Down the Ethereum rabbit hole

I know smart people who react to cryptocurrency with everything from outright skepticism to guarded indifference to enthusiasm. This year I finally looked down the rabbit hole for myself, and was surprised how much potential is there, specifically in Ethereum; this post is an attempt to collect what I’ve learned. So first, links and commentary on the most interesting stuff I’ve read, and then some personal reactions for anyone who wants to know how crypto relates to guns and Alexander Hamilton.

This is a long post (4k words), and it links in enough material to make a book. Here goes.

What is cryptocurrency?

Here’s Wikipedia. Consider reading the Bitcoin Whitepaper; as the first cryptocurrency, bitcoin demonstrates and implements key concepts. It is system whereby a decentralized network of computers can achieve consensus on which anonymous users own how many tokens, which uses a cryptographic proof of work consensus mechanism to maintain security, and rewards those who operate the network by paying them in tokens. For the rest of this piece, I’m going to assume you have some basic concept of what a cryptocurrency is, and how bitcoin works.

What is Ethereum?

If bitcoin is like a currency based on adding machines, Ethereum is like a currency based on macro-enabled spreadsheets. The bitcoin network keeps track of balances, allows sending and receiving of funds, and has limited other capabilities. Ethereum does all that, and furthermore is Turing complete and can execute code (“smart contracts”) of more or less arbitrary complexity. There’s an explainer here, and again, consider reading the whitepaper. Activity on the Ethereum network is paid for in “gas fees” denominated in ether; simple transfers of ether will have relatively low fees, and arbitrarily complex code will have arbitrarily high fees. So Ethereum is a platform for doing all kinds of blockchain, currency, and contract-related stuff. Ether is both the Ethereum system’s currency, and the medium used to pay fees for all its other use cases.

What can I do with Ethereum?

Obviously you can send and receive ether, to buy stuff from any willing counterparty. But decentralized systems are where it gets fun; you can do some pretty fancy financial engineering using smart contracts. Applications include:

Prediction Markets. You can participate in decentralized prediction markets on topics like the 2020 election; Ethereum co-founder Vitalik Buterin made $50K doing this (note that this link includes discussion of decentralized lending and automated market makers, other technologies described below). In my opinion, this is a very promising use case; for many uncertain events (i.e. absent prophecy or other special knowledge), forward or futures markets tell us just about as much about the future as we can reliably know. Last I checked, activity in this area had ground to a halt, as Ethereum fees have gotten prohibitively high, but fixes are in the works for that (keep reading).

Decentralized lending. The most interesting decentralized lending scheme is MakerDAO, which allows you to post collateral (e.g. ether, bitcoin), pay an interest rate, and receive in return a quantity of dai, a currency that is pegged to the U. S. dollar. Then you can use the dai as a low-volatility medium for transactions, such as in prediction markets, or with merchants who don’t want 10% daily fluctuations in their unit of account. If your collateral value drops, you face what’s basically a margin call; either repay the dai, or the smart contract will automatically do a fire sale of your collateral (auctioning it for dai on the open market) in order to maintain the peg. The peg has held pretty well over several years of experience, including the pandemic and various major crypto price moves. Holders of an underlying governance token, MKR, set the requirements around collateral and benefit from the interest charged—so they are incentivized to require decent collateral and charge market rates, lest they blow up the system (and their holding in it).

If you like, instead of using the dai as a stable currency, you can turn around and use it to buy more ether! Leverage! This makes cryptocurrency values prone to flash-crash type corrections; drops in the ether price can generate further selling pressure, as people either liquidate their holdings or get liquidated automatically for being under-collateralized.

On a more positive note, this type of mechanism could theoretically generate synthetic assets that track any price of interest. You could in principle have trustless, decentralized dai-euro or dai-gold or dai-TeslaStock or dai-inflation-adjusted-global-GDP, assuming non-trivially that you have a reliable index to track (the oracle problem). Of course using a volatile asset to collateralize a token tracking another volatile asset will only work if you accept very high collateral ratios or very high risk of liquidation…hopefully we all have a sort of natural intuition that tells us that using Bitcoin as collateral to do a leveraged purchase of synthetic Tesla stock on the Ethereum blockchain is not wise. But still, this is cool from a technical perspective.

Non-fungible tokens. Ethereum allows purchase or creation of non-fungible tokens (NFTs), proving a verifiable right of ownership to…something (right now most NFTs do not actually convey a copyright). Spending millions of dollars on an NFT is roughly analagous to spending millions on an original painting: in either case, anyone can typically look at the art online, and probably even hang a realistic copy of it on their wall, but the buyer gets the mimetic satisfaction of possessing it in a way that others can’t. In both cases, not something I am likely to do anytime soon, but I’m perfectly content to see a new mechanism by which artists creating public output can take patronage from (possibly very silly) rich people.

