5 ms·

With my only credentials being that I once had a single email exchange with Bill Joy, I'd have to disagree and say that the future needs us more than ever. Mac

With my only credentials being that I once had a single email exchange with Bill Joy, I'd have to disagree and say that the future needs us more than ever. Machines are not taking over humans. They are taking over boring tasks, leaving the ever more interesting ones to us. Nanotech and the like are completely devoid of meaning without human input. And where is nanotech these days, anyway?

jiggy2011 14y ago
The big question is , what are these interesting jobs and where are they going to come from? Intellectual output can be much more easily and freely copied than physical labour (e.g software,music etc) so do we really need billions of mediocre scientists/programmers/artists or just a few million good ones who's work everyone else can reap the benefit from?
- michaelochurch 14y ago
  The big question is , what are these interesting jobs and where are they going to come from? My 22nd blog post (of a 23-part series that was supposed be 22, before it was supposed to ~7) is covering that question. The core concept in a lot of my writing recently has been Convexity, which is when the difference between excellent work and mediocre work matters much more than the difference between mediocrity and noncompliance. In engineering we call this the "10X" effect. [snip] Where is there room left for humans? 1. High-end service work. [...] Wealthier people might pay a huge premium for the [high-quality] work they can get from a person as opposed to the [reliable] work done by computers. Then, the payoff function is a convex filter applied to this performance measure, y. Even though it's service work, the "little bit extra" that comes form human involvement makes it convex work on the real market. 2. Creative arts (probably). I doubt that we'll ever see computers that can do this as well as we can. Of course, they'll take on more of the grunt work and eventually be used to supply inspiration, but I think there will be a human edge, in writing novels and composing symphonies, forever. At least, that'll be true at the high end. Of course, there'll be auto-generated trash romance novels in a decade, if there aren't already. 3. Research, science, and programming. These [cheap] reliable machines don't descend from the sky. People have to solve scientific problems, hardware engineers have to build the machines, and software engineers must program them. There's also a place for executives and project managers in that process, as well: bringing coffee so the people doing actual work don't have to break flow to get it. 4. Leadership. People use "leadership" as a euphemism for the rapidly evolving and someday-soon-subordinate set of activities we call "people management", but there will always be a place for people who can actually lead on their own merits. The above (1-3) require not only it, but the most important subcategory (perhaps the only actually important one) of leadership, which is... 5. Education and teaching. This is what I call "hidden convexity" because the wages for these professionals are despairingly concave, but the impact is not merely convex, but fully exponential. One of the most important Buddhist meditations-- and it requires no supernatural beliefs to do it-- is on the kindness of others, as a way of understanding interdependence, and it often comes back to teaching. That is the most important thing we do. Of course, all of this work has something in common. It's convex. Small differences in performance (including noise factors) have major impacts. A few people do those jobs very well and gets lots of rewards, and most get very little. This is why Convexity is such a major economic problem (I contend, the defining problem of the 21st century). The convex world has no hope of sustaining full employment. Would you hire a 17-year-old who just learned Python to write your core machine-learning infrastructure? Hell no, but you'd hire the 27-year-old he turns into after ten years of hard work. Who will pay for that, though? How's he going to eat while he gets there? The convex economy can't support the learn-while-you're-economically-useless-on-your-parents'-dime regime. Learning has to be lifelong, because the world changes so fast, and that also means people will need to be paid while they do it. [/snip]
