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?
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?
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?
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]
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?
http://www.overcomingbias.com/2012/04/em-econ-101-talk.html in case you haven't seen it.
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.
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.
[1]: patio11 had good things about to say about the publisher he worked with, perhaps they could also work with you too.
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.