Arbitrary tokens. You can create arbitrary fungible tokens, such as currencies or points. This means that yes, if you must, you can create YourNameDogeCoin, piggybacking on Ethereum’s security (and paying ether gas fees) to create your very own highly secure cryptocurrency, and trade it with your friends. And since you can, many people have. Thus stories like this, where a low-liquidity (but nominally high market-cap) coin was donated/thrown away away to the tune of several billion dollars.

Automated market makers. You can implement, supply liquidity to, or trade against automated market makers, notably including constant function market makers, for other crypto-related assets. Thus you can make it really easy for people to trade their bitcoin for YourNameDogeCoin, and earn small commissions when they do! Automated market makers can even be used to do things that aren’t obviously useless, like facilitating prediction market trading (previously linked prediction market story did this) or trading in governance tokens (like the MKR token underlying MakerDAO).

Insurance. You can in theory build smart contracts that pay out only if a hurricane hits a given area, or if rainfall is less than X, or if Donald Trump wins the presidency in 2024 AND there is an earthquake in San Francisco that year, or in the case of any event for which reliable data is available to feed an oracle. Mostly this is still hypothetical, but supposedly you can actually buy flight delay insurance, crop insurance is being sold in Kenya, and various other applications are in the works, so that’s something.

Is front-running trades a problem?

Suppose you see a great opportunity for a trade on Ethereum, so you submit it. What happens next? Very possibly, you get front-run, i.e. someone else sees the trading opportunity and grabs it (or at least part of the value of it) for themselves. This is obviously a problem in traditional markets too, but in Ethereum, with generalized code, it gets really weird.

What if you realize your smart contract has a bug, such that anyone who understands the exploit can drain all the currency from the contract? You might be tempted to just exploit the bug yourself, get the currency safely in your wallet, then fix the bug. But generalized front-runners lurk. These are programs that are listening to the market, automatically testing transactions they see—even complicated, obfuscated contract interactions that have never been done before—and front-running anything that gives an apparently desirable outcome. Thus the fascinating horror story of the Dark Forest.

This is avoidable, however, if you collaborate directly with Ethereum miners to get preferential treatment for your transaction (secret execution that cannot be front-run). Thus you can escape the dark forest.

But if miners can give preferential treatment to otherwise front-runnable transactions, they can equally give preferential treatment to frontrunners, or engage in front-running themselves, thus capturing miner-extractable value (or MEV). There’s an amazing paper “Flash Boys 2.0: Frontrunning, Transaction Reordering, and Consensus Instability in Decentralized Exchanges” by Daian et al (link) that goes over the history of generalized front-running (including multiple bots programatically bidding against each other to front-run the same transactions) and the challenge of MEV.

Miners could theoretically band together to make sure that everything front-runnable is in fact optimally front-run, with value split between miners and those who identify the front-running opportunities. And in fact this has happened over the last six months, via Flashbots, which is now used by over 80% of the network’s hash rate (cryptographic processing power).

This is extremely fair, in a sort of inside-out way that mirrors other cryptocurrency phenomena: it is stable, rule-driven, transparent, logically predictable, and secure, with the cost & benefit of resulting in a hard-edged environment and zero forgiveness for mistakes.

You could also write your own bot to exploit any simplistic behavior by generalized frontrunner bots, and indeed this has been done; one Reddit reaction to generalized frontrunning bots losing a ton of money to this was the classic “Play stupid games, win stupid prizes.”

What are the big challenges for Ethereum right now, and what’s to be done?

Ethereum’s transaction throughput capacity is fully utilized, fees for that capacity have been bid up to fairly expensive levels, its underlying mining algorithm is wasteful, and much of the network activity seems frivolous.

Mining currently consumes an immense amount of electricity and a significant fraction of global graphics card production capacity, mostly to do “proof of work” cryptographic calculations that are then immediately discarded. Hence the wastefulness problem. Some people worry about this from an environmental perspective; I care more about the economics—the cost of all that electricity must ultimately be getting extracted from Ethereum network users. The plan here is to transition to proof of stake, so that transactions are secured by validators who prove they own ether (the stake), rather than by miners who prove they own computing hardware (proof of work). Dishonest validators automatically lose the ether they staked, so they can’t stay dishonest for long, and you would need to buy a significant fraction of all ether in existence to attack the network.

