outlaw cash and make all money virtual, easy peasey)

Let's get past all this sissy-ness of concussions, and get the blood pumping. Whose with me?

Mar 25, 2014
Seems that I've always heard boats described as "A whole in the water you throw money into." I have a hard time imagining this as a real solution to any one problem, let alone the spectrum of problems listed.

But enough of that. I want to take on the robo-doubters. Automation HAS disenfranchised a lot of the population. It's not that robots will replace all jobs. It's that ever-increasing automation will squeeze out the least-capable people at the bottom of the economy. There will be a time when a person with an 80 IQ and marginal education won't be suitable for any job in our economy because robots will do labor better than humans, and a person with an 80 IQ isn't worth paying for his thoughts. People like us, in this blog, will adapt to whatever our economy needs, because (presumably) we're not in the well-below-average demographic.

Once robots become a consumer-level technology, the programming and development of their abilities will accelerate. They'll cook well-established foods, do janitorial services and other light maintenance, and landscaping. It may max out there for a good while, but the people in these jobs today aren't going to become robot mechanics and programmers. They're in low-skill employment because they don't have many skills -- not generally because they are rocket scientists who enjoy minimum wage work.
Mar 25, 2014

[...The powers that be have presented many a decreasing percent unemployment number as a wonderful thing that seems to imply job growth, but when you look behind the curtain, the dominant change that is driving the number down is that many unemployed people are giving up and dropping out of the workforce, so they don't get counted...]

That may account for some of it but according to http://data.bls.gov/pdq/SurveyOutputServlet?request_action=wh&graph_name=CE_cesbref1 there is indeed real job growth the past three years.
Mar 25, 2014
You had me at "House Boat!"
+5 Rank Up Rank Down
Mar 25, 2014
Coming from Britain, where a substantial part of the country's canal legacy is still intact or has been restored, I initially found the canal idea an appealing-sounding scheme; but as other posters here have pointed out, there are some major problems with it.

Here are some additional ones I can mention that haven't been discussed so far, or have only been hinted at:

- The Rocky Mountains in particular would present a huge challenge for a genuinely transcontinental network of canals. It was hard enough to construct the roads and railways that traverse its various ranges; imagine how difficult and expensive it would be to do this with canals. (Theoretically, you could also tunnel under the mountains, but that would be even more difficult, expensive, hazardous and vulnerable to geological disturbances.)

- Where's all the water going to come from to fill the high-altitude sections, especially after a dry winter? Water flows downhill, so a LOT of compensatory uphill pumping of water would have to be done every time a lock is opened. This would be extremely energy-consuming: there would have to be lots of pumping stations, all !$%*!$% massive amounts of electricity from the grid (which would itself have to be built out extensively). There would also have to be LOTS of locks, aqueducts and/or boat lifts, many of them having to be built in challenging and near-inaccessible terrain.

- The adverse impact on the environment would be huge. Apart from the massive amounts of soil to be excavated and dealt with, migration routes and vulnerable or unique habitats would be obstructed, impeded, disturbed or destroyed, both during and after construction. Invasive species of every type would have a much easier time colonising new territory, carried there on boats or swept along by the flow of water.

- The amount of energy and water consumed, and CO2 produced, to make and move all the concrete required would be enormous. Even if a lot of that downside could be shifted overseas, the environmental cost of it would be great wherever it took place. Supplying most types of the great !$%*!$%*!$ of energy consumed in connection with the construction of the project would probably strain system capacity beyond its current limits.

- Imagine how much capacity-saturating truck and rail (and maybe even marine) traffic will be generated in connection with the construction phase, and how much air pollution -- not only from the CO2 involved in making concrete and manufacturing, operating and transporting all the equipment needed during the construction works, but the atmosphere over the United States would be laden with excess dust not just for a few months, but for decades, as mountains are blown up and massive !$%*!$%*!$ of soil and rubble were moved around the country. And what would you DO with all that spoil? Some of it could no doubt be made into concrete for the project, but by no means all. You couldn't just spread the excess around on the land -- there would be simply too much of it, and even if you could largely separate the excavated soil from the rocks, most of it would be infertile, pebble-filled and useless for agriculture; not the kind of material you'd want being dumped on good farmland, let alone the landscapes and habitats of a national park.

- There would be massive adverse impacts on the people currently occupying or living near the construction sites: homes and neighbourhoods would have to be demolished, contiguous farmland would be bisected, there would be horrible traffic disruption and accidents, and thriving communities would be sidelined as existing routes through them were made obsolete or less economic. The legal and political headaches that all of this would create would be enough to torpedo the project on their own, especially if the project was opposed by some of the states (and it can be guaranteed that not all of them -- maybe even not ANY of them -- would be in favour of it).

- What would you do if (when) environmental and economic impact studies concluded that the adverse effects in a given locality would be enormously damaging? Would you try to push ahead with that part of the project regardless?

- Once built and colonized by boaters, there would be great seasonal bottlenecks as people followed the weather. Many canal stretches would become clogged with the boats of people wanting to park up for the winter, while others languished, devoid of significant traffic for months.

- Who would police these waterways, and how?

