Tuesday, April 17, 2012

Global Robot Population

A reader points me to this chart - I am not able to independently verify the data source.

If the data are to be believed, our robot brethren experienced a significant die-off in 2010, but then overcame that event by more than doubling between 2010 and 2011.   I guess this would have to be an effect of a lot of robots being retired as a result of the great recession (either that or the data is lousy).

That aside, the robot population seems to have a doubling time of about 2 1/2 years.  If that were to continue, here's the extrapolation (along with the most recent UN projections for the human population):

Needless to say, this extrapolation will prove quantitatively unreliable, but it does serve to illustrate the general timeframe implied by a population of 18m with a doubling time under three years.  The curve may slow down if the robotics industry is unable to develop robots fast enough to find new market niches at the same rate in future as it has in the last decade.

I note that Kurzweil's prediction for the singularity is 2045 - this curve seems to be in broad agreement.

I hate thinking about this stuff because it's so painful.

However, in the interests of not being guilty of intellectual cowardice, here's my best current thinking about how this might play out:
  • This trend will continue because it's in the short-term interests of societal elites.  The median influencer's life can be made better with more robotically produced consumer goods and with service robots to perform tedious chores (or human labor made cheap by competition from robots).  
  • The creative classes can have fun with new toys and with thinking up new uses for the technology.
  • Ever larger numbers of people will continue to be made technologically unemployed by this trend.  
  • Managing the "class formerly known as working" will become an increasing challenge.  More and more of them will present as "criminals", "terrorists", and other undesirable labels since society is not able to provide them with a meaningful way to contribute (and people need meaning).
  • The least disruptive approach to managing this is for the underclass to disappear into technologically mediated secondary universes (whatever TV & video games evolve into).
  • However, the traditional cultural ethics that despise welfare/dependency etc will prevent easy/full use of this solution, and the alternative is to lock up more and more deviants and use more and more sophisticated technology to find and monitor the deviants - managing the risk that they become organized and attempt to overthrow the existing order.
  • Some people will reject the automation trend and there will be an ongoing romantic/back-to-the-land/local food/anti-globalization/anti-technology movement.  To the extent it relies on resources not needed by organized global society, and doesn't oppose "progress" violently or too-effectively, it will be tolerated.
  • Depending on how good the roboticists get how quickly, there's going to become a point where there really isn't enough in it for a sufficiently large fraction of humanity.  I simply see no way this trend can continue without eventually rendering almost all of us irrelevant.  People's basic survival instincts will not tolerate that.
  • However, by that point, there may very well be no easy way back, and all hell will break loose.
Gloomy as hell I know - but whenever I think about this subject I am simply unable to come up with a narrative convincing to myself in which there are many future generations of tolerably happy humans. 


barath said...


I do wonder if it's just true in the short-term (i.e. the rest of this decade) and not beyond that. If what we're seeing in Europe is any indication, at some point it's not so important what social elites want (or what big corporations want, insofar as they have different interests): when debt piles up, energy gets expensive, and the population is unemployed and anxious, even the social elites begin to fragment and turn on each other.

Setting that aside, I think you're assuming that the capital (and energy/resource) expenditure required to build such a vast population of robots will exist. I'm not sure that's the case. Take Greece for example: if you look at what's going on there, many people are going back to the land, and not due to any anti-technology movement. While of course every country will experience the problems of finance, energy, and climate differently, I think Greece might be a rough model of where many developed nations are headed, and presumably these are the same nations that would need to be developing/deploying robots for the projections to pan out.

(On a less related note: I would love to see a chat between you and Tom Murphy on energy and economics, as two physicists who have been thinking about these topics deeply for a number of years but have slightly different viewpoints.)

Brett said...

I don't think it will get that bad. Robots - and especially the kind of robots that would be necessary to actually replace most human labor - aren't cheap, and they rely on a supply of energy, spare parts, and maintenance.