  - jiggy2011 14y ago
    The education example is interesting because it's the only one of those which tends to be heavily unionised in most western countries at least. So that probably explains why the wages are concave. It's also one that benefits greatly from physical proximity, so there will be a demand for a fairly high number of teachers because they don't "scale" as well. Although MOOCs may change that a bit. Based on what I have heard from friends who are teachers, getting paid more as a teacher is usually done via being promoted to a leadership position (i.e headteacher etc) which has more to do with politics than teaching skill. Or getting seniority with experience (measured in years) which does not necessarily correlate with skill either. This means that teachers who teach well do so because they are intrinsically motivated to teach rather than for the money but of course they do need some money to eat. I wonder if this could be applied to other fields. For example, would you rather be paid a guaranteed fixed lower middle class wage to hack on open source projects all day with a 0% risk of starvation but equally a 0% chance of getting rich but you do get a pay rise after 5 years and another after 10 years. Or would you rather work for peanuts at a startup with a 1% chance of becoming a multi-millionaire? And would there be difference in how motivated you were to do your best work in either scenario?
  - randallsquared 14y ago
    How do Robin Hanson's Em economics figure into your thinking on this? http://www.overcomingbias.com/2012/04/em-econ-101-talk.html http://www.overcomingbias.com/2012/04/em-econ-101-talk.html in case you haven't seen it.
    - michaelochurch 14y ago
      I will look into that. Thanks!
  - yuhong 14y ago
    Reminds of of this: https://news.ycombinator.com/item?id=4349781 https://news.ycombinator.com/item?id=4349781 http://www.reddit.com/r/web_design/comments/ifjld/launching_a_new_portfolio_who_isnt_hows_it_look/c23g2xl http://www.reddit.com/r/web_design/comments/ifjld/launching_...
- tomjen3 14y ago
  Nurse plus: most family doctors could be replaced by a friendly nurse and a big computerized database of symptoms and diseases and a relatively simple program to tell her which tests to perform and when to refer the patient to a specialist. When you locked yourself out of your house you don't need a fully trained locksmith, you just need a guy who can open the lock which is easy with a modern tool. In general there are a ton of jobs available through deskilling, guildbusting and creative thinking.
michaelochurch 14y ago
Bingo! This is what inspired me to develop the Convexity theory ( http://michaelochurch.wordpress.com/2013/04/03/gervais-macleod-21-why-does-work-suck/ http://michaelochurch.wordpress.com/2013/04/03/gervais-macle... ). It's a middle-ground between head-in-the-sand "nothing's changing" and the almost theological "Singularity". These machines are powerful levers, but they still need movers, and there's no evidence that they'll become sentient or develop ulterior agendas. We'll just develop more convincing abstraction layers and delegating a greater share of the unpleasant, repetitive work.
- Hitchhiker 14y ago
  Grady Booch thinks otherwise .. http://www.infoworld.com/t/consumerization-of-it/the-age-of-the-sentient-machine-upon-us-214391 http://www.infoworld.com/t/consumerization-of-it/the-age-of-...
- Houshalter 14y ago
  It's still a serious issue for those who aren't engineers or programmers. There is a net benefit of automation, but realistically in our current economy, unemployment isn't that great. And of course in the long run even those jobs can be replaced once we have intelligent machines. Though that would create a completely different world than this one.
- clicks 14y ago
  Can you just write a book already michaaelochurch? Or rather, just compile your blog writings into one [1]. I've been meaning to read everything you've written in your blog... but I think I think I keep pushing it because I generally don't like reading long things on a computer screen. [1]: patio11 had good things about to say about the publisher he worked with, perhaps they could also work with you too.
gnosis 14y ago
The near future might need us, but will it need nearly as many of us?

