Blockchain - a Certainty Machine

There are thousands of definitions of what blockchain is, ranging from the grandiose (the Internet of Value) to the prosaic (a better database). However, most are statements of potential opportunities with little or no recognition of almost certain limitations.

To add to the confusion, the author proposes (yet) another very general definition, that covers both opportunities and constraints.

A blockchain is a machine for reducing uncertainty, in a closed system.

Looking at the first part, there are, of course, multiple definitions of ‘machine’, literal and figurative, so we will pick one from Webster’s, “a mechanically, electrically, or electronically operated device for performing a task”, in this case a computer device (hardware and software) that performs the task of reducing uncertainty.

An elegant machine to reduce uncertainty

The classic blockchain, where the specific task is to transfer Bitcoins, is certainly an elegant machine, consisting of a clever combination of relatively standard components (encryption, Byzantine Fault Tolerance, decentralized governance, transparency, immutability, open software, and proof of work) which are configured in a novel way to reduce the uncertainty that a particular transfer will fail.

What is more, the design of the machine, published openly by the reclusive Satoshi Nakamoto, allows others to understand the internal workings and to build their own copies to perform a different task.

As with other types of machine, such as automobiles, it opens the possibility for replacing or improving components to reduce uncertainty for a particular task, which is one of the reasons why the blockchain pattern has become so attractive. 

By why ‘improve’ not ‘ensure’ certainty?

 We know from quantum physics that absolute certainty does not exist, so all we can do is get closer and closer to certainty in a particular situation and are happy ('satisficed') when ‘close enough’, with current technology. To their credit, the designers of various blockchain machines are usually quite content to highlight areas of uncertainty, a tactic that, in itself, reduces uncertainty by alerting users to limitations (a sort of “do not operate machine in hot/wet/windy conditions”).

The term ‘trust’ is often used with the concept of blockchain, but trust is an effect of the level of uncertainty inherent in a particular machine rather than it's result, as trust relates to an individual’s subjective beliefs rather than the objective quality of the machine itself.

Doubtless as smart, creative people come to understand the ‘blockchain machine’, all sorts of improvements and uses will result. The opportunities to use new blockchain designs appear boundless, but are there limitations (even the best automobile won’t work on the summit of Everest – yet?).

Open and Closed Systems

General Systems Theory (GST) , developed by biologist Ludwig von Bertalanffy in the 1960s, describes two types of ‘systems’: ‘open’ and ‘closed’.

 A closed system is one where interactions occur only among the system components and not with the environment. An open system is one that receives input from the environment and/or releases output to the environment.

Bitcoin is a ‘closed’ system or one that is self-contained. Bitcoins are mined/ ’discovered’ within the system and are exchanged between parties interacting within the system, although the parties themselves are obviously outside of the system. Parties may transfer Bitcoins between one another using the machine and be almost completely certain that the transfer will take place and can see the effects by inspecting the blockchain. Once effected, a transfer cannot be changed - it is immutable – thereby reducing the uncertainty that the transfer can be disputed later.

As at mid-January 2017, there are some 16 million Bitcoins in circulation with an estimated traded value of around $12.6 billion in mid-January. Note that the blockchain underlying Bitcoin only records the number of Bitcoins in circulation and not the value of each Bitcoin (about $780 in mid-January).

The only interface to external systems for Bitcoin, other than the hardware/software systems upon which the blockchain operates, is with the general economic system, whenever Bitcoins are exchanged for 'real' money or goods. A person will often exchange Bitcoins for money or goods/services with another person but those exchanges are not reflected on the blockchain, only the transfer of Bitcoins. Exchanges of money or goods are governed by the laws in place in the wider economic system which is ‘open’ and full of uncertainty, not least about the source of money or the quality of goods.

So, the wider economic system, although subject to extensive regulations, is not as efficient as a blockchain machine in reducing the uncertainty of certain types of economic exchange. On the other hand, it is much more flexible (allowing goods for goods bartering, for example)

In attempting to increase the flexibility of a blockchain, designers have generally looked to loosen constraints. 

The most obvious is restricting access to the blockchain ‘machine’, so-called ‘permissioning’. If we know everyone in the group, and have alternate sanctions then we can reduce transparency and the need for general consensus with only a little increase in uncertainty. Likewise, if a central authority (such as a clearing system) is introduced then payments and deliveries of goods can be ensured with only a small increase in uncertainty.

On the other hand, if we permit the operations of a blockchain machine to be changed, then there is an increase in uncertainty, as occurred recently when the DAO blockchain was drained of $60 million value by a rogue ‘smart contract’.

In other words, the more a blockchain is opened, the less effective it is in reducing uncertainty.

So, there must come a time when relaxing the constraints raises uncertainty to a level that a blockchain machine will become inferior to other solutions for the task in hand.

What does this mean in a practical sense?

The most obvious is that uncertainty is a cost and the ‘value’ of a blockchain machine should reflect its ability to reduce uncertainty and that should be priced into its use .

The second is that fruitful opportunities for using a blockchain machine are more likely to occur in a ‘closed’ system, where it excels, rather than in an ‘open’ system, where uncertainty is rife.

This is why in finance, possible opportunities for blockchain have concentrated on more certain, (mainly) closed, systems such as post-trade settlement systems. Economic systems that are by their nature somewhat closed, such as loan syndication and equity issuance, are also obvious candidates for a blockchain approach.

Proponents of a blockchain approach sometimes take a different tack and propose to ‘close’ an open system with an argument such as, 'if everyone used a cryptocurrency, such as Bitcoin, then all problems would be solved’. Of course, that does not cover the uncertainty when a closed system meets another open system, such as politics.

In short, when considering possible opportunities for blockchain, one should consider: how closed is the system for which the blockchain machine is proposed; what is the task to be performed; and just how much uncertainty will be reduced by the proposed solution.

The answers to those questions should guide the design of the opportunity.