The Rise of the Machines
Via Tyler Cowen, a Robin Hanson paper (PDF) on the economic implications of robots sufficiently sophisticated to act as near-substitutes for a wide range of human labor. Roughly speaking, the consequences would be much faster economic growth, but falling wages. The whole scenario is a bit outlandish at this point, but it's at least possible that this is a limiting case of a developing situation so it's worth thinking about a little bit.
Historically, the share of national income going to capital versus the share going to labor has stayed roughly constant. It fluctuates a bit according to what point in the economic cycle we're in, but the long-run trend is flat. Robots, Hanson suggests, would destabilize this relationship and tilt the playing field decisively in favor of capital. And, in fact, during the past economic cycle we've seen fairly robust economic growth combined with flat or slightly falling wages -- just the sort of shift, albeit on a smaller scale, that Hanson attributes to the robots.
Now, we don't have any artificial intelligence, but we are living through a period of rapidly improving information technology. It's at least possible, in other words, that we're witnessing the beginnings of something like this forecast transformation to a world in which the economy will grow faster-than-usual but wages will grow slower-than-usual or possibly even fall depending on your situation.
Were that to happen, it would have very adverse consequences on most people's living standards, even though the country as a whole would have a much greater quantity of goods and services. That's a bad outcome. Fortunately, though, it's an outcome we could avoid by taking steps to disperse the ownership of capital so we don't have a two-tier society in which a small minority of people own the bulk of the assets and the majority of people overwhelmingly derive their earnings from labor.
Looks like Matt's joining the "Ownership Society."
May 3, 2006 9:57 AM | Reply | Permalink
It's not obvious to me that we would have lower living standards. Many products would be cheaper, and healthcare might be better as well.
You don't even need artificial intelligence to achieve much higher levels of automation. All you need is robots that can manufacture more robots and manufacture other things as well. Living cells achieve this without intelligence. We don't have the technology yet to build an artificial self-replicator, but the obstacle is not really the same as the obstacles to AI.
I agree that one danger is the concentration of ownership as individuals without any remarkable skills may have trouble coming up with anything of economic value to offer, making it impossible to enjoy the benefits of global increases in productivity.
For the situation to be sustainable, we also need to overcome resource limitations as well. I'm imagining a scenario along the lines of self-replicating robots mining the asteroid belts using solar power and sending goods back into the gravity well of earth. It's a dream, but not one that violates the laws of physics or requires fundamentally new technologies.
If we've reached that point, then there is a big question of allocating what should look like a bounty. That's the kind of problem we'd want to have. I don't have the answer, but I've probably wondered about it since I was first exposed to the prospect of total automation as a teenager. Giving individuals ownership shares on capital from which to draw dividends may sound good, but it strikes me as just a bandaid solution. Once the notion of labor becomes irrelevant for the fast majority of humans, we're going to have to rewrite a lot of economic assumptions.
May 3, 2006 10:19 AM | Reply | Permalink
How is this different from Rifkin's the End of Work?
More seriously, the simple solution is the the enforcement of the zeroth and first laws of robotics.
I for one welcome my new cyborg overlords.
May 3, 2006 10:24 AM | Reply | Permalink
The impact of rebots on the economy are so over discussed. Let's think beyond them to when we all have replicators powered by dilithium crystals. What kind of economy will we have then? What kind of economy would we even need?
May 3, 2006 11:15 AM | Reply | Permalink
I don't get it. How do you get order of magnitude increases in growth at the same time as "dramatically falling wages"? The authors say:
These models also suggest that wholesale use of machine intelligence could increase economic growth rates by an order of magnitude or more. These increased growth rates are due to our assumptions that computer technology improves faster than general technology, and that the labor population of machine intelligences could grow as fast as desired to meet labor demand.
Aren't the authors making the startlingly unrealistic assumption that production decisions, and labor demand, are somehow independent of overall demand for consumables? If wages fall dramatically, so does consumption. From where comes the demand that is supposed to drive the booming growth in production?
Also, they seem in some way to assume that the economy is stagnant as to the forms of production that occur, and only changes with respect to the quantity of production. If humans are put out of work by machine labor in established areas of production, then don't those humans put themselves to work in other areas where they enjoy a comparative advantage?
May 3, 2006 1:29 PM | Reply | Permalink
Far from being outlandish, the scenario is very nearly inevitable. I work designing tooling and automation equipment, and several times in the past couple of years looked at doing "lights out" systems for manufacturing this or that part without ongoing human labor. In some cases it might have been cost effective, if we'd had more capital on hand. And I'm part of an early effort to build what could become an open source self-replicating factory, the "reprap" project:
http://reprap.org/
The future looks grim for people who don't have any skills a machine can't replicate. An awful lot of our current workforce is employed in the role of glorified "pick and place" robots, doing nothing even the current generation of machines couldn't do with some good programing.
The obvious question is, if those people lose their jobs, who's going to buy the output of the automated factories? The obvious answer is, these people are NOT as stupid as the jobs they do, and have at least the potential to do something that requires genuine intelligence.
We'd better find a way to realize that potential.
May 3, 2006 2:59 PM | Reply | Permalink
Weighing in as a computer professional long interested in AI and of course as a Transhumanist:
First of all, without a decent simulation of human conceptual processing, replacing humans with robots is not entirely feasible. However, that says nothing about still being able to do a fair amount of replacement of manual labor with robots.
