Technological unemployment is (almost) here

[pawel7777]

It's been almost here for centuries now. This idea is called the Luddite fallacy.
...

What if "Luddite fallacy" is a fallacy on its own? LF is basically based only on belief that if nothing bad happened 100-200 years ago, it won't happen again in the future.

As someone already wrote somewhere in this thread, there only 3 sectors: Farming (extraction of raw materials), Manufacturing and Services. Over the generations the workforce shifted from Farming to Manufacturing and then to Services, but there's nowhere to go from there.

Looking at possibilities of the technology we already know, it doesn't take much imagination to notice that vast majority (eventually all) of work could be fully automated when such technology is perfected.

I think the problem is people only think of automation as automation of production (robots working on factory line etc), while services seems to become automated at much faster rate lately.

So fallacy or not, it wouldn't hurt to have a scenario (plan B) on what to do when LF theory is proven wrong.

Useful links:
http://www.economicshelp.org/blog/6717/economics/the-luddite-fallacy/ (some interesting comments there)
http://en.wikipedia.org/wiki/Three-sector_theory

[1715374688]

1715374688

[1715374688]

1715374688

[1715374688]

1715374688

[1715374688]

1715374688

[Nicolas Dorier]

As someone already wrote somewhere in this thread, there only 3 sectors: Farming (extraction of raw materials), Manufacturing and Services. Over the generations the workforce shifted from Farming to Manufacturing and then to Services, but there's nowhere to go from there.

You are confusing a model with the reality. The categorization of labor into 3 sectors is nothing but a model helping to understand our world. There is nothing holy into that.
If this model fails at explaining reality, then it will evolve into further categories.

Here is my response to luddite fallacy.

Automation will profit for the first movers only. After all competition adapted into it, they will start fighting over price.
This fight to the bottom will bring the price near of the cost of production, which would be, the price of maintaining the AI. (which can't be 0, if it can be zero, then all the population will have their own robot to do all they want for them, so no job will be necessary anymore and I don't see how it can be a problem)

Then, imagine that people are out of jobs because of automation, so they can't buy anything.
Well, it means that AI will not be able to be maintained, and factories staying with AI will just have to go bankrupt, liquidating their AI.

However, people always have needs to be satisfied and are willing to work for it. Which means that labor force will be willing to work into factories, while, at the same time, AI is getting destroyed by lack of maintenance.

The equilibrium depends on the price of getting AI maintained versus the price of physical labor.
In other words, the price of intellectual labor versus the price of physical labor.

In a world where intellectual labor is too costly, then physical labor becomes economically more affordable.

AI will always have a cost, if it does not, then we don't care about having a job because everybody will be taken cared by robots.
If the population can't bear the cost, then the AI will destroyed and physical labor restored. OR people will flow from physical labor to intellectual labor, which will drive down the cost of AI.

[AtheistAKASaneBrain]

Steve Wozniak admitted the problem!

http://uk.businessinsider.com/steve-wozniak-artificial-intelligence-interview-humans-family-pets-2015-3

Similar to what Bill Gates said, during the same interview when he talked about bitcoin iirc.

But the problem of AI taking over is not necessarily directly related to technological unemployment. Maybe AI will employ every human to do some kind of dirty work for 18 hours a day in exchange for a bowl of rice and glass of water...

I don't like Bill Gates views on Bitcoin, he wants to deanonimize it, make it centralized and controlled, and basically do the opposite of what satoshi (and most of us) want.

[Razick]

It's been almost here for centuries now. This idea is called the Luddite fallacy.
...

What if "Luddite fallacy" is a fallacy on its own? LF is basically based only on belief that if nothing bad happened 100-200 years ago, it won't happen again in the future.

As someone already wrote somewhere in this thread, there only 3 sectors: Farming (extraction of raw materials), Manufacturing and Services. Over the generations the workforce shifted from Farming to Manufacturing and then to Services, but there's nowhere to go from there.

Looking at possibilities of the technology we already know, it doesn't take much imagination to notice that vast majority (eventually all) of work could be fully automated when such technology is perfected.

I think the problem is people only think of automation as automation of production (robots working on factory line etc), while services seems to become automated at much faster rate lately.

So fallacy or not, it wouldn't hurt to have a scenario (plan B) on what to do when LF theory is proven wrong.

