They have. At one time, there were millions of horses employed in the transportation sector. As technology advanced, they were unable to retrain in other fields because no matter how bright, a horse is fundamentally not as useful as a car. This hasn't happened for humans yet. That doesn't mean it never will.
Technology is improving. People are not. The fact that an increasing line has not yet crossed a flat line is not evidence it never will.
What happens when we develop technology which can do 99% of what a human can do better, faster, and cheaper?
With previous innovations, there was always something else that a displaced human worker could find to do. What happens when 99% of human jobs are gone?
No it doesn't. Capital becomes obsolescent, labor does not.
>What happens when we develop technology which can do 99% of what a human can do better, faster, and cheaper?
During 99% of recorded history, agricultural labor was the prevailing definition of "what a human can do".
I'm not talking about the prevailing definition of what a human can do. I'm talking about the actual fact of what a human can do.
You still haven't addressed the question of what will happen when nearly everything which a human being is capable of can be done better by a machine.
This holds even for the scenario of AI becoming a very close substitute for humans. The economic pressure, however, is to develop AI in directions that are complementary to human abilities.
If and when the robot dream is finally realized, it would be novel insofar as it involves the unprecedented production of laborers through an economic process. The debate over the effects of bringing new workers into the economy has, however, been played out many times. (E.g. regarding cheap robots for unskilled work cf. Chinese immigration in the 19th century.)
I'm not sure why you ask me for a reference for a claim I didn't make. I claimed that labor and capital are substitutes as inputs to production. I also claimed that an inferior or more expensive substitute may go unused when a better alternative becomes available.
Do you disagree with any of these claims?
If you're edged out of a job as a human it's getting more and more difficult to re-train. If it were easy to re-train in a lot of cases it would be automated already.
What would that do?
It would drive down the salaries of sales reps by 2%. That's it. That's the magic of the market economy. That money saved by replacing people with software goes back into the economy and creates demand for more goods (which needs to be pushed by sales reps, for example.)
And by the way, I would argue that electrification and the industrial revolution were far more disruptive (and happened damn fast) than online shopping and bookkeeping software.
> Ok let's say they don't re-train. Let's say they just go out and try to find a job as
No, the point is what if: 1) The easy to do and train for jobs are all automated away. 2) The available jobs require PhD level expertise to do.
If I've studied for 8 years to get my quantum physics PhD to get my job as a technician at Quantum World, then it's going to be damn hard to retrain for 8 more years to get a PhD in Natural Language Processing to work phone support at AutonomaTech.
People are becoming increasingly specialized. If automation eliminates the easy to learn jobs, it gets harder to jump from one specialization to another.
Some will be hard to automate, because they are fuzzy and depend on interacting with human beings.
It's harder for a robot to cut your hair and put on make up like you direct it to, than to play chess.
Pick another position that scales with production or demand and you'll invariably find similar automation that has been introduced in the past ten years.
This ever increasing march of technology may result in more goods being created, but due to the nature of our capitalist society, it tends to pool the capital in those who have capital to begin with.
The huge disparity between the rich and the poor in the US is not due to "technology" but it is not being helped by it either.
Technology creates efficiencies, but there is no natural way for those efficiencies to benefit those whom it puts out of work.
Only because we are paying people to not work, and because fair pay is illegal under a certain rate. Consequence is that work worth less than minimum wage cannot, by law, be given to people - and so automation moves in.
Stop paying people to not work and they'll find or make work. (There is a difference between a safety net and making lack of arbitrary wealth illegal.)
And "the poor", for most practical purposes, does not exist in the US. If you are living on more - a lot more - than the world median income, if you are flushing your toilet with drinking water, if you have a TV & car & cell phone & AC & thermostat heating & >1000 sq ft housing & ..., if you have over $1/meal/person/day[1], and etc you're not poor.
If your choice is 850 days of food vs. a TV, and you opt for the TV, you're not poor. (Derived from [1] above.) With over two years of energy literally under your belt, paying off that cheap loan and earning far more, even under adverse conditions, leverages you well outside poverty.
"Poor on paper" is unconvincing, given the opportunity for wise choices.
Noblesse Obese
I really don't expect mass unemployment due to technology to be a problem until we get a lot more jobs disappearing to robotics. When you see a purely-robotic McDonalds, start wondering.
Some of those wars might well have been the result of changing economics.
The reason technology doesn't create systemic unemployment is that people don't just shrivel up and die when they lose their job. If someone is replaced with a piece of software they don't just say "oh well, I'm just gonna sit here and starve." They either find a job in a sector that requires minimal training, or they re-train. They have no choice!
The basic premise is this: there is no limit on the amount of work that needs to be done - there is always room for more human labour (either in servicing other humans or creating goods for them, or whatever.) This is just another way of stating Say's Law (Wikipedia link: http://bit.ly/pvD9Hz) which is that goods are paid for with other goods. If $x is saved in one area, it just gets pumped into another part of the economy, creating demand/jobs there. The net effect of technological progress is: the same number of people producing more goods and services.
You don't even have to believe me - just look around! this is exactly what's been happening for thousands of years. Technology has never resulted in systemic unemployment - only temporary frictional unemployment as people re-train.