A Tour of the Ethereum Token Bubble

I was not old enough during the dot-com bubble to experience it in any extent. It must have been something. The scope of it. The market correctly foresaw a leviathan. But it didn’t understand exactly its form. And in its uncertainty and excitement, it went mad.

I’ve read a little about the time, read the archives of GettingIt.com, a grotesquely unprofitable valley-focused webzine published at the height of the bubble, whose offices, on bankruptcy, were inhabited by their former (and then-newly-homeless) editor-in-chief for over six months until finally he was kicked out. Legend has it, he spent most of those months inhaling vast quantities of nitrous oxide, leaving a comically large pile of empty whipit canisters for the landlord to clean up.

The CAPE was at an all-time high of 44, never matched before or since. There were parties every weekend, the launch of each new startup or IPO trigging inconceivable celebrations with open bars, all at the investors’ expense. It was the age of pingpong tables, of ice-cream delivery startups, of eyeballs over dollars and IPO, IPO, IPO! It was a golden time for nerds, and a platinum time for fakers, scammers and wanna-bes.

A hell of party and a hell of a hangover — the granddaddy of all bubbles. Quite an experience to live through I imagine, but one I could never quite understand.

Most of those startups seemed so obviously to be terrible ideas. Why were people buying them? I wasn’t old enough to comprehend this question during the dot-com bubble, but I’m living through the ICO bubble now. And for the first time since I read the archives of Gettingit.com in bemused horror, I finally feel like I’ve had a taste of what its like.

Sure, the ICO bubble is much smaller, but for all that its madness is keener, more concentrated. And the terrible ideas would make petfood.com blush.

What is an ICO? It is not a security, at least not one that resembles any security that existed more than 5 years ago. I suppose you could call them commodities that are purchased overwhelmingly for speculative reasons. But I am getting ahead of myself. First, what is Ethereum?

Ethereum is a blockchain protocol very similar to Bitcoin, but unlike Bitcoin it supports sophisticated forms of transactions known as smart contracts. These are simple programs that define when, how, and in what manner ETH (the name of the native token on the Ethereum network) is distributed. Most importantly for this article, this flexibility allows one to create new tokens on top of the Ethereum platform very easily. And it is these tokens or “coins” that are being speculated on in this bubble, “ICO” being an acronym for initial coin offering.

These coins are sold as vital parts of as-yet-undeployed decentralized mechanisms (in the economics sense, as in “mechanism design”) which are purported to be useful at some point in the future, and their tokens are claimed to extract rent or value from the people who will use these services. As we will see, most of these projects are unlikely to be useful. And of those that have a chance of being useful, most don’t seem to clearly need the tokens that were sold and don’t have a clear path to provide value to the token holders, or are likely to be quickly forked into less rent-seeking forms.

That is, unlike in the IPO bubble, for the most part token holders do not own any direct claim or share in the profits of these mechanisms. It is as if during the dot-com bubble people were not speculating on shares of useless companies but instead speculating on limited-supply coupons that could be redeemed to purchase the services of these companies on some indefinite future date.

I cannot help but think this will end in tears.

The Headless Giants

It is best to illustrate by example. So I am not accused of cherry picking, I will examine the top 8 Ethereum tokens by market cap — Note, I attempted to do this but as this was written over multiple weeks, the ranks changed rather quickly.

We start with Golem.

Golem currently has a market cap of $302,742,047 and intends to become a decentralized marketplace for computing. The team claims it will become cheaper than AWS, which I am highly skeptical of. Decentralized services presume you can’t trust those who provide you the service. Golem, for example, requires redundant layers of computation to ensure people do not game the system with fake computational results. Because of this, it seems very unlikely it will ever be more efficient than centralized services, especially ones as competitive and at such scale as modern cloud providers are.

But I could be wrong; whether Golem will prove $302,742,047 useful is a matter of opinion, but what is not an open question is that the Golem Network Token (GNT) is a clumsy funding hack that decreases the usability of the platform. Here they announce the token:

The Golem Network Token (“GNT”) account is a core component of Golem and is designed to ensure flexibility and control over the future evolution of the project. GNT is created during the crowdfunding period (described in this whitepaper) and, following the first major release of Golem, GNT will be attributed a variety of functions in the Golem network. ● Payments from requestors to providers for resource usage and remuneration for software developers is going to be exclusively conducted in GNT. ● Once the Application Registry and Transaction Framework are implemented, GNT will be necessary for other interactions with Golem, such as submitting deposits by providers and software developers or participation in the process of software validation and certification (as described in the Application Registry section). ● The general conditions for using GNT will be set in the Transaction Framework, but specific parameters of these interactions will be possible to define within each software integration.