Useful links:
http://www.economicshelp.org/blog/6717/economics/the-luddite-fallacy/ (some interesting comments there)
http://en.wikipedia.org/wiki/Three-sector_theory

Fair enough, *if*  LF is proven wrong, we should have some idea of what to do. However, the Op's ideas both come from a far-far-left perspective that would require that almost all individual liberties be given up. I can't support that. If large scale unemployment were to become a huge problem that cannot be addressed, I suspect we could find better solutions, although I admit I can't think of many right now.

[johnyj]

I wish I was born 100 years from now. I would be enjoying a welfare paycheck, I Don't even care if it's not a lot. I would be able to focus on leisure time, work on art and other things that I love that don't give me any money, while the machines work for us and the people that are really needed do their job and get paid a lot.

That's an ideal situation, however, how much resource would allocate to you is the question. Based on today's model, government must heavily tax everyone who have some income to achieve a good welfare condition. Scandinavian countries for example have a tax rate of almost 50%. But with so high tax, companies will be reluctant to stay in the country and move the operation to other countries. Usually enterprises get the most income, they have many legal ways to avoid the tax. So the welfare condition purely depends on the income of the employees, which is constantly being outsourced to low cost countries

[pawel7777]

...
Then, imagine that people are out of jobs because of automation, so they can't buy anything.
Well, it means that AI will not be able to be maintained, and factories staying with AI will just have to go bankrupt, liquidating their AI.

However, people always have needs to be satisfied and are willing to work for it. Which means that labor force will be willing to work into factories, while, at the same time, AI is getting destroyed by lack of maintenance.

The equilibrium depends on the price of getting AI maintained versus the price of physical labor.
In other words, the price of intellectual labor versus the price of physical labor.
...

OK, for the sake of discussion, let's imagine huge factory able to provide food for entire population.

Yes, there's always some operational cost. Yes, if no one buys from them, they'll go bankrupt. But unemployment won't reach 100% overnight. So while unemployment (number of people with zero income) is slowly increasing and demand is falling, the factory could just scale down its operations accordingly, and in theory could exist even with 90% unemployment.

If decline in demand will increase maintenance cost, then those costs will be passed onto customers by raising prices of food.

Now, being a human with no income doesn't mean you're ready to provide your work for a few bucks per month. You won't work unless you know the earnings will be enough to survive on. Therefore the higher costs of food (and all the other goods) proportionally increase the minimum wage you need. That effectively makes you uncompetitive again.

Even if somehow AI died out and got replaced back by human labour, the same cycle will be happening all over again:
progressing automation -> full automation -> collapse of AI -> humans workers are back -> progressing automation...

But since the technology is already known and available, the cycle would be much faster and the shift could happen multiple times during 1 generation, making such reality impossible to live in.

Anyway, the equilibrium at 100% unemployed is purely academic discussion. The 'shit will go down' probably between 30-50%. Desperate people will support any new leader that will promise to nationalise every company and put food on their plates (Hitler v2.0).

[Nicolas Dorier]

But since the technology is already known and available, the cycle would be much faster and the shift could happen multiple times during 1 generation, making such reality impossible to live in.

Good point, if it is quick the current generation won't be able to adapt fast enough. So yes, if it is too fast, there will be some instabilities, but it will be temporary, with technology people are adapting faster and faster. (at the extreme, my job as a developer put me in an environment where I will not use 95% of what I know today in 5 years... and well, I have to adapt)

If decline in demand will increase maintenance cost

Why ? if you AI is making 300 apples, why would the AI cost more to maintain if demands is 100 rather than 200 ?
In a competitive environment, the price of a good tends to be equal to the cost of operation. A race to the bottom.
But for the sake of argument, let's say that producing surplus cost money (for example, in oil, increase of cost of storage)

then those costs will be passed onto customers by raising prices of food.

A decline in demand exert a downward pressure on price, not the reverse.
Take the example of Oil, the surplus of oil have a downward pressure on price.
Because if decline of demand increase "maintenance cost" as you said, then it becomes economically profitable to dump your stock, even at a loss.

In any case, if the pressure on price by customer is so that the cost of maintenance is above the price that customer would be willing to buy, then the business of making the factory run on AI will collapse.

How would it be replaced by human ?
Simple : if customers are poor and can't buy an AI produced apple, then they will simply start growing their own apple. (which by definition will have a lower cost, since they will always afford to grow their own apple trees)
This will bring back humans at work.

