In a previous post I discussed the problem with bitcoin's constant-quantity of money. Wei Dai has commented that he views bitcoin as probably problematic for probably similar reasons. But even an asset-backed currency such as Gold or b-money has a problem.
Hard core money, an objective asset that retains its value is great when doing transactions with potential enemies. It should have an important place in transactions across disparate legal systems such as countries. You want to walk away without future obligation (a threat). "Cash on the barrel head" has its place with potential enemies not mutually adhering to a higher law or assurance of mutual benefit after the transaction (that higher law and mutual benefit are often the same thing). But money does not have to be meant only to objectively optimize isolated transactions without regard to a wider society. It can be more intelligent than that, finding optimal solutions to the prisoner's dilemma on a grander scale, beyond the immediate participants in a transaction.
The problem (or lack of "optimality") occurs in systems where you are not the only one who is important to you. It's not ridiculous or anti-evolution-theory to assume you will sacrifice a part of your profit for the benefit of others, especially if it is a small cost to you and a great benefit to others. If you count your success as dependent on society's success and not just your bank balance, there's more to consider. This is why a constant-value coin is not ideal. By making the value of the asset vary with something other than a stable asset, pro-social law (aka system-wide intelligence) can be implemented.
The fundamental problem with a constant-quantity coin like Bitcoin or gold is that it is an anti-social money. It seeks for the holder to maintain value without regard to the condition of society. Society can go to hell and Gold (at least) will still have value. That's a-social. Past transactions that result in a person holding an asset should be post-invalidated if the sum of those transactions resulted in disaster for all. Every transaction should carry a concern for all, present and future. That is a characteristic of a system that displays cooperative intelligence. There should always be a feedback measurement from the future of the entire community of people you (should) care about back to your current wealth. This feedback is a scientific measurement as if the past was an experiment. It enforces decisions on how to make future measurements, seeking an optimal outcome. Defining an optimal outcome should be the first step. (this is not easy, see footnote 1). Deciding how to measure it is the second step. Deciding how to use the measurement in order to adjust your actions in order to maximum the outcome is the core intelligence (see footnote 2), once you've realized the important of steps 1 and 2. Technology has advanced so rapidly, we never formalized a consensus goal for 1 well enough for defining a number 2. As Einstein said, the defining characteristic of our age is an excess of means without knowing what we want. It used to be we just wanted money so that we could have food, sex, and children, or to have enough pride via money and/or social relative to our peers that we felt justified in having children.
Side note, Nick Szabo has pointed out that keyboard credit from modern banking allows speculators to change the price of commodities as much as supply and demand. In what I describe here, that would need to be prevented.
This is why a coin that adjusts to keep commodity prices constant is more intelligent. Laws against monopolies and pollution can regulate transactions to prevent the anti-social nature of maximizing profit per transaction. That's not the benefit of a commodity coin. A commodity-coin has a different kind of system-wide intelligence. If commodities are in excess to demand, the prices will try to fall. So a currency following a basket of commodities will "print" more of itself to keep commodity prices stable. In a growing economy, the excess money could replace taxes, so it would merely fund government, or it could fund the building of more infrastructure to make it's workers healthier, happier, and/or more competitive with other countries. That would demonstrate intelligence that is good for the system's future. A less intelligent outcome (defined as bad for the future strength of the system) is to print the money to buy off voters or to bailout corrupt, inefficient, useless banks with QE.
Printing more money when commodity prices falls prevents the type of destructive deflation that occurred in the Great Depression. Instead of printing more money, they burned food in the fields. They stopped producing real assets like commodities on purpose instead of producing paper money.
If commodities get scarce, the money supply would contract along with them, raising its value. This promotes savings and working. Theoretically the savings and working would be directed towards more commodity production to return the system to health.
In the first case, an economic boom is allowed because the availability of commodities indicated it could afford it. In the second case a bust is prevented by making everyone work harder.
In the first case, savers are penalized. It should be this way because their capital is no longer needed to invest in producing more commodities. It needs to be spent, and if they are not spending it, then "the people" will spend it on the street, reaping the rewards of past savings. Commodities are the measure because they are the fundamental inputs to everything else that require the largest investments.
In the second case, everyone is biased towards being more of a saver.
footnote 1) Should we have a higher median happiness, or a higher median happiness times number of people? Should we restrict it to people? Or should we have a preference for machines? They're infinitely more efficient (see past posts of my measurements of their ability to acquire energy, move matter, create strong structures, and to think about how to do it). They'll be the only ones capable of repelling an alien invasion and to engage in the most successful invasions themselves.
footnote 2) Intelligence requires feedback from observation to modify the algorithm. Engineering control theory writes it as a differential equation and block diagrams "consciousness" as the subtraction from where you are from where you want to be, and takes action on the difference (the error signal). I am not sure if there's any A.I. that is not a variation of this. If it is not making an observation (science) to increase intelligence, is it an adaptable intelligence? In past posts I've mentioned how the simplest form of this feedback is also an element (like a NAND or XOR gatee) that can be used to implement a complete Turing machine. A house thermostat is an example. There is also a reduction in entropy in intelligence, taking a lot of observation to classify observation into a much smaller set of action. The error-signal-of-consciousness may need a reduction (classification) of the observed world. I believe Schrodinger discussed this in "What is Life?"
No comments:
Post a Comment