Ethereum then has a roadmap for increasing throughput, principally via sharding (splitting the blockchain into a set of parallel, inter-referenced chains to expand capacity) and rollups (aggregating many off-chain actions into a single on-chain entry). I am optimistic about the Ethereum community’s ability to pull this off in the next few years.

I expect that the chain will return to full utilization pretty quickly after sharding is implemented; if it ends up with spare capacity and low fees someone will just use it to emulate Minecraft, in which they will then emulate a computer, in which they will then emulate an entire universe. (Conclusion: you yourself are probably just a simulation inside the computer-Ethereum-Minecraft-computer of a higher universe! Except not.) But in all seriousness, it does seem like available capacity will tend to be used. It feels like building a toll road in a high-demand area: as you expand it, you may end up with much higher throughput (and tolls) and stubbornly comparable congestion.

So much for wastefulness, fees, and throughput; solutions seem to be underway.

But what about frivolousness? Isn’t this all a bubble?

Much of current cryptocurrency activity does indeed appear to be frivolous and speculative. On one level, this is disturbing. If the hope is for blockchains to provide bank-like services worldwide in a censorship-resistant, decentralized, anonymous way, then CryptoKitties are not what success looks like. Arguably crypto activity is no more frivolous or speculative than Hollywood or Las Vegas, though it is younger and less widely understood. And Hollywood and Vegas are here to stay. But that’s hardly a ringing endorsement.

The early internet bubble makes for a slightly more sympathetic comparison. There were many deeply stupid businesses, many “why would anyone want that?” or “why would anyone trust that?” reactions, and many grossly inflated asset values. But meanwhile, Amazon, Google, and similar businesses were beginning to take over the world, via exponential growth from small base levels. Crypto may be at a similar point.

And there are some niches where the crypto dream really is happening. Nigerians are importing goods from China, using bitcoin to bypass a dysfunctional banking and regulatory environment. I mentioned insurance and prediction markets above. And organizations facing legal sanction (whether deserved or not) really can use crypto to anonymously transact. Admittedly some of these crypto users are ransomware organizations who absolutely do deserve legal sanction, especially the ones who make my job harder. But that’s sort of like saying some people who use guns are criminals—it’s even true, but I argue we’re still better off in a world where people have access to such things. It’s getting harder for tyrannies to control their citizens’ economic activity. In one sense ransomware is sort of the bug that proves the feature.

Undoubtedly, though, much crypto activity is pure speculation (e.g. Dogecoin knockoffs). So…

How can I think about ether’s value, vs. other cryptocurrencies or cash?

Most cryptocurrencies will likely be more or less worthless in the long run; they compete directly with each other, and there are winner-take-all network effects. Thus we might expect there will be only one general-purpose cryptocurrency in the same way there’s only one Google or only one Facebook. There might be any number of subsidiary tokens, like the MKR governance token for MakerDAO, but these behave more like equity securities, not currencies in their own right. And there will probably be a Yahoo-analogue, limping along for decades, sustained by a few niches long after it has lost the war. But which one will be dominant?

An obvious first question is “what’s dominant now?” Bitcoin and Ethereum are head-and-shoulders above the rest of the pack. So which of those two is feature-dominant? Obviously Ethereum.

Maybe Ethereum is the MySpace of crypto, and it will be taken down by Algorand or Cardano or some other “Ethereum killer”…but the very phrase “Ethereum killer” points out where the action currently is; Ethereum seems to have a formidable advantage in developer support and adoption. Thus arguments for Ethereum Maximalism (dated, but has aged well), and the view that Ethereum should overtake bitcoin in market cap. Those links lay out the case for Ethereum in more detail.

So assume Ethereum will be dominant. We can then value it based on projected profit from transaction costs and some assumed discount rate. Ethereum fees are currently order-of-magnitude $10B/year. After proof-of-stake operating costs will be low, so the system then looks kind of like a business generating $10B/year of profit. (There are a number of moving parts given pending updates, which will result in fees being burned and stakers receiving newly minted ether, but assume the total incentive to stays on the same order of magnitude.) At a current market cap around $300B, $10B in profit gives a price-to-earnings ratio of 30, which is not too crazy for a tech business experiencing meteoric growth that plausibly aspires to take over significant chunks of the banking, payments, insurance, and tech industries.

The sort of ultimate bull case is that Ethereum, along with its various fee-paying dependencies (including stablecoins), becomes widely used as a blockchain platform, supporting all the applications above and facilitating decentralized economic activity in ways we now struggle to imagine. AirBnB, now a pretty standard kind of internet business, was more or less inconceivable in the first decade of consumer internet; by analogy, there’s no reason to expect we can fully predict the economic consequences of smart blockchains. Maybe it turns out to be CryptoKitties all the way down, but there appears to be some open-ended upside. If it does turn out that Ethereum is comparable to a major currency (with money supply in the $trillions) plus a major tech business (market cap over a $trillion) then its current valuation is cheap.