jiggy2011 14y ago

The big question is , what are these interesting jobs and where are they going to come from? Intellectual output can be much more easily and freely copied than physical labour (e.g software,music etc) so do we really need billions of mediocre scientists/programmers/artists or just a few million good ones who's work everyone else can reap the benefit from?

michaelochurch 14y ago
The big question is , what are these interesting jobs and where are they going to come from? My 22nd blog post (of a 23-part series that was supposed be 22, before it was supposed to ~7) is covering that question. The core concept in a lot of my writing recently has been Convexity, which is when the difference between excellent work and mediocre work matters much more than the difference between mediocrity and noncompliance. In engineering we call this the "10X" effect. [snip] Where is there room left for humans? 1. High-end service work. [...] Wealthier people might pay a huge premium for the [high-quality] work they can get from a person as opposed to the [reliable] work done by computers. Then, the payoff function is a convex filter applied to this performance measure, y. Even though it's service work, the "little bit extra" that comes form human involvement makes it convex work on the real market. 2. Creative arts (probably). I doubt that we'll ever see computers that can do this as well as we can. Of course, they'll take on more of the grunt work and eventually be used to supply inspiration, but I think there will be a human edge, in writing novels and composing symphonies, forever. At least, that'll be true at the high end. Of course, there'll be auto-generated trash romance novels in a decade, if there aren't already. 3. Research, science, and programming. These [cheap] reliable machines don't descend from the sky. People have to solve scientific problems, hardware engineers have to build the machines, and software engineers must program them. There's also a place for executives and project managers in that process, as well: bringing coffee so the people doing actual work don't have to break flow to get it. 4. Leadership. People use "leadership" as a euphemism for the rapidly evolving and someday-soon-subordinate set of activities we call "people management", but there will always be a place for people who can actually lead on their own merits. The above (1-3) require not only it, but the most important subcategory (perhaps the only actually important one) of leadership, which is... 5. Education and teaching. This is what I call "hidden convexity" because the wages for these professionals are despairingly concave, but the impact is not merely convex, but fully exponential. One of the most important Buddhist meditations-- and it requires no supernatural beliefs to do it-- is on the kindness of others, as a way of understanding interdependence, and it often comes back to teaching. That is the most important thing we do. Of course, all of this work has something in common. It's convex. Small differences in performance (including noise factors) have major impacts. A few people do those jobs very well and gets lots of rewards, and most get very little. This is why Convexity is such a major economic problem (I contend, the defining problem of the 21st century). The convex world has no hope of sustaining full employment. Would you hire a 17-year-old who just learned Python to write your core machine-learning infrastructure? Hell no, but you'd hire the 27-year-old he turns into after ten years of hard work. Who will pay for that, though? How's he going to eat while he gets there? The convex economy can't support the learn-while-you're-economically-useless-on-your-parents'-dime regime. Learning has to be lifelong, because the world changes so fast, and that also means people will need to be paid while they do it. [/snip]
- jiggy2011 14y ago
  The education example is interesting because it's the only one of those which tends to be heavily unionised in most western countries at least. So that probably explains why the wages are concave. It's also one that benefits greatly from physical proximity, so there will be a demand for a fairly high number of teachers because they don't "scale" as well. Although MOOCs may change that a bit. Based on what I have heard from friends who are teachers, getting paid more as a teacher is usually done via being promoted to a leadership position (i.e headteacher etc) which has more to do with politics than teaching skill. Or getting seniority with experience (measured in years) which does not necessarily correlate with skill either. This means that teachers who teach well do so because they are intrinsically motivated to teach rather than for the money but of course they do need some money to eat. I wonder if this could be applied to other fields. For example, would you rather be paid a guaranteed fixed lower middle class wage to hack on open source projects all day with a 0% risk of starvation but equally a 0% chance of getting rich but you do get a pay rise after 5 years and another after 10 years. Or would you rather work for peanuts at a startup with a 1% chance of becoming a multi-millionaire? And would there be difference in how motivated you were to do your best work in either scenario?
- randallsquared 14y ago
  How do Robin Hanson's Em economics figure into your thinking on this? http://www.overcomingbias.com/2012/04/em-econ-101-talk.html http://www.overcomingbias.com/2012/04/em-econ-101-talk.html in case you haven't seen it.
  - michaelochurch 14y ago
    I will look into that. Thanks!
- yuhong 14y ago
  Reminds of of this: https://news.ycombinator.com/item?id=4349781 https://news.ycombinator.com/item?id=4349781 http://www.reddit.com/r/web_design/comments/ifjld/launching_a_new_portfolio_who_isnt_hows_it_look/c23g2xl http://www.reddit.com/r/web_design/comments/ifjld/launching_...
tomjen3 14y ago
Nurse plus: most family doctors could be replaced by a friendly nurse and a big computerized database of symptoms and diseases and a relatively simple program to tell her which tests to perform and when to refer the patient to a specialist. When you locked yourself out of your house you don't need a fully trained locksmith, you just need a guy who can open the lock which is easy with a modern tool. In general there are a ton of jobs available through deskilling, guildbusting and creative thinking.

michaelochurch 14y ago

Bingo! This is what inspired me to develop the Convexity theory ( http://michaelochurch.wordpress.com/2013/04/03/gervais-macleod-21-why-does-work-suck/ http://michaelochurch.wordpress.com/2013/04/03/gervais-macle... ). It's a middle-ground between head-in-the-sand "nothing's changing" and the almost theological "Singularity". These machines are powerful levers, but they still need movers, and there's no evidence that they'll become sentient or develop ulterior agendas. We'll just develop more convincing abstraction layers and delegating a greater share of the unpleasant, repetitive work.

Hitchhiker 14y ago
Grady Booch thinks otherwise .. http://www.infoworld.com/t/consumerization-of-it/the-age-of-the-sentient-machine-upon-us-214391 http://www.infoworld.com/t/consumerization-of-it/the-age-of-...
Houshalter 14y ago
It's still a serious issue for those who aren't engineers or programmers. There is a net benefit of automation, but realistically in our current economy, unemployment isn't that great. And of course in the long run even those jobs can be replaced once we have intelligent machines. Though that would create a completely different world than this one.
clicks 14y ago
Can you just write a book already michaaelochurch? Or rather, just compile your blog writings into one [1]. I've been meaning to read everything you've written in your blog... but I think I think I keep pushing it because I generally don't like reading long things on a computer screen. [1]: patio11 had good things about to say about the publisher he worked with, perhaps they could also work with you too.

gnosis 14y ago

The near future might need us, but will it need nearly as many of us?