The equilibrium of AI factories versus Human factories only depends on the cost of AI versus cost of human labor.
However, the lower the cost of AI will become, the higher the chance that everybody will be able to profit from it, and, at an extreme point, working for money would not be required anymore and the unemployment problem is not relevant. (I don't think it will ever happen)

Anyway, the equilibrium at 100% unemployed is purely academic discussion

It is, but jobs is not what provide wealth, products and services are (robot or human produced), by keeping that in mind, 100% unemployement might not be a big deal.

[johnyj]

If lots of people were pushed to unemployment, and government don't provide them any kind of income, then they will form a small society and have their own economy, helping each other to produce food, cloth, housing etc...

Once they don't have any fiat money income, they will successfully get rid of the slavery of the existing fiat money system, they would start to use some community currency, bitcoin most likely

It just like kicking out Greece out of EU will make them stronger, I guess central banks would never allow that to happen

[giantdragon]

AI will always have a cost, if it does not, then we don't care about having a job because everybody will be taken cared by robots.

These robots still will need resource to work (land, energy, raw materials) which mostly are owned by "1%" now!

Fair enough, *if*  LF is proven wrong, we should have some idea of what to do. However, the Op's ideas both come from a far-far-left perspective that would require that almost all individual liberties be given up. I can't support that. If large scale unemployment were to become a huge problem that cannot be addressed, I suspect we could find better solutions, although I admit I can't think of many right now.

You, conservative fanatics, must understand that it is absolute game over for you when Luddite Fallacy will break down! Other options are simply impossible!

But with so high tax, companies will be reluctant to stay in the country and move the operation to other countries. Usually enterprises get the most income, they have many legal ways to avoid the tax. So the welfare condition purely depends on the income of the employees, which is constantly being outsourced to low cost countries

The government could stop this by imposing prohibitory "exit tax" or nationalization.

So while unemployment (number of people with zero income) is slowly increasing and demand is falling, the factory could just scale down its operations accordingly, and in theory could exist even with 90% unemployment.

You forgot about security costs when 90% of population are unemployed and have nothing to lose!

[aso118]

If lots of people were pushed to unemployment, and government don't provide them any kind of income, then they will form a small society and have their own economy, helping each other to produce food, cloth, housing etc...

Once they don't have any fiat money income, they will successfully get rid of the slavery of the existing fiat money system, they would start to use some community currency, bitcoin most likely

In effect, this trend of globalization would reverse. Already, countries are putting up barriers to trade. This would extend to local communities as well.

[pawel7777]

then those costs will be passed onto customers by raising prices of food.

A decline in demand exert a downward pressure on price, not the reverse.

In current economical conditions - yes, as it enforces factories to reduce the costs (and often quality) and enforces... automation.
But assuming the automation/optimalisation reached its peak, fall in demand would cause increase the price of single 'unit of food' (ie 1 apple). Simple example:
If you own factory, you need min of $1000/month profit to survive. Currently you're producing 10,000 apples which makes you that $1k profit. If the demand falls by half, you have to increase cost/apple accordingly, otherwise you'll starve to death.
If there's competition who can keep the low prices - you're going down. But if your factory is on 100% of its cost/production efficiency and all the competitors are in the same position - then the market has to accept higher prices.

In any case, if the pressure on price by customer is so that the cost of maintenance is above the price that customer would be willing to buy, then the business of making the factory run on AI will collapse.

How would it be replaced by human ?
Simple : if customers are poor and can't buy an AI produced apple, then they will simply start growing their own apple. (which by definition will have a lower cost, since they will always afford to grow their own apple trees)
This will bring back humans at work.

Are you assuming that every average, long-time-unemployed Joe will always have a piece of land/tools/seeds/skills and 1-2 years of spare time (during which he needs no food/shelter etc) required for his first apple trees to grow?

With majority of population (in western countries) concentrated in big cities and owning nothing (they would sell everything in order to survive) you shouldn't be expecting them to grow/produce anything out of thin air. But you can surely expect them to rob and steal from the ones that still have something.

As history proves, during big financial crisises, when people started to starve, there were no massive 'migrations to the woods' to hunt and fish and create alternative economies/societies, but there were revolutions.

The equilibrium of AI factories versus Human factories only depends on the cost of AI versus cost of human labor.
However, the lower the cost of AI will become, the higher the chance that everybody will be able to profit from it, and, at an extreme point, working for money would not be required anymore and the unemployment problem is not relevant. (I don't think it will ever happen).

Not without changes to the current "capitalism-but-not-quite" system. There will be no benefit for general population without adjusting either ownership or distribution laws.

Therefore if you want everyone to profit from technological progress, you would have to move towards:
- planned economy
- unconditional income
or came up with some other solution.

