The Trouble with Ethereum (Part I)

The project is a fascinating latticework castle built on a flawed foundation.

Will Ethereum ever overtake Bitcoin?

"I'm seeing a lot of talk these days about a flippening between BTC and ETH. Is there a scenario in which you ever see that happening?" - SH

An excellent question but also a difficult one. I’ll do my best! None of what follows is financial advice - you should never get financial advice from anyone who asks you to like and subscribe. On a related note, please like and subscribe!

The flippening refers to a hypothetical future where Ethereum passes Bitcoin to become the most valuable cryptocurrency in the world (usually by market cap). That’s kind of a strange target to focus on because it sounds good for Ethereum but it isn’t necessarily - if Bitcoin fell ~50x and Ethereum fell only ~2x then Ethereum would "flip" Bitcoin but neither would have been a great investment.

So the argument about the flippening isn’t really an argument that Ethereum is a good investment. It’s an argument that Ethereum is a better investment than Bitcoin. The reason people care about the comparison between Bitcoin and Ethereum is because many people (myself included) believe that the cryptocurrency market is a natural monopoly where the winner will eventually take all. The people focused on the flippening are people who assume cryptocurrency has already won and hence becoming the best cryptocurrency is the only competition that matters.

I’ve written before about Ethereum generally and about why I think currency markets are winner-take-all - but I haven’t specifically set out to do a head-to-head comparison between Ethereum and Bitcoin directly. I think Ethereum is fascinating technology and I recommend people explore and understand it - but I do not think it will ever "flip" Bitcoin in investment value. I personally maintain 100% of my crypto investment portfolio in Bitcoin and I only own enough Ethereum to try interesting new applications out when they emerge.

Here is why.


  1. What is a blockchain for?

  2. What is Bitcoin for?

  3. What is Ethereum for?

  4. Ethereum as Money

  5. Part II: Ethereum as World Computer


What is a blockchain for?

Blockchains (contrary to popular belief) are not a very useful data structure. Everything you store on a blockchain needs to be broadcast to and stored by every single node on the network - it is an almost comically wasteful approach.1 Blockchains are slower and more expensive than basically any other data structure we know, and they are almost totally useless except for one interesting property - they have no center.

Centralized systems can create efficiency by coordinating the network - but if you can coordinate a network, you can also control it. So decentralized systems deliberately sacrifice efficiency in order to wrest control away from any central authority. A blockchain is an incredibly expensive tool with a very narrow purpose: making a system harder to change. That’s the only thing they’re good for.

A blockchain under the supervision of a central authority is a like a submarine with a screen door: expensive and pointless. Any authority that can be trusted to control a blockchain can perform the same services much more cheaply without it.

Axiom #1: Blockchains are *only* useful for resisting change.

Decentralization is a difficult thing to measure directly, but one way to reason about it is to ask how cheap it is for a new node to join the network. The harder it is to join the network for yourself the more control existing nodes have over your access. If the cost to route around them gets high enough, they effectively control the system.

Unfortunately every node in a decentralized network needs to download and store every transaction on the network, which is expensive. Everything anyone does on a decentralized network makes running a node on that network more expensive, so all cryptocurrencies place limits on how fast new transactions can be added to avoid overwhelming nodes. Those limits are what force users who want to transact into a bidding war to compete for the limited space.

In other words, all blockchains face a fundamental trade off: allowing lots of transactions on the network quickly makes the network cheap to use but expensive to run, allowing only a few transactions makes the network cheap to run but expensive to use. There is no such thing as a free lunch.

Axiom #2: True decentralization is inherently expensive.


What is Bitcoin for?

Bitcoin is also a tool with a very narrow and specific purpose: to be the best possible money. People often think the key feature that makes Bitcoin better money is the limited supply of 21M - but that’s not quite right. The number 21M doesn’t actually mean anything by itself - Bitcoin could have launched with 2.1M tokens or 210M tokens and it would have worked basically the same.

The thing that makes Bitcoin valuable is not how many (or how few) bitcoin there are, it is how hard it is to make more. Satoshi viewed central banks as weak points that invited corruption, so he designed Bitcoin so that central banks were unnecessary and corruption was impossible.

"The root problem with conventional currency is all the trust that’s required to make it work. The central bank must be trusted not to debase the currency, but the history of fiat currencies is full of breaches of that trust." - Satoshi Nakamoto

Bitcoin cannot be seized, censored, counterfeited or debased. In other words Bitcoin is valuable not because of what it can do, but because of what it cannot do. Bitcoin uses a blockchain to prevent control of the system so that no one can force it to do any of the things it is designed not to do.

Axiom #3: Bitcoin value does not come from what it can do, but what it can’t.

Control of money is the kind of power wars are fought over, so Satoshi built Bitcoin to be a military-grade, combat-ready currency. The network is extremely limited in order to make running a node as cheap as possible. Bitcoin sacrifices efficiency and expense to make sure it can never be controlled.


What is Ethereum for?