- Who would pay for such a massive amount of infrastructure construction and maintenance, and on what timescale? It has proved hard enough in the USA to persuade voters to fund public education adequately; that being so, the opposition to funding such a massive project as this can be anticipated to reach insurrectionary levels.

Sorry, Scott -- it will be better for all of us if I kill this project now, before it even starts. :(
+5 Rank Up Rank Down
Mar 25, 2014
[I can't believe people are still debating whether robots are about to cause massive unemployment]

Maybe it's the robobloggers who are keeping the debate alive so that nobody starts scrapping them.
+1 Rank Up Rank Down
Mar 25, 2014
One word: NIMBY!
Mar 25, 2014
Hey guys - guys - the perfect fish to plant in the canal would be salmon. Because then you could attach fiber to their tails and they could lay the fiber as they swam upstream. And you could plant dill alongside the canal.
Mar 25, 2014
C'mon, Scott. This one is beyond farcical. Can't you be at least a little bit real? Do you hold us in so much contempt that you put forth something as inane as this, and pretend it's a real proposal?

Stop wasting our time.
Mar 24, 2014
"I also think your premise is flawed; robots wont cause massive unemployment in the future any more than automation causes massive unemployment now.

[What is your prediction for how unskilled labor finds work when robots are a better option? -- Scott] "

Notice that this is the same reason why a high minimum wage freezes out the lowest tier of the unskilled labor pool from ever getting a job, because their labor value is below what an employer must pay. An employer will only hire a worker whose labor is worth what he must pay for it. Like the human worker that Scott refers to above, if an unskilled worker's labor costs more than the equivalent labor of a robot, the human simply won't be able to hang onto a job.
Mar 24, 2014
Prior to the invention of the railroad and the steam engine, canals weaved all over the United States moving goods on very small boats pulled along by teams of horses. Most of these canals were simply abandoned, and eventually filled in to make room for the expansion of the highway that was immediately next to it, which was required for the horses. Some sections of the canals can still be seen though. When I lived in Cincinnati, there were small towns north of Cincinnati that you could visit and still see a section with water. The canal was only about 12 feet wide, and about four feet deep. The one that I saw connected the Great Lakes to the Ohio River basin.

And to the poster above, I can't speak for Canada, but every river in the United States East of & including the Mississippi River, is indeed dammed and locked every 50 to 100 miles. I used to live within a few miles of one of the largest; McAlpin Locks & Dam just West of downtown Louisville, Kentucky. There is a 10 megawatt hydroelectric plant there as well, and it all sits parrallel to the natural Falls of the Ohio.
Mar 24, 2014
"As robots take more jobs from humans we can expect a massive unemployment problem"

Couldn't get past your first line. The laws of economics put no limits on the number of possible jobs that can exist. Just as software wasn't a job 100 years ago, so too will new jobs appear in the future that robots can't do.
Mar 24, 2014
I just don't see how the canal plan, even if rendered plausible, addresses unemployment.

Ditch-digging, boat-building and the rest will be contracted out to the same corporations who already keep most of the benefits of automation from reaching employee or consumer. They'll import foreign workers and/or pay so little that taxpayers will have to subsidize their below-cost labor with government food programs -- exactly as we do now.

What's needed is a return to actual capitalism, where private costs aren't palmed off on the public and there are penalties for activities that reduce competition and/or endanger the capitalist economy itself.
Mar 24, 2014
1. What is your offer for Canada water? I can guarantee you, if this problem become real, and not just as a society that need to change, Canada will be affected and will not just let you use our water.

2. Why would this kind of unemployment be a problem?
Let me explain, the dream of communism, and by the way I'm not remotely on the left and I'm merely talking about the deep philosophical belief behind communism, not the cheap brand offer by Cuba, North Korea of the ex Soviet Union; is if everyone pitch in and own collectively the mean of production, everyone can live a pretty nice life. The labor we are talking about is the basic employment of our system: farming, mining and processing raw material in something useful.

If those jobs, which are the easiest and most straight forward were taken over by robots, as long as those robots are not sentient, we essentially just said: we can support X number of people. We can feed them, lodge them, give them electricity and water and all the basics facilities of the modern world. Robots are actually taking over sector after sector in the manufacturing world and I'm still baffled that we did not destroy India/China economy by simply producing robots for textiles and basics electronics.

At this point, our base are covered and the only limiting factor is that we need to put an upper limit to the population. Yes, that mean limiting reproduction... If we determine that we can support on above average standard of living 10 billions of peoples, we should make sure to not exceed that population.

Taking this in account, why should you care if everyone is working, or at least an important part of the workforce. People with talent, motivation and the drive to work and get an extra in the society, like travelling around the world could do so. Other who do not really see a point in living a grand life could just bored themselves to death.

3. Stop babbling about hypothetical plan, relying on other countries resources and just make those damn robots happen. The worse that could happen, someone will create a WMD with them: just do not forget that we already have nuke for this.
+6 Rank Up Rank Down
Mar 24, 2014
If only it would work. A better idea would be dirigibles. Way too many unintended consequences and environmental BS to ever make a national canal system feasible.
Get the new Dilbert app!
Old Dilbert Blog