Skip to content

When robots are better at everything

September 11, 2011

This post is in reference to the following articles: and

Professor Hanson gave a fascinating talk awhile back about the Economics of Nanotech and AI. Essentially, he attempts to predict what will happen to the economy after the singularity. His prediction for what will be the root cause for the singularity is “Ems” – whole brain emulations that can accurately replicate human thought. When that happens, mental work now done by humans will be able to be performed by specialized Ems, and thus humans will no longer have a comparative advantage in mental tasks that computers are currently unable to perform. He highlights the major distinction of complements and substitutes. If a production technology is complementary, it makes another factor of production more productive. For example, a hammer makes a carpenter more productive. If a production technology is a substitute, you can use it instead of another production technology. For example, cars are substitutes for horses. Technology has until this point substituted for some human labor, such as when automated looms displaced human weavers. It has also been complementary to human productivity, like how spreadsheets help accountants keep track of the money flows in a firm, or how backhoes help construction workers dig holes.

Here is his graph:
As Robots get better, they displace more human workers
The blue-green area is done by humans, and the dark blue area is done by machines.

I do not think that is a good metaphor for technology (all economic stories are to some degree metaphors). Technology is the recipe humans use to turn raw materials into things we like to use. There are scarce resources that we have to make the stuff we want: land, minerals, energy, plants, animals, and most importantly: effort. The amount of land and raw materials humans have had to work with has not changed since the dawn of time. People get richer by minimizing the effort it takes to turn raw materials into goods.

Capital produces goods in a roundabout way. Instead of doing a calculation by hand, you can program a computer to do it. One way is direct, the other indirect, but in both situations, the scarce resource being used it human effort. Capital is not a substitute for effort, it is a form of effort. Capital itself does not need to be rewarded, only the effort it takes to produce it. No one pays the computer, but they do pay the programmers and the humans who manufacture the computers. The price of the capital is determined by the cost of supplying it and demand for the products it helps create. It’s just like normal supply and demand, except that the demand comes from the goods the capital can create rather than the goods themselves.

In the post singularity, any universal constructor will be able to make copies of itself out of raw materials, and so the cost of UCs will drop to the price of the raw materials it takes to build one. If humans can make an EM and EMs can do anything humans can do, than EMs are self replicating. EMs make human effort non-scarce. Non-scarce goods have a 0 price. Thus, unless the holders of EMs have a monopoly on them, they will not be able to extract surplus rents. All rents will accrue to holders of natural resources (and goods EMs can’t make). Robin Hanson states that “most income in the world is going to the machines” in post-EM society (25:20). It won’t – because relative prices will shift.

Relative price shifts are always how groups of humans are made worse off by technology. When the Luddites were “replaced” by the machines, they could still make the same amount of clothing per worker. The reason their wages dropped was because they could not sell their output for the same high price they were used to. While the price of some goods can fall, the relative price of all goods cannot.

Robin’s article assumes that robots will have such a strong absolute advantage that the relative prices of human produced goods will be driven so low that humans will not be able to earn enough *even at the very high productivity levels in the future* to pay for their own subsistence. Frankly, I don’t find that plausible. Humans will have a clear comparative advantage in things that no one would use a robot for today – social interaction, art, sex, child care, and hospitality. In a post singularity work, robots will be able to make robots, so if a holder of a million robots would be willing to part with one of them for the frivolous use of a subsistence level human, that would be all that was required to lift them into the heights of luxury by modern standards.

Further reading:
Nick Rowe
Karl Smith
Andrew McAfee

9 Comments leave one →
  1. September 11, 2011 8:52 pm

    You seem to not understand the concept of an em if you can write “Humans will have a clear comparative advantage in things that no one would use a robot for today – social interaction, art, sex, child care, and hospitality.” I accept that resources may gain a large fraction of income in the long run.

    • September 12, 2011 12:41 am

      Perhaps I don’t. It is my understanding that Ems won’t have bodies, which will make some of the tasks more difficult. I envision a world which is not quite so close to the Malthusian edge as you have written about. In such a world, I think people would care a lot about what other flesh and blood humans thought about and would want to associate more with them for status reasons.

      • chris permalink
        September 13, 2011 2:48 am

        PerhAps it would just be a star trek universe where the physical necessities were worthless and thus available to all. The point would be personal development and enjoysmemt. Sound like predictions of how we would be living in the twentyfirst century made by futurists in the fifties……

  2. September 12, 2011 4:12 pm

    Once we’re talking about theoretical post-scarcity, obviously most current theories don’t apply. It’s more interesting, I think, to think about the periods of transition.

    You have the creative/knowledge economy (those creating the “recipes” as you put it) and the service economy (those providing “social interaction, art, sex, child care, and hospitality”, except for perhaps the “art” part). Those providing the recipes are rewarded far more than those providing the services, precisely because the recipes provide so much benefit but it is so hard to come up with them.

    There seems to be less and less middle ground. Is this inequality “good”, or at least “okay”?

  3. steveroth permalink
    November 18, 2011 11:11 pm

    Thanks for the link, James. I’d be interested in hearing your thoughts on my more recent musings (and those of Eric Brynjolfsson and Andrew McAffee):

    If we add to the thinking a limit on human cognitive capacity, does that change our predictions for the future, and explanations of the present?


  1. The Coming Battle over Intellectual Property « azmytheconomics
  2. Relative Price Shifts and Creative Destruction « azmytheconomics
  3. In Praise of Robin Hanson | azmytheconomics

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: