« FDA Shock | Main | New Cato blog »
Giant fire-breathing robots with rocket boots and laser eyes
So requests one MR reader. Will he settle for the economics of robots?
Start with Robin Hanson's paper. Robin argues that robots will become close substitutes for unskilled human labor. That requires the wage rate to fall to the cost of robot production. Capitalists become extremely wealthy but laborers might die off or at least go hungry. Someday a robot might be as cheap as a laptop is today.
But will this happen? Since the Industrial Revolution, there have been numerous predictions of falling real wages due to the advent of machines. But across any thirty-year time horizon (some would say fifteen-year, but not I) real wages have risen and in the long run they have skyrocketed. The marginal return to capital has not gone up much if at all.
Even if we have really, really good robots (I still think Deckard was a replicant), they won't substitute for all forms of unskilled labor. Maybe they can drive a car, but will they fluff your pillow? The remaining poor will fill jobs robots cannot handle, own small bits of capital, or live off of charity and transfers. Don't forget, we are talking about a ridiculously wealthy and scientifically advanced world. A small capital investment might carry you through the rest of your life. Plus if robots will be so good, can't they help the rest of us learn some skills or acquire some capital?
We will see a "cost disease" for services which cannot be handed over to robots, but so what? Low productivity sectors may take up an increasing share of the economy in real terms, but again this is most of all a symptom of plenty.
The robots also have to compete against technologically augmented humans, whom I suspect will be the real force of the future. Complex biology is hard to master, so let nature handle that and just purchase the mechanical add-ons, no?
So I don't worry about the special features of robot economies. It is simply fears of Malthusian overpopulation but with metal rather than flesh. The difference is that there is a more obvious profit incentive to produce lots of robots, since they can be owned for profit. But modern technology would have pushed up wages even if we had not seen the falling birthrates as of late.
In Battlestar Galactica they call robots "toasters." In their world that may be a morally dubious judgment, but they seem to have the economics just about right. Now if we let robots vote, or if they have torrid affairs with top DOD research scientists who hold the secret computer codes to our planet's defenses, that is another matter...
Here is my previous post on robot economies.
Posted by Tyler Cowen on May 3, 2006 at 07:51 AM in Economics | Permalink
Comments
You're exactly right. Dr. Hanson might be right about falling real wages for unskilled laborers, but people are not locked into the position of being either a laborer or a capitalist. Under such circumstances, a small investment in capital would yield substantial returns. Even if the poor don't invest in capitol, charity would be easy in this world. Give a man a fish, he eats for a day; give a man a fish-catching robo, well, you get the idea.
Posted by: joshg at May 3, 2006 8:15:58 AM
The robots also have to compete against technologically augmented humans, whom I suspect will be the real force of the future. Complex biology is hard to master, so let nature handle that and just purchase the mechanical add-ons, no?
Right -- and don't expect the 'add ons' to be exotic or far off into the future. Instead think bar-code readers and U-Scan checkout lanes, automatic toll booths, automated parking lot gates, RFID tags, online shopping, etc.
In cases where automation is not possible, look for self-serve to play an ever increasing role. You already fill your own drinks at fast food joints -- if entry-level labor gets expensive enough, you'll also be filling your own sandwich (think about it--why stand around waiting doing nothing when you could get your food faster and save money if you did it youself? And you'd never get olives if you didn't want olives) Or give me a lower rate, and I'd be happy to straighten (or not straighten) my own hotel room (to be honest, I don't really want a maid messing with my stuff anyway). Just vacuum and change the sheets and towels when I leave.
Posted by: Slocum at May 3, 2006 9:16:21 AM
I wonder what the ability to create robots (and stationary AIs) will do to energy demand. At $10/hour (say), you can afford to use on the order of 100 kW of power per unit. Ultimately, the population of robots on Earth may be limited by environmental impact of all that energy (and waste heat) production. The first hint of this may be seen with server farms, where dealing with local heat dissipation is increasingly important.
Posted by: Paul Dietz at May 3, 2006 9:19:27 AM
That ain't a hint, it's the real thing. Moore's law has died a heat death because we're not ready to move out of mosfet. One of the things that stalled the late 90's-early 00's tech boom was the inability to obtain coolers for the data warehouses.
