So, one place where I think Scott errs is that he fails to make a distinction between coordination problems
and public goods
. To see the difference, consider the following two situations.
1. The English language spelling system is horribly bad. Just about every letter has multiple potential sounds that it can map to, and vice versa, there are combinations of letters that are completely silent in some cases but not others, and even the best possible formalizations of English spelling rules work only 85% of the time
. Hence, we could try to switch to a system that is better, yusing the saem leters to olways refur tu the saem sounds so yu kan noe hou tu pronouns a wurd from looking at it just liek mor siviliesd languages. If we all switch over, then that is better for all of us. But if you switch over and no one else does, then you are stuck writing posts that no one can understand well, and so you suffer.
The payoff matrix looks like this:
|You don't switch||You switch|
|Other people don't switch||(1, 1)||(0, 0.99)|
|Other people switch||(0, 1.99)||(2, 2)|
Where the values are (your utility, society's total utility). Hence, even though everyone would switch over, no one benefits from switching over unilaterally.
2. You run a factory. Do you install filters in order to reduce the amount of pollution you emit? You suffer slightly from your own pollution, but not enough to notice; however, everyone's pollution together significantly affects your health and you would rather no one did. But then, installing filters is expensive.
Now, we have:
|You don't pollute||You pollute|
|Other people don't pollute||(2, 2)||(2.1, 1.7)|
|Other people pollute||(0.9, 1.3)||(1, 1)|
Here, in all cases it's 0.1 units better for you to pollute, regardless of where others do. Coordination problems are solvable through many mechanisms; assurance contracts are perhaps the simplest one, and dominant assurance contracts also work well. Public good problems, however, are much more tricky, because there is always an incentive to defect; it's not just a matter of finding a way to move a rock from one valley to another, it's a matter of moving a rock up a 45 degree slope to a point at the top where the slope is still 45 degrees, and keeping it there
. I think that solving coordination problems specifically really should be a primary objective of the crypto-mechanism-design community.
I also expect that interested consumers would voluntarily pay for services from expert private organisations to report on safety and environmental research in certain areas relevant to them.
Well, the problem is that safety and environmental research is a public good. Although, note that this argument is a rationale for the existence of a government-funded certification agency, not for making its certifications mandatory (for that you have to appeal to either irrationality, or a rather weird countersignalling argument which is really clever and cool in its economath but that I'm not sure applies that strongly in many situations).
That's an issue with the entire approach, given it's based in consequentialism. Unless he argues that he is guided by an objective and superior value system, the whole argument boils down to taking away people's freedom simply because he wants to. If all value is truly subjective, it's hard to point to any problem with that.
For that we have his other FAQ
The problem with everyone voting to pass a law is the same "coordination problem" that already existed. There is a rational ignorance where the cost of learning how to vote responsibly is greater than the value of the vote. Thus no one learns how to vote responsibly.
Now I wonder why you are so keen on DPOS