200 years is only a blip in human history. First the machines automated physical labor, then they started automating mental labor, and now they're even starting to automate creative labor. As a friend of mine is fond of saying, society has yet to deal with all the ramifications brought by the steam engine. And we're moving way beyond that.
The big difference now is that the progress is fast enough for people to start to understand that their job may be automated in their life. Smart truck drivers in rich countries know that they have only at most 10 years left. Children from villages need to move to big cities to survive, as they understand the trend that farming jobs will worth less over time. Also more people that I know are getting interested in programming, though they do that for the money, not because they would understand the power of creation with code.
> The big difference now is that the progress is fast enough for people to start to understand that their job may be automated in their life.
The other big difference is that US hourly wages peaked in the 1970s, and there are enough people experiencing the divergence between expectations of continued improvement and reality since then that its no longer a prediction of harms that might occur in the future, but an explanation of harms that are being experienced in the present.
I don't believe that's what the parent comment was saying.
Also, in a way, I believe that supports his argument. Why is it that the world's population has tripled since 1950? It's not just because people are having more sex. It's because of the changes in technology in the past 200 years. ie innovations in agricultural, healthcare, and energy production. [0]
While I won't argue that technology remained stagnant before that, I will argue that technological growth was very very slow. I would argue that many places in the world looked very similar at 400AD and 1400AD (in a technological sense). But in the past ~200 years, we've broken through whatever cultural barrier was holding us back technologically, and as such, the world is a vastly different place than it was 200 years ago.
I disagree at least in part with the article, jobs have always been replaced by machines (as well as globalization, which I don't have time to get into). But just as the technological pace of the world has quickened, so has the pace of job loss.
What is different is that at the "middle" of industrialization, it would be normal for someone to lose the job, but possibly find a new job, especially in the service industries. (And thus we have post-industrialization). However, what I would argue is happening now is that it becomes more and more difficult (especially with ever more specialization in jobs)to lose one job, and then quickly replace it with another job. That's one of the difference between then and now. As the parent posted, once the service industry jobs have been automated away, and then "creative" jobs, what then? "Post-Creative Jobs?" (Seriously, if anyone can answer this for me, please let me know.)
The only jobs that probably won't be automated away is programming. Even if you build layers and layers of abstraction, you still have to code on top of the highest layer...and debug the machines when they go wrong (and they will go wrong; machines aren't perfect).
I can imagine a future where the lucky few who still have jobs don't actually do any work. Instead, they speak "New COBOL", a debased imitation of natural language, to machine learning algorithms, convincing them to do the stuff that previous generations would have to do 'manually'. "Please make the shark fiercer", these job-holding people would say to the algorithm, before feeding it huge troves of data that would teach the machine basic concepts such as "shark" and "fierce".
Obviously, some media commenters would claim that programming has been rendered now obsolete by the rise of New COBOL, but in reality, these lucky few speaking 'New COBOL' are the new programmers of that era. In a world where algorithms eat the world, someone still has to babysit the algorithms.
200 years is over an order of magnitude longer than the time people typically predict we have left before robots "take all our jobs", which is usually like 5 or 10 years [1]. It is also several times longer than it takes each time for the economy get better, unemployment to fall, and this fear to subside, as demonstrated by the article.
This suggests that the current rise in fear will likely reverse long before it becomes any more of a problem than it was the previous many times. That's the point of the article.
What's your point? Why is it relevant to compare against whatever arbitrary definition of "human history" you chose? (Since the "civilization"? Since the Agricultural Revolution? Homo sapiens? Homo erectus?)
When AI replaces programmers, mass employment is over. There will be nothing that can be done by a human that a machine cannot do better. Fallout-style "synths" will be able to do everything we need our feet and fingers for.
My only real concern is that the more autonomously powerful and self-aware these new intelligences become, at what point will they perceive humanity to be a net liability? What happens when they don't need us anymore? What if by then (esp with climate change and antibiotic resistant bacteria) we can't live without them?
We'll be lucky if we get the Neuromancer ending, where machines travel the universe while humanity happily farts around on earth. Lucky.
It seems like a category error to worry too much about machines deciding that humans are an "inefficiency" or a "liability." The drive towards maximizing efficiency is alien to computer systems that have not been explicitly constructed with that in mind. Organic creatures have an evolutionary impetus to maximize efficiency; machines (at this point and in the foreseeable future, at least) do not.
On the plus side, a neo-feudalistic dystopian future scenario where the capitalists own and control everything and 99.999% of humans are peasants eking by, should only last until the machines take over all the wealth. The old shirt-sleeves saying suggests this will take about three generations.
Then we will all be equals.
(I'm hoping for a pseudo-post scarcity society brought by the singularity).