[Nicolas Dorier]

If the demand falls by half, you have to increase cost/apple accordingly, otherwise you'll starve to death.
If there's competition who can keep the low prices - you're going down. But if your factory is on 100% of its cost/production efficiency and all the competitors are in the same position - then the market has to accept higher prices.

No, the result is that the market will find alternative to AI and/or apples for meeting the demand, making AI apple producer out of business.

Are you assuming that every average, long-time-unemployed Joe will always have a piece of land/tools/seeds/skills and 1-2 years of spare time (during which he needs no food/shelter etc) required for his first apple trees to grow?

I am advocating that long time unemployed Joe will start selling his labor for producing apples because it became profitable. It might be on someone's else floor who would be happy to rent it for exploiting a new usage on his land. (that previously had none)

As history proves, during big financial crisises, when people started to starve, there were no massive 'migrations to the woods' to hunt and fish and create alternative economies/societies, but there were revolutions.

It is interesting to look at history to see how those starvation happened. And you will see that it happened EVERY SINGLE TIME when a state used force to prevent the market for doing its job either via central planning or war. (great famine, vietnam famine, north korea famine, russian famine, soviet famine, chinese famine)
In some case, the initial cause might be a disease, but you will see that state used force to prevent the market for fixing the problem (embargo, war, banning ownership)

Not without changes to the current "capitalism-but-not-quite" system. There will be no benefit for general population without adjusting either ownership or distribution laws.

Therefore if you want everyone to profit from technological progress, you would have to move towards:
- planned economy
- unconditional income
or came up with some other solution.

I agree that the current "capitalism but not quite system" is widening the gap of wealth by force.
Planned economy would lead to famine, as always, and inefficiencies. Thinking one know more than the market, and is superior to the human race (devoid of corruption), is a dangerous fairy tell.

Unconditional income might be more reasonable. But this will not solve the problem.
Unconditional income is a periodic wealth distribution taken from the savers to the debtors.
From the saver's perspective, it only means that he will either : adjust the interest rate for compensating the loss, OR investing in other medium not subject to this problem.
And such alternative medium might well become money.
People won't let themselves get robbed, especially when it is well known in advanced.

But I agree that unconditional income is better than the current capitalist-but-not-quite system which make people with political connections and insider informations richer on the back of both : those who produces and those who does not.

[giantdragon]

Just found interesting report from the UK Ministry of Defence which describes trends up to 2045 year. Technological unemployment issue also mentioned in it.

Robots and automated systems have the potential to be near-ubiquitous in 2045, capable of carrying out extremely complex tasks with no human involvement. Potentially, machines could: carry out all manufacturing and agricultural tasks; repair themselves; as well as clean, cook and tidy. Advances in artificial intelligences could make machines so life-like that they are able to answer phones in call-centres, take care of children and even teach. Computers could potentially diagnose and treat almost every medical condition. If progress on this scale is made in the future, it is possible that there would be very few jobs still carried out by human beings, who could, in practice, be almost totally excluded from the workforce.

In richer countries, this large-scale automation of work would be likely to have a mostly positive effect, as governments would probably be able to provide their citizens with all the material comforts they need.  However, many people may initially struggle to achieve a sense of purpose and social status without work, with possible rises in cases of depression. Education systems may need to be totally redesigned to enable people to self-motivate and to gain satisfaction from activities other than work. Over time, when populations had become more used to a life without work, they may fill their time playing sport, painting, reading and composing music.

In poorer countries, however, large-scale automation of work could stall economic development, perhaps even reversing it. For example, foreign companies would be far less likely to employ cheap labour, as machines would be so much more cost-effective and efficient.  Some employment could initially remain, as poorer countries would be less likely to afford to use machines to carry out roles in their internal labour markets. Eventually, though, automated equipment would probably become cheap enough for even these countries to afford. Poorer governments may not be able to provide more than an extremely basic level of subsistence, meaning that people would have no way to improve their living conditions. This could lead them to become deeply frustrated and angry. Nevertheless, citizens would still receive some benefits from large-scale automation, such as charity-funded machines capable of providing very cheap diagnosis and treating disease. Mass protests and civil unrest could still develop, with ‘anti-robot’ movements becoming increasingly powerful.

P.S. With the phrase "governments would probably be able to provide their citizens with all the material comforts they need" report's authors see large-scale redistribution, which seems inevitable even for right-leaning/conservative people (which I think these military analysts are).