What will most likely happen is that we'll reach some type of equilibrium between robots and "augmented" humans in the labor market. "Augmented" meaning everything from attachable devices like those Google Glasses, to full-blown cybernetic implants.

TiradeFaction said...

"managing the risk that they become organized and attempt to overthrow the existing order."

Doesn't that happen in any social order? And while it's often successful, isn't "profound" political/socio economic change (whether good or bad) dependent on the few times it isn't?

Just saying...

Stuki said...

Assuming Moore's law like increases in AI processing power, by the time robots are capable enough to replace the average worker, they are also only a few years away from replacing even the most "capable" ones.

After all, and unless my knowledge is completely outdated, there aren't orders of magnitude differences between the processing power of individual human brains, along any measurable dimension. Neuron count, synaptic count, or whatnot.

Much of what is "genius", is simply positional. Einstein discovered general relativity. If he hadn't, some other dude (or dame) would have; a few months to years later. Due to the extreme cheapness of copying ideas and discoveries brain to brain, first mover advantages are absolutely overwhelming; hence it's easy to attribute greater differences between the "geniuses" and the rest of humanity than really exist in the flesh.

sunbeam said...

Regarding some of the last points you made:

"Depending on how good the roboticists get how quickly, there's going to become a point where there really isn't enough in it for a sufficiently large fraction of humanity. I simply see no way this trend can continue without eventually rendering almost all of us irrelevant. People's basic survival instincts will not tolerate that.

However, by that point, there may very well be no easy way back, and all hell will break loose."

What does

"there's going to become a point where there really isn't enough in it for a sufficiently large fraction of humanity"

mean exactly?

I don't know whether it is elitist, condescending or what have you. But I don't think the average citizen of the US is even capable of understanding the problem.

I just can't imagine there being a mass uprising or disruption of anything in this country by it's citizens, unless it were some kind of racial thing.

I know we had stirrings in the 1930's with communists and fascists and plain old populists. To me at least, I can't imagine these things happening any more.

The US is firmly and totally under the control of... something. A class of people, corporations, elites it doesn't really matter.

It's really kind of spooky. I'm not a world traveller, but it seems to me we are the most passive population in the world.

I just can't imagine any kind of serious resistance to anything that doesn't involve black people in this country. And even with that, the angry ones always wind up doing what they are told anyway.

I have a mental image of a conveyer belt conducting people sitting on sofas into an incinerator. They are oblivious to everything but eating cheetos and watching Jerry Springer on a widescreen tv, which was thoughtfully supplied to them.

In short, I just don't think the American people have it in them to do much of anything about it, even if they did notice.

Even if you somehow got a message out there in comprehensible terms to them, one minute of Talk Radio, or a few jokes by Jay Leno and the whole thing will be marginalized again.

Orwell said something about the last hope being the proles.

These ain't those proles.

Heck the old Soviet Union didn't do it as well as we do. None of them believed it, from the victims to the ones who profited from the system.

Here the victims believe it.

If you are looking for some kind of mass uprising or social movement it won't happen here.

Unless something happens to turn off television, and the internet now I guess.

Seth said...


We're likely to 'lose our cars' before we 'lose TV/internet'. And by 'lose our cars', I really mean be forced to buy lighter/smaller, more fuel-efficient ones and become actually *careful* about how much we drive them. And from there to "maybe I'd better ride my bike" or "OMG, maybe I'd better live closer to work", etc.

That kind of pain -- trivial as it might seem from a certain POV -- will do A LOT to adjust our received cultural ideas about "the American Way Of Life". You know, the thing Dick Cheney told us was 'non-negotiable' as he lied us into Iraq.

Well, once lack of cheap energy starts *making* it negotiable even Americans might actually become less passive.

Greg said...

This post is unexpectedly gloomy coming from you. Stuart. Here are two, unrelated, reasons to think that things may not be so bad.

(1) It may be that the robot population grows faster than this projection for a while -- and that we will want it to.