I can't find the video, but a very powerful example of the folly in this way of thinking is in the mind of a horse. When the car came out, horses could have looked around and said "I know they're taking our jobs today, but for the past 10k years, every time one of our jobs got replaced, we found another one." Yet, despite such a lengthy history of adaptive value, after cars became prevalent, horses were never useful an laborers ever again. It's not far fetched to think that humans are just a better horse, able to stave off economic obsolescence slightly longer than they did.
This is a clever collection. I'm sure at least one article, on the eve of the moon landing, asked "We've been wanting to go to the moon forever. What's different this time?"
Concerning robots, I can answer that question. This time, we have cheap sensors and connectivity, powerful processors, and machine learning.
It's early, but it's a sea change. Since the beginning, robots have been instructed what to do. Now, increasingly, they can decide what to do. They can react to the environment. Recover from failure. Handle variance.
Most of the focus has been on autonomous driving, but the same sensors and algorithms are applicable to so many varied tasks that have traditional been unapproachable by machines. (The moment this technology enables a laundry folding robot, I'm buying it, not matter what it costs.)
And it's not just manual labor but knowledge work as well. We laugh at strange errors from Siri and Google Now, but I have also started to notice that more and more of my calls don't involve a phone tree and don't have to get escalated to a human.
Sure, but the article is about fear. In those "more than 200 years" there's been a worldwide explosion in prosperity, and each time a rise in unemployment was blamed on robots, it later fell again without any retreat by the robots. This suggests that your claim that "Robots have been taking all the jobs" is nothing to fear.
Thinking about robots is thinking about cheaper, smaller and more energy efficient devices. For example think about smart phones and communications, think about information storage and transformation. Those aspects are not a single extrapolation of the past, they are a revolution. I see many advances in nanotechnology, machine learning and miniaturization that could give rise to a new revolution, this is like the transistor, we are at the verge of a big change.
As for the 1930s and onwards, I once described the Japanese war in China as indefensible even from a narrowly economic perspective, because developing Manchuria was enough to absorb Japanese industrial productivity for the next thirty or forty years. That phrase is historians' standard way of describing what the Japanese wanted to do, but it was an absolute revelation to a non-historian friend of mine, who realized from that phrase that industrial productivity has been a waste product for the past eighty years or so...
Also, can we stop calling it "robots." I know it's silly, but I always picture a sci-fi style humanoid robot when they tell me they're taking my job. But we're not talking about that, we're simply talking about machines and computers with vastly more automation than previously.
We already have almost "self driving cars". They are called busses, where 30 people are almost automatically driven by one so it's already 96% automated. We're automating that one last job. If you think about trains, ships and even farms, they are already automated to a higher degree than buses.
And if people are afraid they will be jobless in an economy dominated by robots, there is a way out - becoming more independent. There are a number of ways a person, a community or a city could become more independent from imports and big corporations.
1. agriculture - if the person or group has land, they don't depend much on other external agents
2. solar and wind energy - they can be built on premise and don't depend on a central energy production system which could be hard to pay for, in the long run, if you don't have much cash
3. 3D printing - making us independent from factories which are often on another continent and owned by big corp
4. open source - anything we put there becomes our common legacy; fortunately AI research publishes almost everything, and we hope to keep it like that
5. capital - who has capital can invest and have a steady stream of income
So, if big corp doesn't want to hire us and we don't have money to buy their products, we still have our hands, our brains and help from the community to find solutions. We can be independent even from jobs provided by big corps.
By the way - a robot and a 3D printer could build more robots and 3D printers. Think about that. When they can automate the process to a high degree, then any community can self-bootstrap into economic viability without depending on continuous delivery of tech and products from large corporations. That would be great for independence.
30 comments
[ 2.8 ms ] story [ 63.0 ms ] thread(sorry, couldn't resist ;-)
The other big difference is that US hourly wages peaked in the 1970s, and there are enough people experiencing the divergence between expectations of continued improvement and reality since then that its no longer a prediction of harms that might occur in the future, but an explanation of harms that are being experienced in the present.
Also, in a way, I believe that supports his argument. Why is it that the world's population has tripled since 1950? It's not just because people are having more sex. It's because of the changes in technology in the past 200 years. ie innovations in agricultural, healthcare, and energy production. [0]
While I won't argue that technology remained stagnant before that, I will argue that technological growth was very very slow. I would argue that many places in the world looked very similar at 400AD and 1400AD (in a technological sense). But in the past ~200 years, we've broken through whatever cultural barrier was holding us back technologically, and as such, the world is a vastly different place than it was 200 years ago.