[marcus_of_augustus]

they assume that notion of "government" is still relevant in an era of widespread decentralised networks disintermediating every necessary function that governments currently perform badly. The concepts of 'employment' and 'workforce' will also be redefined in the presence of widespread shared information, arbitration and capital allocation resources.

Basically it is nonsense to even contemplate what influences technology might have on society when you have no way of knowing what the full suite of those technologies will be.

Like the AT&T ad from the early nineties that has the guy sending a fax from the beach ... envisaged global connectedness but missed fax obsolesence to email/messaging?

https://www.youtube.com/watch?v=TZb0avfQme8

[Gyfts]

This post has a lot of truth to it, unfortunately I want to say. A notable example is Amazon. They are starting to automate their delivery and packing services more and more over the past years. Slowly and slowly do workers labour become more worthless after being replaced by high powered machines that pay in the long term. Not necessarily blue collar jobs as mentioned in the OP, but those will soon follow as computers can outwork and "out think" an intelligent worker in multiple aspects. Next decade we'll really see how technology can take the jobs of people.

[romjpn]

Basic income does not mean "very high tax" (well it depends on the political view of the people you talk with ).
On the english page of wikipedia about the basic income, there is not much information about financing it. The french page is more interesting about that.

- Cancellation of the "RSA" in France (free money program if you search for a job, around 400 euros I think). Anybody without a job can apply (with other conditions of course).
- Other social welfare program such as "housing help", or giving money for each baby born.
- Monetary creation, meaning not giving money to banks but directly to people (See QE for people).
- Then we come to taxes (eeeww are saying libertarians ^^) with property taxes and capital revenues and if possible, low "active" income taxes.
- Money spent on checking if people search for a job etc. Coming down to ZERO.
- Revenues from "nature" (like in Alaska with Oil).

I'm pretty confident that it will be at the center of the political debate in the forthcoming years.

[leen93]

solution: cut taxes and gov spending
for people to lazy to work thinking they can just steal it or do other criminal things, there's a death penalty for that 

[giantdragon]

solution: cut taxes and gov spending
for people to lazy to work thinking they can just steal it or do other criminal things, there's a death penalty for that  

Most likely "death penalty" will be for you, conservative fanatics, conducted by unemployed crowd! Good news are that these people won't waste $100K+ of taxpayer funds on lethal injection but choose cheaper method instead!  

[DanielT]

Much before robots become that useful, most people will already have bought one or two of them (it is worth surviving a massive economic transformation), and they will probably make robots works for livelihood. In the worst of hypothesis, people will associate to others who have robots.

It seems that, economically, there is a challenge: huge increase in robot-capital prices that would inevitably come in a robot-dominated economy. Technological advancements can't stop this. All the massive economic management we do for many, many stuff now, would have an analogue in the future. That is a non-trivial economic problem. As the current wealth of the world is inflated by imaginary assets (=fiat), there is not enough capital for the massive transformation (that is why the 1% can't just tell fuck off to the 99%, kill 95%, and become the kings of robots), so spontaneous processes would favor human beings undertaking robot-economy projects.

No need for government.

[pawel7777]

Basic income does not mean "very high tax" (well it depends on the political view of the people you talk with ).
On the english page of wikipedia about the basic income, there is not much information about financing it. The french page is more interesting about that.

- Cancellation of the "RSA" in France (free money program if you search for a job, around 400 euros I think). Anybody without a job can apply (with other conditions of course).
- Other social welfare program such as "housing help", or giving money for each baby born.
- Monetary creation, meaning not giving money to banks but directly to people (See QE for people).
- Then we come to taxes (eeeww are saying libertarians ^^) with property taxes and capital revenues and if possible, low "active" income taxes.
- Money spent on checking if people search for a job etc. Coming down to ZERO.
- Revenues from "nature" (like in Alaska with Oil).

I'm pretty confident that it will be at the center of the political debate in the forthcoming years.

Well, if you think about it, the countries which already have benefits for unemployed (without time limits), benefits for the disabled, child benefits, state pensions are not that far from unconditional income. Pretty much every adult, whether employed or not gets his money one way or another.

Much before robots become that useful, most people will already have bought one or two of them (it is worth surviving a massive economic transformation), and they will probably make robots works for livelihood. In the worst of hypothesis, people will associate to others who have robots.

Doubt that. Your robots would likely end up sitting on the couch with you and watching day-time tv. If I have a factory, why would I employ your robot if I could just buy/lease my own? Your robot will always have higher cost as you'd need to include not only its maintenance/depreciation but also costs of your living.