5 ms·

It's not so much that there won't be a labor market, but rather than radical changes to the labor market are accelerating. Two centuries ago, well over 90% of

by beat 11y ago

It's not so much that there won't be a labor market, but rather than radical changes to the labor market are accelerating. Two centuries ago, well over 90% of income earners were farmers. Today, it's like 2% and dropping. But unemployment didn't become a more massive problem than it was then. Labor became available for other purposes, enabling the industrial revolution.

ThomPete 11y ago
Yeah but it's accelerating at such a rate that it's important to look at the endgame which IMO is no labour market. I agree that even before we reach that level the issue is thre and so it will serve as a help for those who are "left behind".
- nightski 11y ago
  If that ever happens (which I highly doubt, there is little to no evidence) it is a very long ways off. Hundreds of years at a minimum.
  - ThomPete 11y ago
    There is a lot of evidence that this is how things are moving. What most people get wrong about a job is that they confuse the human doing them with the job they human is doing. Most if not all jobs only require you to be a subset of a human to actually be useful. And so there is a general misconception that computers would need a general intelligence to take over a job when in fact most of the jobs we do require only a fraction of our abilities as humans and normally can be broken down in smaller subsets.
    - notahacker 11y ago
      There was a lot of evidence that this was how things were moving a century ago, but the predictions made at the time look more than a little bit wrong. Not only have the sons of the increasingly obsolete production line employees mostly found jobs, but so have their daughters! Trouble with breaking down jobs into smaller subsets is that you need an awful lot of robots, and suddenly the rental on those doesn't look an awful lot better than the cost of hiring unskilled labour. I mean, the core role of the average fast food service employee in exchanging cash for heated product was rendered obsolete by the heated coin-operated dispensing machine more than a century ago, but it turns out that they're also quite good at upselling to larger items, dealing with unusual requests, getting rid of unruly customers and removing stains from locations you'd need a robot with really impressive articulated limbs and computer vision to adequately deal with. And they don't cost very much per month compared with specialist business process software licenses, never mind hardware. So fast food vendors, with some of the most obsessive attention to process optimization imaginable and some of the least highly-regarded employees, continue to be huge employers
      - ThomPete 11y ago
        No there wasn't a lot of evidence because back then the human intellect wasn't being outcompeted by the machines. This is what is happening now. Machines are gradually moving further and further up the abstraction ladder. So unless you are aware of another human ability used to do a job it kind of stops there.
        nightski 11y ago
        As someone who is trying very hard to keep up with the latest improvements in machine learning and artificial intelligence, I see very little evidence to say we have made strides in reasoning on the level of a human intellect. Machine learning is still unfortunately limited to very specific tasks. Even though the means through which those tasks are accomplished may be very general (neural networks, Bayesian statistics, etc...) the end result is still very limited. Let's not fool ourselves with how far along the technology is and instead embrace it for what it is.
      - ThomPete 11y ago
        Fast food vendors are already now starting to look for ways to automate as they continue to push the prices down on their products. Prices that needs to keep going down in order for them to 1) survive the competition, 2) deliver affordable food from the majority of their customers who are making less and less money.
        notahacker 11y ago
        The Automat has been around since before most of the fast food chains, yet half a century of continuous innovation in process optimization has continued to conclude that paying a bunch of people to mostly stand behind counters exchanging cooked food for cash is worth the money. And unlike the average service industry employee the fast food worker isn't even expected to be a particularly good generalist, is mocked for perceived lack of social skill and isn't operating in a role where new tasks can be invented to supplant tasks which have largely been automated away.
chrisco255 11y ago
We had 25%+ unemployment during the Great Depression, at a time when hardly any women were in the workforce. I would say that was a much more difficult time than this great stagnation we find ourselves in currently. Our economy is now largely service-based, but we are still quite a ways off from automating most services. There's a lot of speculation that the economy will imminently suffer rapid upheaval with few jobs to replace the ones that are automated, but it remains speculative.
- ThomPete 11y ago
  What you and others get wrong in that analysis is that you don't take into account that just one algorithm have to learn something then all the others can do the same. Point machine learning towards a problem and things that normally was cumbersome and took several people to do is suddenly done in the fraction of a second. It's like claiming that if an alien race landed on planet earth with superior skills and superior technology it wouldn't effect the human ability to survive because we endured battles with other species before. We can't use history to teach us much since the opponent is different. This is not speculation anymore. This is unfolding right in front of us and we better take it serious.
  - chrisco255 11y ago
    Machine learning is still far from automating most services. If anything, machine learning will create as many jobs as it will destroy.
    - ThomPete 11y ago
      You must be able to provide some examples of jobs that will be created. So when machine learning will remove the need for radiologist and many surgeons as it's already well underway to. What areas is it that these doctors will then be able to work in thats going to be of value? It's ironic how steadfast many of you are in this idea that more jobs will be created when there is no evidence it's the case and the new jobs that are created require much less than whatever it replaced. You are also confusing globalization with actual job creation in the west.