That is to say, there is already a substaintial (if very local) environmental impact from the power useage, and it has already changed the direction of uprocessor evolution. We will need to become much more efficient with our handling of waste heat just to keep uprocessors evovling.
But in the end, if we can develope high enough temperature uprocessors, we can probably figure out how to use uwaves to get rid of most of the heat locally. That gives us lots of possible things to do for the next dissapatory step.
Posted by: Nathan Zook at May 3, 2006 9:48:43 AM
Deckard a replicant? In Bladerunner, yes. In Do Androids Dream of Electric Sheep? I'm not so sure.
Posted by: Michael Stack at May 3, 2006 10:04:59 AM
Can anyone explain this statement of Tyler's:
"Robin argues that robots will become close substitutes for unskilled human labor. That requires the wage rate to fall to the cost of robot production."
Would substituting "As a result, the wage rate will fall..." for "That requires the wage rate to fall..." be another way to say this?
As written, it implies that robots will not become substitutes for unskille labor until the wage rate falls to the cost of robot production, which doesn't make sense.
Posted by: ziel at May 3, 2006 10:42:32 AM
I wonder if robot development will get a large boost from the emerging needs in the nursing home/assisted living industry. A lot of unskilled labor will be needed to change diapers and feed me and my cronies in about 20 years or so.
Posted by: Rich Berger at May 3, 2006 10:43:45 AM
Slocum: The salad bar and many cafeterias are already there.
Everyone else: if I can give a little perspective from academic computer science, the robot take over is not going to happen, and people who think it is are wrong in the same way the people in the '50's were wrong when they talked about computers the size of buildings. The reality is that humanoid robots will be unable to compete with embedded computing and nano-devices.
To take Rich Berger's example, you will eventually wear diapers that can change themselves, rather than buying a robot to change old-fashioned diapers for you. And yes, Tyler, your pillow will fluff itself.
Rich is exactly right that assisted living/nursing homes will be a key place for this. Many Japanese companies and American academics are investing a lot of research into technology for the elderly. Most of the research, however, is focused on embedding computing power in ordinary home devices rather than in robotics. Imagine a floor that automatically detects and calls for help when you fall on it and can't get up. At Georgia Tech (where I was a Phd student in Computer Science), some CS professors are working on real examples like a toilet that checks the sugar in your urine automatically every day and clothing that monitors your temp and pulse.
p.s. Robin is very smart to talk about "computers" and "machines" in his paper instead of about "robots".
Posted by: DK at May 3, 2006 11:11:11 AM
The Japanese have been investing in robot technology while Americans have been importing illegal immigrants:
http://www.vdare.com/sailer/japanese_robots.htm
We shall see which proves wiser.
Posted by: Steve Sailer at May 3, 2006 11:18:34 AM
And yes, Tyler, your pillow will fluff itself.
This sounds right. And I think robot labour will eventually dominate many of the industries where we have cheap human labour today -- garment sweatshops, for example, or agricultural labour and whatever it is called in the meat-processing plants.
On the other hand, I'm pretty well certain that no matter how cheap robots become, there will still be a sizeable portion of society that prefers to have humans involved in the manufacturing process. There are, for example, people who will pay a huge premium on their clothes to ensure that every part was hand-made, nothing machined. And this is to say nothing of the psychological pleasures of mastery and command -- the delight in the thought that one has myriads of servants at one's beck and call. Robotic servants and self-fluffing pillows cannot satisfy this desire, but cheap human labour can. Well, and replicants, maybe.
I think there will always be some demand for cheap, (largely) unskilled human labour. The only question is whether there will continue to be a supply.
Re: The robots also have to compete against technologically augmented humans, whom I suspect will be the real force of the future.
I don't see why one would expect the entire population to benefit from technological augmentation. I suppose poor people might have their augmentation paid for my prospective employers -- an indentured servitude-type relationship, perhaps? -- but even then, I don't see why we wouldn't still have a significant residue of unaugmented human beings.