I disagree at least in part with the article, jobs have always been replaced by machines (as well as globalization, which I don't have time to get into). But just as the technological pace of the world has quickened, so has the pace of job loss.
What is different is that at the "middle" of industrialization, it would be normal for someone to lose the job, but possibly find a new job, especially in the service industries. (And thus we have post-industrialization). However, what I would argue is happening now is that it becomes more and more difficult (especially with ever more specialization in jobs)to lose one job, and then quickly replace it with another job. That's one of the difference between then and now. As the parent posted, once the service industry jobs have been automated away, and then "creative" jobs, what then? "Post-Creative Jobs?" (Seriously, if anyone can answer this for me, please let me know.)
[0] https://en.wikipedia.org/wiki/Demographic_transition (note: a number of people argue that there is also a stage 6)
Hi, my name is sgnelson, I'm a post-creative. I need a job.
I can imagine a future where the lucky few who still have jobs don't actually do any work. Instead, they speak "New COBOL", a debased imitation of natural language, to machine learning algorithms, convincing them to do the stuff that previous generations would have to do 'manually'. "Please make the shark fiercer", these job-holding people would say to the algorithm, before feeding it huge troves of data that would teach the machine basic concepts such as "shark" and "fierce".
Obviously, some media commenters would claim that programming has been rendered now obsolete by the rise of New COBOL, but in reality, these lucky few speaking 'New COBOL' are the new programmers of that era. In a world where algorithms eat the world, someone still has to babysit the algorithms.
This suggests that the current rise in fear will likely reverse long before it becomes any more of a problem than it was the previous many times. That's the point of the article.
What's your point? Why is it relevant to compare against whatever arbitrary definition of "human history" you chose? (Since the "civilization"? Since the Agricultural Revolution? Homo sapiens? Homo erectus?)
[1]: e.g. http://www.nbcnews.com/tech/tech-news/robot-may-take-your-jo...
My only real concern is that the more autonomously powerful and self-aware these new intelligences become, at what point will they perceive humanity to be a net liability? What happens when they don't need us anymore? What if by then (esp with climate change and antibiotic resistant bacteria) we can't live without them?
We'll be lucky if we get the Neuromancer ending, where machines travel the universe while humanity happily farts around on earth. Lucky.
Then we will all be equals.
(I'm hoping for a pseudo-post scarcity society brought by the singularity).
Concerning robots, I can answer that question. This time, we have cheap sensors and connectivity, powerful processors, and machine learning.
It's early, but it's a sea change. Since the beginning, robots have been instructed what to do. Now, increasingly, they can decide what to do. They can react to the environment. Recover from failure. Handle variance.
Most of the focus has been on autonomous driving, but the same sensors and algorithms are applicable to so many varied tasks that have traditional been unapproachable by machines. (The moment this technology enables a laundry folding robot, I'm buying it, not matter what it costs.)
And it's not just manual labor but knowledge work as well. We laugh at strange errors from Siri and Google Now, but I have also started to notice that more and more of my calls don't involve a phone tree and don't have to get escalated to a human.
You mean this? https://www.youtube.com/watch?v=QDljo6UUWps
They just haven't finished yet.
The question is, for how long will that realistically continue?
As for the 1930s and onwards, I once described the Japanese war in China as indefensible even from a narrowly economic perspective, because developing Manchuria was enough to absorb Japanese industrial productivity for the next thirty or forty years. That phrase is historians' standard way of describing what the Japanese wanted to do, but it was an absolute revelation to a non-historian friend of mine, who realized from that phrase that industrial productivity has been a waste product for the past eighty years or so...
And when ready I can travel to my own planet, terraform it and live out my artificially enhanced lifespan there.
(A dumb criticism, I know.)
And if people are afraid they will be jobless in an economy dominated by robots, there is a way out - becoming more independent. There are a number of ways a person, a community or a city could become more independent from imports and big corporations.
1. agriculture - if the person or group has land, they don't depend much on other external agents
2. solar and wind energy - they can be built on premise and don't depend on a central energy production system which could be hard to pay for, in the long run, if you don't have much cash
3. 3D printing - making us independent from factories which are often on another continent and owned by big corp
4. open source - anything we put there becomes our common legacy; fortunately AI research publishes almost everything, and we hope to keep it like that
5. capital - who has capital can invest and have a steady stream of income
So, if big corp doesn't want to hire us and we don't have money to buy their products, we still have our hands, our brains and help from the community to find solutions. We can be independent even from jobs provided by big corps.
By the way - a robot and a 3D printer could build more robots and 3D printers. Think about that. When they can automate the process to a high degree, then any community can self-bootstrap into economic viability without depending on continuous delivery of tech and products from large corporations. That would be great for independence.