by beat 11y ago

ThomPete 11y ago

Yeah but it's accelerating at such a rate that it's important to look at the endgame which IMO is no labour market. I agree that even before we reach that level the issue is thre and so it will serve as a help for those who are "left behind".

nightski 11y ago
If that ever happens (which I highly doubt, there is little to no evidence) it is a very long ways off. Hundreds of years at a minimum.
- ThomPete 11y ago
  There is a lot of evidence that this is how things are moving. What most people get wrong about a job is that they confuse the human doing them with the job they human is doing. Most if not all jobs only require you to be a subset of a human to actually be useful. And so there is a general misconception that computers would need a general intelligence to take over a job when in fact most of the jobs we do require only a fraction of our abilities as humans and normally can be broken down in smaller subsets.
  - notahacker 11y ago
    There was a lot of evidence that this was how things were moving a century ago, but the predictions made at the time look more than a little bit wrong. Not only have the sons of the increasingly obsolete production line employees mostly found jobs, but so have their daughters! Trouble with breaking down jobs into smaller subsets is that you need an awful lot of robots, and suddenly the rental on those doesn't look an awful lot better than the cost of hiring unskilled labour. I mean, the core role of the average fast food service employee in exchanging cash for heated product was rendered obsolete by the heated coin-operated dispensing machine more than a century ago, but it turns out that they're also quite good at upselling to larger items, dealing with unusual requests, getting rid of unruly customers and removing stains from locations you'd need a robot with really impressive articulated limbs and computer vision to adequately deal with. And they don't cost very much per month compared with specialist business process software licenses, never mind hardware. So fast food vendors, with some of the most obsessive attention to process optimization imaginable and some of the least highly-regarded employees, continue to be huge employers
    - ThomPete 11y ago
      No there wasn't a lot of evidence because back then the human intellect wasn't being outcompeted by the machines. This is what is happening now. Machines are gradually moving further and further up the abstraction ladder. So unless you are aware of another human ability used to do a job it kind of stops there.
      - nightski 11y ago
        As someone who is trying very hard to keep up with the latest improvements in machine learning and artificial intelligence, I see very little evidence to say we have made strides in reasoning on the level of a human intellect. Machine learning is still unfortunately limited to very specific tasks. Even though the means through which those tasks are accomplished may be very general (neural networks, Bayesian statistics, etc...) the end result is still very limited. Let's not fool ourselves with how far along the technology is and instead embrace it for what it is.
    - ThomPete 11y ago
      Fast food vendors are already now starting to look for ways to automate as they continue to push the prices down on their products. Prices that needs to keep going down in order for them to 1) survive the competition, 2) deliver affordable food from the majority of their customers who are making less and less money.
      - notahacker 11y ago
        The Automat has been around since before most of the fast food chains, yet half a century of continuous innovation in process optimization has continued to conclude that paying a bunch of people to mostly stand behind counters exchanging cooked food for cash is worth the money. And unlike the average service industry employee the fast food worker isn't even expected to be a particularly good generalist, is mocked for perceived lack of social skill and isn't operating in a role where new tasks can be invented to supplant tasks which have largely been automated away.

chrisco255 11y ago

We had 25%+ unemployment during the Great Depression, at a time when hardly any women were in the workforce. I would say that was a much more difficult time than this great stagnation we find ourselves in currently. Our economy is now largely service-based, but we are still quite a ways off from automating most services. There's a lot of speculation that the economy will imminently suffer rapid upheaval with few jobs to replace the ones that are automated, but it remains speculative.

ThomPete 11y ago
What you and others get wrong in that analysis is that you don't take into account that just one algorithm have to learn something then all the others can do the same. Point machine learning towards a problem and things that normally was cumbersome and took several people to do is suddenly done in the fraction of a second. It's like claiming that if an alien race landed on planet earth with superior skills and superior technology it wouldn't effect the human ability to survive because we endured battles with other species before. We can't use history to teach us much since the opponent is different. This is not speculation anymore. This is unfolding right in front of us and we better take it serious.
- chrisco255 11y ago
  Machine learning is still far from automating most services. If anything, machine learning will create as many jobs as it will destroy.
  - ThomPete 11y ago
    You must be able to provide some examples of jobs that will be created. So when machine learning will remove the need for radiologist and many surgeons as it's already well underway to. What areas is it that these doctors will then be able to work in thats going to be of value? It's ironic how steadfast many of you are in this idea that more jobs will be created when there is no evidence it's the case and the new jobs that are created require much less than whatever it replaced. You are also confusing globalization with actual job creation in the west.