On the other hand, re: joshg's point right at the top ("fish-catching robo") the welfare state would be much less problematic if there were a robot slave army powering our economy. Morally speaking, sure, there may be scolds criticising a society of idleness and plenty, but strictly speaking, the malign effects of welfare (or, on the other end of the wealth spectrum, trust funds) are the recipients' problem, not mine.
Posted by: Taeyoung at May 3, 2006 11:33:47 AM
The robots also have to compete against technologically augmented humans, whom I suspect will be the real force of the future. Complex biology is hard to master, so let nature handle that and just purchase the mechanical add-ons, no?
Heck, this is just what pacemakers and prostheses do. We were mechanically augmenting humans well before we were building robots. Augmentations are more valuable than robots now, and are being developed at least as quickly; I'd expect augmentation to outpace robots for decades to come.
Posted by: Grant Gould at May 3, 2006 11:54:36 AM
Great post, but have you considered the effect of the unemployment of robots? See a great picture of an out of work "toaster" and this video about their plight. Also, here's a BSG Wiki about "toasters" and a look at the latest and greatest in toaster technology.
Posted by: Brian at May 3, 2006 12:24:14 PM
Tyler:
As far as I can tell, the risk free return on capital has fallen tremendously over the last 170 years, but financial markets have improved in such a manner as to allow individuals to capture a larger fraction of this risk free return and to more precisely select the level of risk they want.
Wage rates though, for the lower paid 2/3rds of the US population, seem to have been rouchly constant over this time except for an increase in hourly wages of about 50% between WWI and the depression and a rough doubling between WWII and 1970. If one includes the impact of taxes on real wages the results are even more pathetic.
Welfare has been improved by technology. Today's beggars can get antibiotics, and our welfare recipients have TV and recorded music, but these benefits are almost independent of the return on labor.
It seems to me that we need a new framework for understanding the cumulative effects of long term economic development.
Posted by: michael vassar at May 3, 2006 12:33:37 PM
quote from http://hanson.gmu.edu/econofbrainsimulations.pdf :
"[So far,] machines have mainly helped humans be more productive at tasks that machines cannot do. By complementing humans, machines have so far raised the value of most human labor. ... Previous trends need not continue, however. ... Slowly improving machines have two effects on human labor. First, machines get better at the tasks machines do, which makes all the other tasks more valuable. This complementary effect raises the demand for human labor. Second, some marginal tasks switch from humans to machines. This substitution effect lowers the demand for human labor. So far humans still do most tasks worth doing, and so the net effect has been to raise human wages. But this picture changes dramatically if machines can do almost all the tasks that people can do. Human wages can then fall with the falling price of machines. ... human population growth no longer limits economic growth. Simple growth models can easily allow a new doubling time of a month, a week, or less."
Yes starvation can be prevented by charity, by a demand for things "handmade," and by owning small amounts of capital. And I focus on a brain simulation scenario, which allows robots to do pretty much everything humans can do.
Posted by: Robin Hanson at May 3, 2006 1:25:48 PM
Of course Robin, in the extreme Malthusian situation you allude to charity and small amounts of capital are of limited efficacy and "hand made by natural humans" may imply costs inconcievably greater than "hand made by uploads".
I think that Malthusian scenarios invoke value judgements. Which world do we want? One with greater aggregate utility or one with greater utility for the average observer? Many people are uncomfortable even with "dominant" outcomes where the utility for the top 1% or so is greater than that of anyone alive today and that top 1% is far more numerous than today's total population but where the average utility is positive but lower than that today. Arguably all rational concerns about population growth have always been driven by such discomfort. Ironically, this may also drive romanticism. I suspect that in Roussea's time the average "savage" had a greater lifetime utility than a typical Frenchman, but a lower lifetime utility than a 90th percentile wealthy Frenchman. Obviously the carrying capacity associated with a hunter-gatherer lifestyle + slash&burn agriculture was less than 10% as great as that associated with French agriculture.
Nick Bostrom would remind us to pay attention to the utilities we associate with different types of existance driven by selection pressures and would council us not to burn the "Cosmic Commons".