Notice how the value of GNT is defined not by what it enables, but what other tokens are proscribed from doing. GNT will be valuable because people will be prevented from purchasing Golem’s services with any other token. Instead of paying for your computations in ETH, you will have to go to an exchange, lose ~1% in exchange fees, this possibly on top of what you lost buying ETH, hope the volatility of GNT doesn’t screw you, then pay for your computations.

And if this is worrying for users, it should be doubly worrying for investors. As Golem is entirely open source, there is nothing stopping anyone from forking Golem, pulling out the Golem Network Token and creating a better product that lacks the rent-seeking, using ETH instead. The classic response to this line of argument is “network effects”. But Golem is not a social network. It is a marketplace for a commodity with very low switching costs. The value of a computation platform doesn’t scale with its users in the same way the value of a social network does. The switching costs for both users and providers of computational services is very low comparatively.

So GNT’s value comes at the expense of Golem’s users, offering no benefit over ETH. What about the next on the list, Augur’s REP token? REP gets a pass, as it looks to be a straightforward implementation of TruthCoin and I’ve always thought TruthCoin was very clever. Unlike Golem, it could not function without a second token. Though I doubt it is a good investment, it at least is not obviously a stupid idea.

But back to the superfluous tokens. Next up, GNO. I hate to pick on Gnosis, as they are creating infrastructure for prediction markets, and this is God’s work. Nonetheless, their token, GNO, is duct-tapped into their contracts in an even more wacky way than Golem’s. Here they answer that perennial question, what are Gnoisis Tokens?

There is a fixed number of GNO tokens (10 million) . Up to 90% will be sold in an auction.
There is a 0.5%(*) platform fee for creating predictive assets (requirement for trading).
GNO tokens generate WIZ (short for Wisdom) tokens . The more WIZ that is used on the platform, the more it will be generated per GNO
WIZ can be redeemed to pay $1 worth of fees (very similar to KFEE).
Fees can also be paid with the token that is used in the market (ETH/BTC token/…) . Once they are paid, the fees will be sent to a third party auction contract and auctioned off for GNO.

Simple right… They explain further, and articulate my concerns about how easily such projects are forked to the benefit of the user and the perdition of the investor:

How Can Gnosis Remain Viable if Participants Choose Not to Pay in WIZ? A core value proposition of Gnosis (and decentralization) is to guarantee future characteristics of platforms to both users and developers without relying on the trustworthiness of an operating company. In order to do this, elements including fee rates, must be codified into the software itself. It is expected that WIZ will be the overwhelmingly predominant method for paying fees in the Gnosis ecosystem. In the unexpected event that this is not true, and users are paying in BTC or /ETH, the platform may become vulnerable to low-fee copycats or potentially even illegal forks of the Gnosis codebase. These alternative platforms may logically cause erosion of the Gnosis userbase, subsequently triggering justified loss of developer confidence that their created markets and applications will remain viable on Gnosis. In order to avoid this scenario, we designed a fee-reduction mechanism to bolster competitiveness of the Gnosis platform. The result is added confidence for developers and partners that Gnosis is the infrastructure they should be building markets on.

So GNO is necessary to create WIZ, and WIZ can be used to pay trading fees. But unlike Golem’s token it is not artificially required to be used to pay for anything. However, they have a janky incentive structure designed to make WIZ more appealing, called the fee reduction mechanism. And this is necessary because if it wasn’t necessary someone would create a fork where WIZ wasn’t necessary, which would be disastrous.

They elaborate:

Two core requirements for the fee reduction mechanism is that it is both decentralized and costly. The mechanism must be costly in order to eliminate spam or manipulation. The core functionality of the mechanism is as follows: All fees paid in BTC/ETH/Tokens go to an auction contract outside the control of the Gnosis team. If fees exist in the auction contract, any GNO token holder can submit a bid, bidding their held GNO against some amount of fees contained in the auction contract. If the bid is accepted, the GNO will then enter the auction contract and the user will receive the fees specified. When the user’s GNO enters the auction contract, the fee reduction mechanism will be triggered causing a reduction in fees on Gnosis proportional to the total amount of GNO held in this auction contract. The auction contract is one-way and GNO cannot leave this wallet

There is a notion in design of elements emerging from the process exactly. Of carving a tool just so, to solve the problem simply and naturally. Reading their whitepaper where everything seems to work fine with just ETH, I seriously wonder if the team at Gnosis can honestly say the WIZ and GNO tokens need to be there for any reason other than to provide a fig-leaf for raising a bunch of money from some very unknowledgeable investors.