Consider demographic projections for the next 40 years. The population of the developed world (and China) is expected to age fairly dramatically, and the dependency ratio, the ratio of old (and young) non-working people to workers is expected to increase equally dramatically.

Consider also that at present about 20% of 80-year-olds have some form of clinical dementia and that this goes up by a few percent per year of age. Sub-clinical loss of mental function is probably more widespread, as are physical problems that prevent elderly people cooking and cleaning.

The working age population of these countries is not expected to grow, and young adults will be spending more time in tertiary education, so the number of workers will decline. There won't be enough people to do all the home help and nursing that will be required. IMHO neither the old nor the young will want people employed as home help if there is a workable alternative. So we may welcome two billion robots, as home help for elderly Baby Boomers who are still in their own homes, and nursing support staff for those who are in care.

We may need another one or two hundred million as tutors and teaching support staff, from early childhood to tertiary undergraduate level.

Korea is already experimenting with both of these uses of robots, and so far the experiments seem to be working. I expect China to start working on this soon, as the long-term results of its dramatic one-generation fall in family size* start to come home to roost.

(2) You appear to allow no scope for the exercise of democracy. How (by what means) do you think that "the class formerly known as working" will be disenfranchised so completely?

Already discussion of the problem of rising inequality in advanced economies is gaining momentum. I don't expect this to bear fruit very soon, but I do expect that it will in the medium run.

Income inequality may well continue to get worse before it gets better, but I do expect that it will get better. Perhaps later in the United States than in less ideologically-driven countries, but even there, eventually.

* The One Child Policy is often given the blame/credit for this, but China's total fertility rate (lifetime number of children per woman) was falling before the OCP, and the policy didn't affect the rate of decline.

Aimee said...

I see one big brake on this vision. Aren't robots extremely energy intensive? From the embedded energy of their manufacture to the energy it takes to keep them "alive?" I think there will not be a big enough surplus of energy to sustain the doubling and redoubling of the robot population. I think we are going to rededicate that energy to things like keeping the lights on. Unless, of course, the people for whom robots make money (or the robots themselves!) are powerful and ruthless enough to simply appropriate the energy, consequences for the rest of society be damned.

wygrif said...

"I don't know whether it is elitist, condescending or what have you. But I don't think the average citizen of the US is even capable of understanding the problem."

Are you serious? What do you think the autoworkers have been complaining about all these years, exactly? It's the folks who relied on manufacturing jobs to survive who have been dealing with this for years, not the educated classes.

sunbeam said...

Yes, I am serious.

I've tried repeatedly to bring up issues I'm concerned with and it's like talking to an uncomprehending, unbelieving brick wall.

I live in the American South.

People here:

1) Do not believe in Climate Change. They notice and frequently remark on how screwed up the weather is (if they are old enough and paid attention to that kind of thing) but they don't believe it.

2) There is no such thing as peak oil. It is the oil companies fault (minority opinion here), they are keeping that magic carburetor secret. Or the government won't let them drill. Or it is the arabs fault in cahoots with Obama.

3) Automation. Birds chirp. No idea what to say. If they do it's something like "It won't be that bad." "Someone will have to fix the robots." "Cool! I didn't want to paint my house anyway!"

They don't listen. They don't want to listen. Bring up something like what drought actually means and there is no real comprehension of what that could mean. No mental image, nothing.

Talk about Lady GaGa though...

dr2chase said...

It's not at all clear to me that we can keep on invoking Moore's Law. Notice that clock rates have pretty much stalled; we can put more stuff on a chip, but we can't run it faster, and we're obtaining a lot of our energy improvements by turning things off when we are not using them.

Insulators (oxide layers) are getting so thin that they're not insulating that well, either.

The sorts of tricks I know of (that I know I can talk about) include things like capacitively coupled "contacts" (avoid the energy/time/space expense of honkin' big wires connecting chips) and capacitively coupled "transmission lines" to get signals across chips more quickly.