What can go wrong for the network?

What if people break current cryptographic methods, e.g. by quantum computing? This has been considered; reportedly quantum-resistant cryptographic methods are available, and simply not in use because they are currently less efficient. (The author of that comment is an Ethereum cofounder and has around a billion dollars in value riding on himself being correct, so he should not be casually dismissed.)

What if huge bugs are uncovered in Ethereum? Fortunately the community is paranoid and key network components have multiple implementations, so hopefully they do not all suffer the same bug simultaneously, and those that remain functional hold up the network. There’s also a precedent from one “too big to fail” bug, the issue with “The DAO,” that resulted in a hard fork. This means that most of the network agreed to adopt a sort of arbitrary ex post facto fix (violating the normal logic of the system) that successfully resolved the issue, while the rest of the network split off and became Ethereum Classic, which has relatively little value or adoption.

What if there’s a successful proof-of-work attack attempting theft or denial-of-serice? It’d cost a lot to do – you’d have to either buy a significant portion of global GPU production for several months, or pay a whole set of miners (who already have been buying a significant portion of global GPU production) to kill the goose that lays their golden eggs. And the attacker would have to start soon in order to do it before the transition to proof-of-stake.

What if there’s a successful attack after proof-of-stake? An attacker could buy a significant portion of the ether market cap and then spin up enough validators to execute an attack, either stealing funds or locking up the network. As mentioned above, this would be expensive, and a fair amount of thought has gone into ways the balance of the network could penalize such an attacker (principally by taking their staked funds away) and recover.

What if crypto is declared super-illegal everywhere, and the full force of the law comes down on ISPs to make it impossible to use? Then we’re probably well on our way to a dystopian police state, but I think this is quite unlikely in the U. S., where significant portions of the tech industry and a good-size retail investor base would be howling if there was any indication of going down that road.

What if it’s technically legal and doesn’t get attacked out of existence or dominated by a competitor, but suffers some combination death by a thousand regulatory cuts, slow adoption, significant technical hurdles, delays to the development roadmap, and a continuing lack of non-frivolous use cases? This sort of thing is the most likely failure mechanism in my mind. However, I am optimistic that the forces of ingenuity will triumph over the technical obstacles, outrun effective regulation, and find value-generating ways to use such a powerful new tool.

How does this relate to your broader philosophy?

I would love to see people in dysfunctional countries have alternatives to using local currencies that suffer from inflation, capital controls, demonetization or expropriation risk, etc. I can imagine smartphone-based stablecoin usage becoming significant in much of Africa, which could be a very good thing. It could act like M-Pesa, but with potentially lower fees and/or reduced exposure to government interference.

I also broadly favor free trade and commerce, domestically and internationally, so to the extent cryptocurrencies and blockchain technology facilitate that, that’s a good thing.

Furthermore, I favor free citizens having access to tools that make them more or less ungovernable by an oppressive regime. Thus I think citizens of a free country should be able to own guns, communicate privately using strong cryptography, associate online or offline as they choose, transact without anyone’s permission using media of exchange of their choice (gold coins, cryptocurrency, tantalum bars, whatever floats your boat), and so on. Cryptocurrencies could be an important tool in the arsenal of future dissidents, and their existence thus tends to make freedom more secure—a good thing. The last few paragraphs are more or less the libertarian cypherpunk case for cryptocurrencies.

From a different angle, cryptocurrency is fascinating as an experiment in natural law and institutional design. I believe in natural law morality: there are some things that are inherently right or wrong, independent of any specific divine decree, and that God observes and promotes these laws rather than imposing them ex nihilo. Torturing people for fun is wrong, and God both will not and cannot change that. (The Book of Mormon talks about this sort of thing, in one case stating “Now the work of justice could not be destroyed; if so, God would cease to be God.”) So there are right and wrong ways to conduct human affairs, that need not depend on any central authority. Simultaneously, I am aware that men are fallen creatures, and that they are likely to reach the best outcomes within institutions that incorporate checks, balances, incentive design, etc. such that they can’t easily oppress each other.

The crypto community is working towards institutional arrangements that are cognizant of cheating, yet can nonetheless create common goods and implement natural law (Thou shalt not steal double-spend! Thou shalt not bear false witness against thy [blockchain]!), all without reference to any externally imposed law. In this view, Ethereum founder Vitalik Buterin is potentially a sort of 21st century Alexander Hamilton-as-hacker: a heroic figure driving forward groundbreaking institutional and economic developments that ultimately promote freedom and prosperity around the globe. I would watch that Lin-Manuel Miranda musical any day.