Bitcoin doesn’t do very much - you can pretty much use it to send and receive bitcoin. Ethereum was originally designed as a response to that constraint. Unlike Bitcoin it is Turing-complete, which means the Ethereum blockchain can perform any calculation that a computer can do. Ethereum can do more interesting and complex things, but that also means Ethereum nodes are more expensive and difficult to run. Using a blockchain directly is expensive (Axiom #2).

Since it is more expensive to run an Ethereum node, fewer projects bother and more rely on centralized third-party services like Infura. Ethereum trades away some difficult to quantify level of decentralization in exchange for more flexibility at the base layer. You could keep making similar trade-offs exchanging decentralization for efficiency, which is what many other cryptocurrency projects have done.

Ethereum can run any program that a computer can run - but the reverse is also true: there is nothing you can do on Ethereum that you couldn’t do 1000x more easily and cheaply using an AWS server. Running applications on a blockchain instead of a traditional server makes them more expensive (Axiom #2) but it doesn’t actually change what they are capable of, only how easy they are to change. (Axiom #1).

Bitcoin’s architecture treats decentralization as mission critical but also a precious resource to be used sparingly on only the most critical parts of the system. Ethereum’s architecture is more flexible and permissive, making it easier for developers to decentralize their applications across the blockchain but in turn making nodes more expensive to run.

Bitcoin is built so that only money itself can be decentralized but that decentralization is as absolute as possible. Ethereum is built so that lots of things can be decentralized but the level of decentralization is less complete. The Bitcoin blockchain is like the Supreme Court - slow to generate new decisions but very difficult to overturn. The Ethereum blockchain is like district courts - more capacity, but less finality.

Arguments in favor of investing in Ethereum generally take one of two forms:

  1. Ethereum as Money (e.g. Ether: the Triple Point Asset by David Hoffman)

  2. Ethereum as World Computer (e.g. Own the Internet by Packy McCormack)

Those arguments overlap and intertwine, but I think it is useful to consider them individually to analyze them on their own merits.


Ethereum as Money

Ethereum can do everything that Bitcoin can do and a host of things that Bitcoin cannot do, so it is understandable why so many people start with the assumption that Ethereum is like Bitcoin, only better. But Bitcoin is actually only useful because of what it cannot do (Axiom #3) so it’s not at all obvious that "doing more" is even good. Money that can surprise its owner is actually profoundly bad.

Consider, for example, the growth of ETH supply over time:

Predicting how many ETH there will be in the future is a genuinely difficult task because Ethereum genuinely retains and demonstrates the ability to change their monetary policy. Even the graph above (the most recent I could find) is already out of date with the launch of EIP-1559.

Ethereum supporters often argue that this is a good thing - so far the changes to Ethereum’s supply schedule have actually reduced the projected supply of ETH.2 In fact with EIP-1559 the supply of ETH is now actually decreasing. If Bitcoin’s limited supply of 21M makes it sound money, the argument goes, then Ethereum’s decreasing supply should make it "ultrasound money".

Unfortunately, the promised number of tokens is not the same as the strength of the promise. Bitcoin’s monetary strength doesn’t come from the number of bitcoin, it comes from the certainty that number cannot change. Ethereum’s token supply is decreasing right now - but the history of Ethereum’s development shows how easily that could be updated. Etherean’s describe their monetary policy as Minimum Necessary Issuance but there isn’t much clarity about what 'minimum necessary' will mean or who should get to decide.

Bitcoin uses a blockchain to enforce the promise that no one can decide to create more bitcoin. Ethereum’s promise is that someone will decide the right amount of new ETH to create. Different people can make different arguments about who is making those decisions, but the history of Ethereum very clearly shows decisions being made.3 Investing in Ethereum means trusting the powers that control it to do a good job.

That’s a fundamental problem because the whole point (and only purpose) of a blockchain is to eliminate trust. If there is a central bank empowered to change the supply of money a blockchain is a pointless and expensive waste (Axiom #1).


Next Issue: Ethereum as World Computer

Ethereum can do a lot of interesting things and as a result a lot of interesting things have been built on top of it. Ethereum is the leading platform for decentralized applications today as measured by developer activity, capital invested or total transaction fees. Ethereum investors talk about ETH as though it were equity shares in the things built on top of ETH ("own the internet") but that’s not a great analogy for reasoning about a decentralized platform …

Continued in Part II

1

The fact that blockchains are so terribly ugly and inefficient is part of why they took so long to discover. Security researcher Gwern has an essay on this from 2011 called Bitcoin is Worse is Better.

2

Good for Ethereum holders but not as obviously good for Ethereum miners.

3

Bitcoin has also updated and changed over the years but those updates have been slow, conservative, opt-in, widely supported and backwards-compatible. In contrast Ethereum’s history is full of rapid (difficulty bombs), controversial (DAO rollback) and fundamental (EIP-1559) changes and more are promised (ETH2.0). Whether you agree with those changes or not they collectively suggest the existence of coordination and hence, control.