The exponential trends look pretty impressive going back 40-some years; look at the last 5-to-8 instead.

Stephen said...

How is the robotic population defined and determined. I have ten time keeping robots in my house they all individually much more effective than a slave counting the seconds or repeatedly turning an hour glass.

Chris Vernon said...

This song is relevant here:

Are robots energy intensive? What's their productive output per kWh?

Economics are mucked up here with high taxes on labour and low, or even subsidies on energy. If it was the other way around, move all the labour taxes onto energy and energy subsidies onto labour does that change the picture?

The Rational Pessimist said...

Stuart. Have you read Martin Ford's book 'The Lights in the Tunnel' (http://www.thelightsinthetunnel.com/)? Ford has a background in some ways similar to your as a computer engineer and analyses the impact of technology on the future economy. His central argument is that technology is now allowing capital to be substituted for labour in ever larger areas of the economy, which in turn is killing demand. The situation differs from earlier periods, such as the industrial revolution, where technology allowed the creation of new capital that required additional labour to operate. He also argues that this modern trend has just as serious implications for white collar as blue collar occupations.

Ford has the economics profession firmly in his sight for criticism, arguing that their complacency over the impact of technology is due to a backward-looking interpretation of the data. Questioning of the social and economic benefit of technology is perpetually labeled ‘the Luddite fallacy’ by economists, but Ford would really characterize this now as the economic fallacy of the Luddite fallacy. As an economist by training myself, I do accept many of Ford’s arguments, but would stress that parts of the economics profession are starting to realize that ‘something strange is going on’. Take Raghuram Rajan’s excellent book ‘Fault Lines`. Rajan saw the easy money and financial liberalization moves of the 90s and 00s as a response to the loss of purchasing power of huge swathes of the lower and middle classes. Only through the creation of debt could the system create enough demand. The culprit for this was deemed to be globalization by Rajan, but globalization itself is really just a product of technology. Maybe one could argue that, up to now, the benefits to workers in India and China more than outweighed by the losses to workers in OECD countries, but as technology marches on I doubt that this argument can be sustained much longer.

With technology perpetually undermining demand, the idea that ‘austerity’ can bring us back to equilibrium appears laughable. And that does not mean that Keynesian pump-priming is the answer either. Ultimately, technological progress should mean that we can produce more goods and services with less inputs (putting peak oil and climate change to one side for a minute). We therefore have to work on creating some kind of political framework that allows this to take place without consigning an ever-wider section of the population to the economic scrap heap.

Axel Boldt said...

I agree with your premise in general terms: human labor will surely become redundant before long. It has happened already in agriculture, it is currently happening in manufacturing (robots), and it will soon start to happen in the service sector (artificial intelligence).

Now here's a more positive narrative about how this may play out.

Once the overall development has become obvious to economists and then to journalists, populist politicians will start to appear who offer the simple solution: tax the robots, and distribute the revenue to the non-working masses. These politicians will be elected, obviously.

What are the welfare-receiving masses going to do with their lives? How are these people, who strictly are not needed anymore, going to create meaning in their lives?

People will still need ways to attract life partners and friends. Most can't do it with money anymore, since they live on welfare. So instead they cultivate all sorts of leisure activities: they learn to sing or play an instrument, they write interesting books or video games, they study philosophy to become intriguing conversation partners, they attempt to make interesting discoveries in science, they develop gardening skills, they become good cooks, etc. There's an infinite variety of activities that increase your chances of attracting partners, friends and admirers and thereby create meaning in your life.

And so they lived happily ever after...

Tim at SHYFAQ said...

"...whenever I think about this subject I am simply unable to come up with a narrative convincing to myself in which there are many future generations of tolerably happy humans"

You must not have read much of Kurzweil, then, because he addresses all of your objections in The Singularity Is Near. He may not be right about everything, but at least he tackles all of these issues. Check it out, and then you can be more specific about those areas in which you agree or disagree.