Posted by: michael vassar at May 3, 2006 2:29:15 PM
Robots are already taking over some unskilled labor as we see in auto-mobile manufacturing today and yes I agree that it will hurt unskilled laborers because of this. On the other hand this can be a huge benefit as it'll push unskilled laborers to actually go out and become 'skilled'. Robots can't operate on their own(just yet) they still need programmers to feed them instructions and mechanics to build and repair them as they break down. This calls on the need for skilled workers to perform such tasks. Robots can do the cheap brainless labor, such as mowing the lawn, while the humans can focus on leisure time with the kids and family. Robots taking over unskilled labor will hurt unskilled laborers for a bit, but in the long run this will only improve society.
Posted by: ddang at May 3, 2006 2:38:20 PM
Of course Robin, in the extreme Malthusian situation you allude to charity and small amounts of capital are of limited efficacy and "hand made by natural humans" may imply costs inconcievably greater than "hand made by uploads".
I think that Malthusian scenarios invoke value judgements. Which world do we want? One with greater aggregate utility or one with greater utility for the average observer? Many people are uncomfortable even with "dominant" outcomes where the utility for the top 1% or so is greater than that of anyone alive today and that top 1% is far more numerous than today's total population but where the average utility is positive but lower than that today. Arguably all rational concerns about population growth have always been driven by such discomfort. Ironically, this may also drive romanticism. I suspect that in Roussea's time the average "savage" had a greater lifetime utility than a typical Frenchman, but a lower lifetime utility than a 90th percentile wealthy Frenchman. Obviously the carrying capacity associated with a hunter-gatherer lifestyle + slash&burn agriculture was less than 10% as great as that associated with French agriculture.
Nick Bostrom would remind us to pay attention to the utilities we associate with different types of existance driven by selection pressures and would council us not to burn the "Cosmic Commons".
Posted by: michael vassar at May 3, 2006 2:48:05 PM
Real wages for men have been falling over a period much greater than 15 years.
Posted by: Half Sigma at May 3, 2006 3:57:39 PM
Oscar Wilde had a particularly interesting take on machine labor in The Soul of Man under Socialism (emphasis in original):
"Now, as the State is not to govern, it may be asked what the State is to do. The State is to be a voluntary association that will organise labour, and be the manufacturer and distributor of necessary commodities. The State is to make what is useful. The individual is to make what is beautiful. And as I have mentioned the word labour, I cannot help saying that a great deal of nonsense is being written and talked nowadays about the dignity of manual labour. There is nothing necessary dignified about manual labour at all, and most of it is absolutely degrading. It is mentally and morally injurious to man to do anything in which he does not find pleasure, and many forms of labour are quite pleasureless activities, and should be regarded as such. To sweep a slushy crossing for eight hours on a day when the east wind is blowing is a disgusting occupation. To sweep it with mental, moral, or physical dignity seems to me to be impossible. To sweep it with joy would be appalling. Man is made for something better than disturbing dirt. All work of that kind should be done by a machine.
And I have no doubt that it will be so. Up to the present, man has been, to a certain extent, the slave of machinery, and there is something tragic in the fact that as soon as man had invented a machine to do his work he began to starve. This, however, is, of course, the result of our property system and our system of competition. One man owns a machine which does the work of five hundred men. Five hundred men are, in consequence, thrown out of employment, and, having no work to do, become hungry and take to thieving. The one man secures the produce of the machine and keeps it, and has five hundred times as much as he should have, and probably, which is of much more importance, a great deal more than he really wants. Were that machine the property of all, every one would benefit by it. It would be an immense advantage to the community. All unintellectual labour, all monotonous, dull labour, all labour that deals with dreadful things, and involves unpleasant conditions, must be done by machinery. Machinery must work for us in coal mines, and do all sanitary services, and be the stoker of steamers, and clean the streets, and run messages on wet days, and do anything that is tedious or distressing. At present machinery competes against man. Under proper conditions machinery will serve man. There is no doubt at all that this is the future of machinery, and just as trees grow while the country gentleman is asleep, so while Humanity will be amusing itself, or enjoying cultivated leisure which, and not labour, is the aim of man - or making beautiful things, or reading beautiful things, or simply contemplating the world with admiration and delight, machinery will be doing all the necessary and unpleasant work. The fact is, that civilisation requires slaves. The Greeks were quite right there. Unless there are slaves to do the ugly, horrible, uninteresting work, culture and contemplation become almost impossible. Human slavery is wrong, insecure, and demoralising. On mechanical slavery, on the slavery of the machine, the future of the world depends. And when scientific men are no longer called upon to go down to a depressing East End and distribute bad cocoa and worse blankets to starving people, they will have delightful leisure in which to devise wonderful and marvellous things for their own joy and the joy of every one else. There will be great storages of force for every city, and for every house if required, and this force man will convert into heat, light, or motion, according to his needs. Is this Utopian? A map of the world that does not include Utopia is not worth even glancing at, for it leaves out the one country at which Humanity is always landing. And when Humanity lands there, it looks out, and, seeing a better country, sets sail. Progress is the realisation of Utopias."