The Japanesem, however, some years ago discovered in developing "lights out" factories that it wasn't that easy - their robots tended to be too inflexible and had to be re-engineered too often to deal with changing business and manufacturing needs.
That said, I believe it was Dr. James Albus that first promoted the concept twenty or thirty years ago - ah, here we go, via Google:
Peoples' Capitalism:
The Ecnomics of the Robot Revolution
Original 1976 Book Online
And his Web site where you can link to the book and his other documents.
His proposal was that robots would make everything and everybody would get a share of the wealh produced as s sort of dividend from the Gross National Product or something - read the book, I never bothered. As I recall from the articles I read about it, he figured everybody would be making half a million bucks a year or so in due time, and would then be able to go off and do whatever they wanted with their leisure time - I think he expected an explosion of creativity once the drudgery of "earning a living" was removed from the populace.
Anyway, it's mostly irrelevant since nanotechnology is going to have an even more "destructive" (from the liberal - or perhaps human - perspective, if not from the Transhuman perspective) effect on the future economy within the next three to five decades and beyond (yes, we are that close - you don't have much time to think about this stuff.)
People are warned not to assume technological change happens in a vacuum and everything else stays the same. It doesn't work like that. Human society changes under the impact of only two things - war (i.e., death) and technology (which is why the two tend to go together.)
If you have massive robot manufacturing, or massive AI, or massive nanotech - YOU are going to change just as much as the economy around you. At least you will if you - or your politicians - don't insist on trying to keep everything as it was in 1900...or 2000...
Nanotech in particular is going to RADICALLY reshape the future into something that is barely recognizable to most people living today. In fact, eventually, in this century, it is likely to RADICALLY reshape YOU into something that would be barely recognizable to a human today.
This is why we call ourselves Transhuman...
If you haven't read any nanotech-based sci-fi (and I omit Crichton's works since he is a ridiculous critic of the whole notion of science while posing as a scientist), you should (while remembering that the authors still have to relate their reality to their readers, so things still seem the same - unless the author is VERY good.)
Start by reading K. Eric Drexler's Eighties non-fiction book, "Engines of Creation", which introduced the concept. Note the statement particularly where he says he will describe applications of nanotech in contexts that nanotech will likely make obsolete.
Believe me, robots are the least of your worries.
May 3, 2006 4:04 PM | Reply | Permalink
Thanks for the reprap link - hadn't heard of that one.
As for people being not as stupid as the jobs they do, our real issue is to see to it that they aren't as stupid as the lives they lead and the people they've been electing to office. The jobs are just part of it.
Because people like Bush are the greatest threat to technological progress in the world.
Fortunately, nanotech will lead to neurotech which will lead to AI which will lead to effective computerized education as well as neural enhancements.
Smart will then be feasible - as long as it's still not socially unacceptable.
This is why I'm a "radical" Transhumanist - I don't have a lot of faith in human social systems accepting advanced tech that threatens the human primate model of society.
May 3, 2006 4:11 PM | Reply | Permalink
This probably contradicts my self-identification as liberal, but I sometimes fear that once the drudgery of "earning a living" is removed, an enormous fraction of people will be driven by boredom into destructive behavior. If we're lucky, it'll just be self-destruction through drugs and high risk amusements. But half a million dollars per person per annum buys a lot of guns, so it's also possible that some will use their new leisure time to assemble heavily armed gangs with no purpose but to inflict violence on other heavily armed gangs and bystanders.
Scarcity is one of the banes of human existence--and one that I'm optimistic that technology can solve. But I think a far more serious problem for humans is the existence of other humans with the means and inclination to cause harm. I'm not sure that this has a technological solution.
May 3, 2006 4:20 PM | Reply | Permalink
The real test will be to see if LDS Robots will be able to mesh seamlessly with their non LDS co brethren. Mormon Robots are a strange breed.
Sorry for the bad joke, but whenever disucssions turn towards the future and robots, I start to think a little silly
May 3, 2006 4:31 PM | Reply | Permalink
I'd say your assumptions are based on what I warned against elsewhere - assuming that when everything changes, everything else stays the same.
Not that I'd ever vote against humans doing the stupid thing automatically, you understand...
Albus' concept I think was based on the notion that you get new art, inventions, etc. when you have aristocrats sponsoring that sort of thing - so when the robots are the sponsors, you get the same effect.
Obviously, the real answer is: all of the above. You'd have some people inventing, some people doing art, some people becoming permanent couch potatoes, and some people trying to take over the world with their lousy half a million...
So what? How is that different from now? Except employment is gone...
People who assume that "everybody" will "come together" to do ANYTHING are always wrong. Always.
By the way, people in this thread should try to find a copy of Bob Black's delightful polemical work, "The Abolition of Work" - when he destroys the liberal notion of work as being valuable.
Oops, here via Google is the typed in version of his essay.
The Wikipedia entry about Black and his works is here.
May 3, 2006 4:40 PM | Reply | Permalink
Slashdot jokes might get you banned here...:-)
In Russia, robots enforce YOU!
1. Build millions of robots.
2. ?????
3. PROFIT!!!
May 3, 2006 4:42 PM | Reply | Permalink