Why There Will Never Be Just One Cryptocurrency

A friend put an argument to me that I have since heard in several forms. There is only one network layer for the internet, the Internet Protocol, and it won because of the network effect: the more people use one protocol the more valuable it is to use that protocol, until the alternatives are not worth running. Cryptocurrency, the argument goes, is the same kind of thing. There will be one, it will be Bitcoin because it was first and has the momentum, and everything else is noise.

I was unconvinced then and remain so. The analogy fails, and the way it fails tells you something about which technologies converge and which do not.

Nobody Argues With Quantum Mechanics

Consider two scientific theories of comparable standing. Quantum mechanics is confirmed to more decimal places than anything else humans have measured. Evolution by natural selection is about as well established as a theory about the past can be.

Nobody outside physics departments argues about quantum mechanics. There are no school boards demanding equal time for a rival account of the hydrogen atom. Evolution, by contrast, has been under public attack for a century and a half and still is.

The difference is not evidential. The difference is that quantum mechanics touches nobody's interests, and evolution touches religion, identity and how people raise their children. A theory converges to a single accepted version when nobody has a stake in the alternatives. When people do have a stake, they keep alternatives alive indefinitely, whatever the evidence.

The Internet Converged Because Nobody Cared

Now apply that to the Internet Protocol.

Why is there one? Because nothing anybody cares about happens at that layer. Nobody's business, politics or identity depends on how packets are addressed. Everything people fight over on the internet — what may be said, who sees what, who gets paid — happens far above, at the level of the applications and platforms. And notice that at that level there is no convergence at all. There are many social networks, many messaging systems, many marketplaces, and they compete, split and die constantly.

So the network effect produced one protocol precisely because the protocol was boring. It had no distributional consequences. The moment a layer starts deciding who wins and who loses, the network effect stops being the only force acting, and the losers start building alternatives.

Money Is the Least Boring Layer There Is

A currency is not like the addressing of packets. It is like evolution: it reaches into everyone's life, beliefs and pocket.

Which currency wins decides who is rich. Whoever holds the early coins of the winner acquires a large share of the world's wealth for nothing, and everyone else pays for it. I have made that arithmetic elsewhere, when writing about what the advocates leave out. A person who understands it, and who did not buy early, has every reason to prefer a different coin, a new coin, or no coin at all. States have the same reason multiplied by their entire tax base. So do banks, payment companies and anyone whose current position depends on the current money.

A layer with that many interested losers does not converge. Each of them can fund an alternative, and the technical cost of launching one is trivial, as the last decade has demonstrated by launching thousands.

Momentum Is Not Destiny

The reply is that momentum wins anyway, that the first mover's head start becomes unassailable. That is true of boring layers and false of contested ones.

Being first gave Bitcoin the name recognition and the largest holder base. It also gave it the most concentrated distribution of any candidate, which is exactly the property that makes it unacceptable to everyone who is not already inside. The head start is also the reason for the resistance. A late entrant with a fairer distribution, or one backed by a state, or one designed for a use Bitcoin handles badly, has a constituency that Bitcoin's success creates rather than removes.

What I Expect Instead

Not one currency. Several, competing along the lines that already divide people: state-issued against private, fixed supply against managed, anonymous against identified. Each will have a constituency whose interests it serves, and none will be allowed to become the only one, because the only one would be too consequential for the losers to tolerate.

That is not a failure of the technology. It is the technology being important. Things that matter do not get to be the Internet Protocol.