Or again, maybe it’ll just be cryptokitties all the way down, or fail (or languish) for regulatory or technical reasons. We shall see.

Are you invested, and how?

Yes; I have a small holding of Ether via Coinbase Pro.

My personal investment theory is that in general, I am not smarter than the market, so I should have most of my net worth in (1) home equity and (2) very broad, very low-cost stock index funds (e.g. VTI). And that is basically what I do. I hold a bit of company stock to be a good sport, and then allow myself discretionary investments with a small percentage of the cost basis of my non-checking, non-emergency-fund, non-retirement financial assets—whence my small Ether holding.

I expect to hold for perhaps 5-10 years (diamond hands? lazy hands?), to let the current development roadmap play out and see what people do with the resulting capabilities. Recent market turbulence doesn’t concern me too much; I bought before the peak, the investment thesis is that Ether could plausibly be an order of magnitude more valuable within 5-10 years, and nothing about that has changed.

I purchased (and hodl) in Coinbase Pro. There are many other exchanges, and some with lower fees, but Coinbase seems to be the dominant safe option. They have managed their business to a level where U. S. banks, auditors, regulators, and so forth seem to have some confidence that they aren’t just a ponzi scheme for absconding with client funds. They also allow withdrawals in either cash or Ethereum—so cryptographic keys can be had if I decide to use the Ether directly, which is good, but I don’t have to keep track of them myself, which is probably also good. Pro looks to have lower fees than the base version of Coinbase.

My investment is in part a self-conscious act of capital allocation; I think Ethereum is the sort of thing that is worth developing and funding, and I am willing to assume risk of total loss in order to (hopefully) help it along in some small way. Despite a significant risk of total loss, I do think it’s expected-value-positive, which is as close as I’ll come to assuming a distribution of probabilities for different Ethereum outcomes.

I want to go deeper…

I originally saw a bunch of these links via Marginal Revolution, my favorite economics blog (not crypto-specific, and worth following in general). I’ve also been getting the Week In Ethereum News, which gives a nice weekly cross section of what’s going on, particularly on the development side.

But what I really want are cryptocurrency rap videos.

Here you go: bitcoin and dogecoin.

If indeed you have figured something out here, why did it take you so long?

I think one factor was that I never wanted to transact with a sketchy exchange; finding out Coinbase was there, as an option to get exposure without obviously losing my shirt to fraud or fees, was a big deal. So thanks, Coinbase. I also (correctly) figured the ICO boom was pretty shady and (incorrectly) viewed bitcoin as a sort of silly curiosity in 2012-13, when I first recall becoming aware of it. (I spent 2010-11 mostly off-grid as a missionary in Brazil.) And frankly with my diversified investment strategy and limited means immediately after college (graduated 2014), I have not historically contemplated investing in such things. Coverage on Marginal Revolution this year got me interested enough to do the work to understand Ethereum, leading to my current opinions. We’ll see how I feel about it in ten years.

Write a Comment


  1. Very helpful. I particularly like the philosophical background, which I find compelling:

    “ From a different angle, cryptocurrency is fascinating as an experiment in natural law and institutional design. I believe in natural law morality: there are some things that are inherently right or wrong, independent of any specific divine decree, and that God observes and promotes these laws rather than imposing them ex nihilo. Torturing people for fun is wrong, and God both will not and cannot change that. (The Book of Mormon talks about this sort of thing, in one case stating “Now the work of justice could not be destroyed; if so, God would cease to be God.”) So there are right and wrong ways to conduct human affairs, that need not depend on any central authority. Simultaneously, I am aware that men are fallen creatures, and that they are likely to reach the best outcomes within institutions that incorporate checks, balances, incentive design, etc. such that they can’t easily oppress each other.”

  2. Thanks Tom. This answered a lot of questions. It also created a lot of questions, but more on that later. 😉

  3. Very interesting (and for me, mighty complex) article, Tom. I need to reread it about twenty more times 🤔 to arrive at a clear understanding of what you’ve written. I can’t say I agree with your idealistic stance on cryptocurrency providing a mechanism to eliminate oppression but one can always hope!

    • Thanks! To rephrase my claim, I figure that all else equal, dissidents are better off if they can transact via cryptocurrency in a way the state can’t easily monitor or prevent. (I wouldn’t emigrate to North Korea for any amount of bitcoin, and agree it won’t “eliminate oppression,” but every little baby step is positive.)