Hypercapitalism: Information Theory and Bayes

It has been over a week since I launched the kickstarter for building hypercapitalism on the testnet  at the Texas Bitcoin Conference.  It isn’t going so well to be honest.  It was always a long shot, but we still have some time left so I’m going to do my best to make a case for why we need a new kind of money on as many fronts as I have time for.

Today I’d like to talk a bit about a line of thinking that has kicked in since I attended the conference.  One of the speakers, George Gilder, spoke quite a bit about entropy, time, and information theory.  This was intriguing as I had quite a bit about this in my my talk as well.

One of hypercapitalism core ideas is that since everything else in our world is subject to entropy, that our money should be as well.  Entropy, it turns out has a lot to do with information and how much information available in the economy.  I hadn’t really thought much about the kind of information, but the more I consider it, the more it make sense to me.

Let me start out with a quick caveat here.  Gilder’s latest book ‘Knowledge and Power’ applies information theory to the economy. It is a fascinating idea that I think has a lot of parallels with what I’m trying to achieve with hypercapitalism.  While I love the parts where he discusses the information theory of economics, I do not agree with many of his conclusions and I’d caution from buying too much into some of them.  I’m not sure where the anti-environmentalism and zion centric angle comes in, but I’d suggest a dixie cup church approach.  Drink the good stuff inside and throw the cup away.

The core ideas of the information theory of economics are based on scientific findings of Shannon  that a high information system needs a low entropy carrier.  Information is the surprise and new information that comes along the carrier.  One of the reasons that most data migrates toward the electromagnetic spectrum is that it is the carrier with the least amount of entropy.  It is dependable and thus can carry a significant amount of data.

Gilder argues that the economy can do this if we just get out of the way and let the information in the system be the entrepreneurial spirit.  He’d have government get out of the way and let the market sort it out.  I think he’s missing something obvious about the qualcomm chips that he talks so much about.  An massive amount of design and architecture goes into those chips to make sure that electrons don’t start jumping channels and corrupting the calculations and readings that are operating the CDMA network.  Jumping electrons cause a lot of entropy.

I think there are plenty of ways in which we can regulate the  economy and establishes courses for it to flow through while reducing the entropy in those channels as well. In particular I think we should replace the entropy filled idea of inflation with the planned consistent entropy of a demurrage currency.  If we know that our money decays at a rate 6% then we can make much better plans.  In practice the rate will float, but it can float in a much more consistent manner than uncontrolled inflation.

Gilder’s work led me to some other work by Mishael Milakovic.  Milakovic’s work uses information theory to get to r > g.  The paper ‘A statistical Equilibrium Model of Wealth Distribution’ shows that inequality increases in an economy where where the average growth of wealth is increasing and/or in an economy with higher turnover activity.  In Hypercapitalism we want to do both of these and yet not have inequality rise.  But how?  We must manage them.  We must take the entropy out of the system to maximize growth and we must have some system of redistribution that minimizes inequality.

Another way that we can reduce entropy in the economy is by reducing the complexity of the economic system.   Because we are using the blockchain to record all the transactions that occur in our system, its complexity is greatly reduced.  One of the key parts of a ‘complex system’ is that it subject to both positive and negative feedback loops that can cause a system to go off the rails if the right stimulus is thrown in.  A system where this can happen has more entropy than a system that doesn’t.  Thus the complex system can carry less information.

When we have a financial failure in hypercapitalism we are able to ‘fold the blockchain’ and connect all the places a failing entities money went with all the people that provided the cash in the first place.  While this does not make the investing entities whole immediately, it will make them whole overtime and this drastically reduces the risk involved.  It also dampns the effect of negative feedback loops.  Thus our system can carry more information.

What information are we trying to carry anyway?  Gilder argues that the information in our economy is the novel invention of entrepreneurs.  When new information is injected into the economy it its rewarded with profit until the system catches up and returns the entrepreneurs profit to the average risk adjusted return.  I tend to agree but I think we should go one step further and make the information we are trying to convey with hypercapitalism the entrepreneurial system.  It isn’t just the entrepreneur that generates the value. They are key, but must also procure initial capital somewhere, and have a market for what they produce.  Instead of just rewarding and ‘solving the economy’ for the entrepreneur ‘information’, I want to solve for the ideal systems. 

These Ideal system produce more value and new ‘information’ than existing systems.  We want more of them.  The more novelty producing system we have, the sooner we cure cancer, extend live, end alzheimer's, get to mars, and then go beyond.

This leads us to Bayes theorem and the ways in which hypercapitalism seeks the optimum information in the economy.  Because we pass back the decaying currency to the person that spent the cash we are in a sense creating a recursive loop.

What is the prior probability that you will spend your money on something that ends up changing the world for the better and creating massive value?  Is it a total crap shoot or do we have some confidence that we can make good, knowledgeable decisions.  Provided that it is not completely random, then a recursive bayesian filtering scheme will move us toward a system with the most information in it.  Hypercapitalism passes these decaying fees back to the person that seeded the original transactions, thus enabling them to do it again.  Some will fail, but if our assumption is correct and it isn’t random, more will reinforce the original probability and we will approach a system with more information.

This redistributing back to source nodes is better than redistributing to all nodes because of the probability that the source nodes of generated wealth will be better at doing it again a second, third, fourth, etc time.

I think this will help us get to local maximums more quickly.  The jump to non-local maximums will be handled by science, discovery, and the limited life span of the nodes on the network(both humans and legal entities).

I’ve tried to leave the math out of this and, to tell the truth, I don’t understand all of the equations.  I’d love to have some help and laying them out by a real professional mathematician.  I’m confident that the ideas are sound.