Next up, DigixDAO. DigixDAO tokens are just a normal security that offers people a share of transaction fees from a tokenized gold business. That is, Digix will allow one to redeem gold in exchange for a token (DGX) it creates. The holders of DigixDAO will get a percentage of the transaction fees required to send said tokens to others. Like Augur, this doesn’t strike me as obviously economically crazy. It’s a perfectly ordinary security really. However, there is an enormous amount of counterparty risk involved, as well as monstrously huge regulatory risks. Still, assuming hypothetically this is not a scam and ignoring the extremely enormous regulatory risks (and these are gigantic gimmies), their token as described at least makes economic sense.

It is sad that being not obviously economically crazy is enough to put you in the top decile of projects in an ecosystem, but nonetheless, in this space this is the case.

Next on the list is Round, a mediation platform for online videogame tournaments, which is so vaguely specified it’s hard to say much about it, save that I recommend the team learn LaTeX. But I ask you, even if you assume their idea makes sense, which it really does not seem to, does a mediation platform for online videogame tournaments that has no users and in fact no working product sound like something that would have a market cap of 86 million dollars in a sane market?

Then we have SingularDTV, with its awful pun name. Its horrible site, a singularity of bad web design, has no link to a white paper, and I can’t find any information describing how it works or even what it is exactly. I have no idea what it does and after doing a little Googling I’m not convinced anyone else does either. All I know is it’s worth $97,801,800.

Iconomi is a bit of an ouroboros, as it appears to be an Ethereum index fund of Ethereum tokens. One wonders if the best way to outperform this market is to not participate at all.

Next, Melonport.

In the first version of this article, I conflated its “green paper” with Iconomi’s white paper on recollection and even had the gall to make fun of them to be too similar! However, I’ve just read their green paper, and will rectify my error now.

Much of it is at least interesting. Basicly, Melonport describes a protocol that allows one to create baskets of tokens and manage them with less counter party risk (though it does not eliminate said risk because the collateralized assets themselves in these baskets, obviously, would have some). In brief, they separate the role of managing a fund from that of holding its assets, allowing the creation of index funds and mutual funds with lower fees and theoretically lower risk. In a world where these tokens represented shares of real companies, this would be pretty useful.

However, we don’t live in that world. We live in the world of tradable coupons for undeployed mechanisms. But this could certainly change if regulations change drastically.

Still even if you grant it is useful, from the point of the investors it still looks to me they’ve been shortchanged. Reading the green paper, you will be stricken by the near absence of their token, MNT. It isn’t until you get to a footnote at the end that you see this:

These Melon token[s] can then be used to use the core, as all usage fees are collected in Melon token. Since module developers invest time and effort into building these modules and since many of them will have an active cost of running, such as for example server costs of running a price feed, there needs to be a way to incentive them. Melon solves this by incentivizing module developers in the same way as almost every blockchain incentivizes its miners. By giving them the transaction fees — or usage fees in the case of the Melon protocol — and by creating an amount for them created through inflation. The Melon protocol will do the same thing and thus effectively future proofing development

And so my above points about GNT and GNO seem to apply just as much to MLN. It, too, seems to be a lame funding hack, and just as subject to forking. Further, I would make an even stronger claim than I made of GNT and GNO: because the shares of these baskets are ERC20 tokens and their value is entirely that of the assets they hold, there is no possible moat or network effect that could prevent an MLN-less fork from taking over.

Beneath The Hype

The above tweet is from Vlad Zamfir, a very prominent developer of the Ethereum project, second only in prominence to its creator Vitalik Buterin. Ethereum and smart contracts are very interesting technologically, but they are very immature. This is no secret and people like Vlad and Rick Dudley are happy to proclaim it from the rooftops. But they are shouting into an abyss.

People on both sides of these transactions smell easy money. The result is a potent combination of half-brained investment schemes and brainless investors trapped in a Keynesian beauty contest. Undoubtedly, this is a bubble. Undoubtedly, it will pop. But as with all bubbles, the trick is timing exactly when. And with that I am not able to help.