Lindy effect
|
The JC’S favourite Big Ideas™
|
“Traditional systems, like wood-plank keeled boats, have an advantage over innovative systems, like the then-novel plywood trimarans, in that the whole process of maintaining traditional things is well-explored and widely understood: old systems break in familiar ways; new systems break in unexpected ways.”
The “Lindy effect” — formulated by Benoit Mandelbrot and popularised by Nassim Nicholas Taleb — proposes that the life expectancy of an idea is proportional to its current age. The longer an idea has survived — the more stress and attack it has survived — the longer it is likely to continue to last. If exposure to shocks and stresses is essentially randomised, the longer an idea has been out there in the world, vulnerable to the sorts of things that could kill it, and it has not been killed off by them then the less likely the statistical record says those events become — or the more likely the thing is to be antifragile to the sorts of shocks that do occur. Ideas might — and, actually, almost always do — mutate to reflect their maximal utility in a given environment.
Blockchains as a case study
There are different levels and which ideas operate.[1]
At the time of writing, a permissionless blockchain is a novel solution to the problem of definitively recording ownership and registering transfer of goods. It is yet to be fully tested in lots of different operating conditions, so we do not as yet know how robust it is. But “a means of recording and registering title to movable assets” is not a new idea — title registries have been around since at least William of Normandy commissioned the Domesday Book, a thousand years ago. We have all manner of ownership registries for financial assets, land, ships, planes, corporations, intellectual property — any class of asset worth protecting will have some kind of title registry.
Just as blockchain is an instance of the abstract form “ownership registry”, an ownership registry is just an instance of the abstract form “database”.
So will permissionless blockchain sweep away all other forms of ownership registry?
The Lindy effect says when gauging survivorship prospects of competing ideas, all other things being equal consider relative longevity. Absent a technological development that has enabled a new idea which was not possible before, expect the older technology, eventually, to win out, because it is “road-tested” in more battle conditions. If it was going to be found out, it would have been found out by now. So ask: what advantages does a blockchain have, and what disadvantages, as against a traditional register and how important are they?
Blockchains are more secure than conventional registers — virtually unhackable, we are told, and more definitive — what is done cannot be undone, dispense with the need for intermediaries, and operate even in an environment totally lacking trust (where one could not trust a registrar). On the other hand they are slow, inflexible, have capacity constraints, and consume enormous amounts of energy.
Are the advantages the blockchain offers game-changers?
We would say no: at least until the emergence of the internet, the world has lived with the security vulnerabilities of conventional registries — which is why we have them. The flexibility a trusted intermediary registrar offers over pre-programmed code, however smart — is, in most cases, an advantage, not a disadvantage, and trust is also a feature, not a bug, of a commercial environment.
Having a “human” registrar should there be a security compromise is also an advantage. Permissionless blockchains are no more immune to fraud than any system having gormless meatware in the technology stack (the blockchain itself may be unhackable. but the users who transact across it aren’t, and you are only as safe as your weakest link). [2]
Commerce does not work without trust, so a commercial technique that can function even where commerce cannot is not much of a selling point. If you are running a business in Mogadishu, the security of your ownership registries should be the least of your concerns.
These are not compelling reasons to abandon conventional registries, all of which are now in electronic (if less encrypted) form. Not only have conventional registries been knocked by the millennium into robust shape, but blockchains transparently have not. Flash loans, XSS vulnerabilities, sybil attacks, rug-pulls malicious javascript —
See also
References
- ↑ A sound idea might be articulated in a fashionable way, and so become unfashionable, while at a deeper layer the idea is still sound. One can render a twelve bar turnaround in a modish new romantic way which goes out of fashion, while the basic song remains sound. This idea propels Scott Bradlee’s Postmodern Jukebox.
- ↑ See, as a catalog of perfidies, Web3 is going just great ...and is definitely not an enormous grift that's pouring lighter fluid on our already-smoldering planet.