Posted by: Erik at May 3, 2006 10:19:53 PM
I am so impressed with Mr. Hanson's integrals and equalities! One would think that the entire field of economics had been duped into believing that the more equations, derivatives, and power laws one uses (always with exactness -that little "=" sign), then, by all means, the more believable one's work might be considered.
Am I the only one who finds Mr. Hanson's math to be out of touch with reality and with the current state of knowledge of the affairs spoken of? Would it not help if, as he integrates and differentiates, Mr. Hanson pointed out the certainty of the various factors he uses? Would it not be better if he used less mathematics and more English when describing systems whose parameters one hasn't the slightest idea about?
Posted by: glardo at May 4, 2006 1:04:35 AM
But in the end, if we can develope high enough temperature uprocessors, we can probably figure out how to use uwaves to get rid of most of the heat locally.
Um... what? Microwaves? This makes no sense at all.
The ultimate limit will come from direct thermal pollution of the Earth. There's only so much heat you can add to the atmosphere before the temperature increase becomes too large. The solution then will be to move servers off-planet, where there is no significant limit to how big their radiators can be made.
Posted by: Paul Dietz at May 4, 2006 9:03:54 AM
Servers will never again be a dominant source of heat. By all means, we will move as much processing off-planet as we can afford, but embedded devices do and will dwarf the amount of heat being generated by big iron.
But if we can convert that waste heat into uwaves efficiently, we need merely to select a wavelength which has low atmospheric absorbtion, and transmit it into space.
Before that happens, however, I expect that we will get serious about capturing waste heat from buildings, including homes.
Posted by: Nathan Zook at May 4, 2006 9:43:34 AM
But if we can convert that waste heat into uwaves efficiently, we need merely to select a wavelength which has low atmospheric absorbtion, and transmit it into space.
This likely runs afoul of the second law of thermodynamics, unfortunately. Narrow-band radiation has low entropy (depending on bandwidth and brightness). Waste heat has high entropy.
One could imagine a scheme where an intense very narrowband microwave beam is directed at the Earth, where entropy is added to it, and the somewhat randomized photons are allowed to escape to space. A scheme like this is used in laser cooling of certain solids by anti-Stokes scattering. But you'd need almost no absorption for this to come out ahead.
Posted by: Paul Dietz at May 4, 2006 11:44:52 AM
Two words: productivity gains.
Prices keep falling and utility rising while wages remain constant. The average American today is richer in most ways than a milionaire circa 1900.
Posted by: TallDave at May 4, 2006 12:10:11 PM
This is an excellent discussion.
I suppose my original question was about perceptions.
Look at what TallDave said just now: "Two words: productivity gains. Prices keep falling and utility rising while wages remain constant. The average American today is richer in most ways than a milionaire circa 1900."
Tell that to people against free trade!
My fear is that research will be restricted by robots taking very visible jobs, like cashiers and cabbies and roofers.
For the doubters as to the extent of robotics, I'm going to try and post a timeline shortly of realistic expectations at my blog here: I.K.bot
Posted by: Ivan Kirigin at May 4, 2006 9:14:46 PM
