The economic law of comparative advantage states that cooperation between individuals of differing capabilities remains mutually beneficial. Suppose you are highly skilled and can make eight widgets per hour or four of the more complicated doohickies. Your neighbor Joe does everything the hard way and can make one widget or one doohicky in an hour. You work an eight-hour day and produce sixty-four widgets, and Joe makes eight doohickies. Then you trade him twelve widges for the eight doohickies. You get fifty-two widges and eight doohickies total, which would have taken you an extra half an hour in total to make yourself; and he gets twelve widgets, which he would have taken four extra hours to make!This argument, if valid, would be very comforting.
In other words, even if AIs become much more productive than we are, it will remain to their advantage to trade with us and to our advantage to trade with them.
There is something wrong here because the conclusion seems to say that companies never go out of business and species never go extinct. If we trace that contradiction back into the argument that supports it, the argument starts to appear broken and unfixable. The argument fundamentally addresses the wrong question. The interesting question is whether the addition of a highly productive individual (named "you" in the scenario) can eliminate Joe's role in the economy. However, Joe apparently manufactures widgets or doohickies from nothing, so in this model there is nothing that could possibly eliminate Joe from the economy.
In a more realistic scenario, Joe would need to spend something to produce widgets or doohickies. Even if we assume Joe owns his tools of production, a mine with an unlimited supply of raw materials, and farms to grow food to feed himself, he will still need to spend resources to defend these assets from his neighbors.
Within the context of a functioning country, the defensive expense might be a tax used to support a shared police force, or money spent to hire a private security force, or time spent by Joe patrolling his borders with a shotgun. If "Joe" is a country, then Joe has to spend resources on a military, and the amount of resources he has to spend on a military increases if his neighbors get a more effective military, either through advancing technology or by spending more resources themselves. For the argument, the nature of Joe's minimal required expense doesn't matter much, so long as it is positive.
To make this concrete, suppose we initially have two individuals named A and B. Everyone needs to eat one potato per day. Assuming A gets fed, A can build one widget per day. B is a farmer and, if he receives a widget, he can grow three potatoes, one of which he must eat. People also like to squander resources when they can: they can experience two units of pleasure by wasting a potato, or one unit of pleasure by wasting a widget.
The stable state in this economy has A producing one widget per day and giving it to B in exchange for one potato. B consumes one potato and one widget, and wastes his other potato and experiences two units of pleasure per day.
Now suppose we add a new individual C who also eats one potato per day, but produces 10 widgets. B has a monopoly on food production, so he can get all of C's 10 widgets. B has to give one potato to C for C to survive and generate widgets, and B has to eat one potato, leaving B with one potato to either waste or give to A. The best possible outcome from giving the potato to A is to get a widget, and wasting a widget is less fun than wasting a potato. So instead B will waste the potato and experience two units of pleasure from the spare potato and 9 units from the spare widgets. A will starve, and C will give all of his widgets to B every day.
This outcome is not all that sensitive to the details. The main problem for A is C has flooded the market with all of the things A can produce, and in the presence of C's production, the marginal utility to others of A's production is less than the marginal utility to others of A's required consumption. So long as this property holds, we can add more commodities to the scenario or eliminate B's monopoly on potatoes, and C's productivity will still cause A to starve.
On the other hand, I've seen this advocated independently by two different intelligent people, so maybe I'm missing something. If you see a fix, please contact me at firstname.lastname@example.org.
Notes added 24 Nov 2008:
There are many ways that the economic assumptions can break down. What comparative advantage points out that the mere existence of an effective edge doesn't matter *as much* in an economic context as in, say, a military one.I agree with Hall's present stand. I do not know whether Hall believes his present stand differs from the one described in his book; I think it does differ.
There's a clearer explanation in A Farewell to Alms: A Brief Economic History of the World by Gregory Clark of why we shouldn't find comparative advantage very reassuring: horses were a clear example of laborers who suffered massive unemployment a century ago when the value of their labor dropped below the cost of their food.I have not read "A Farewall to Alms".
Note added 27 Mar 2011: Nobody has emailed a fix. There seem to be likely scenarios that will doom many or all of us. I hope to write more about this before then.