To put it in Daniel Larimer’s own words, he and Vitalik Buterin fundamentally strive for the same goals. Loosely phrased, both want to minimize corruption and maximize freedom in society. What is really interesting is that two geniuses try to solve the same problems with two different approaches. In this article we will give a short overview of these two approaches. The main difference stems from the social assumptions that the two make about their users. Whereas Dan assumes that most people have good intentions and at least ⅔ of the network is honest, Vitalik uses Ariely/Kahneman-inspired behavioral economic models, assuming that “we all cheat just a little”. Parting from these two differing assumptions, Vitalik and Dan design two fundamentally different consensus mechanisms, namely Ethereum’s Casper FFG and EOS’ masternode-based DPoS. In their recently published blog posts both refer to one another discussing the limitations of Delegated Proof of Stake (DPoS) and cryptoeconomics.
To begin with, let us have a look at what Vitalik has to say concerning DPoS and generally coin-holder voting. In this case voters cast their vote by making a transaction and in some cases their voting weight correlates with their coin balance. Vitalik’s general stance towards coin-holder voting is most clearly expressed in his blockchain governance post. An important and beneficial characteristic of coin-holder voting is that it counters Sybil attacks, i.e a flooding of the network with malevolent actors. By considering the vote of actors as a function of their stake, an attack on the network becomes significantly more expensive.
However, coin-holder voting in the form of DPoS also leads to undesirable outcome of “plutocracy”. The people who hold the highest amount of coins and thereby get the highest amount of influence in network decisions, are not necessarily identical with the people who use the network the most and are reliant on a functioning network. Coin-holder voting does not include all stakeholders equally, and worse even disproportionately rewards long-term coin holders to the disadvantage of those who have come to adopt the platform service. Such is the problem with Ethereum long-term holders who do not use their assets for decentralized computing.
Another problem with coin-holder voting in general is that people with low asset holdings have no incentive to vote given their relatively low influence on the final outcome. This is reflected in the DAO carbon vote which had a voter participation rate of only 4.5%. As a consequence of low voter participation, relatively little voting power needs to be held in order to significantly affect big decisions.
As a potential remedy to this problems, in his most recent blockchain governance blog post Vitalik considers DPoS as an alternative to make every coin-holder’s voice be heard. However, he comes to a rather sobering conclusion with regards to the overall merits of DPoS. The incentive structure of DPoS puts the masternode operators in constant competition with each other over who can pay more generous fees to the voters. Vitalik thinks that the payment of fees (“bribes” in his words) would inevitably lead to centralization and the formation of cartels to kick out those who cannot afford to go along with high fee payments. He specifically attacks EOS for currently having delegate rewards of 5% annually. To fetch the highest amount of delegates and accumulate influence, Chinese EOS supernodes have already started to buy votes through promising higher payouts. This kind of “bribery” may be defended by arguing that the supernodes which can verify transactions the cheapest will ultimately win, thereby increasing overall efficiency of the system. However, Vitalik argues that the gained efficiency may ultimately stand in the way of decentralization. As an alternative to achieve consensus he proposes savvy cryptoeconomic design, meaning the deliberate incentivization of honest verification of transactions through punishments and rewards.
For Dan Larimer, DPoS represents a system where votes can only be bought with approval from the community. If coin-holders disapprove of the plutocratic practices of a supernode, they can delegate their coins to another supernode. Evoking a comparison with a radio station he argues that in order to control the transmitter one needs to have coin-holder approval. In the remainder of his blog post Dan examines the limitations of crypto economic governance. Fundamentally, he argues that cryptography cannot be used to make subjective judgement calls, but only to prove logical consistency. Therefore, if the majority of validators decided to ignore the cryptographic evidence needed to impose punishments and rewards, the consensus system would become malevolent.
Both, Dan’s and Vitalik’s approaches have their advantages and disadvantages. DPoS centralization tendencies described by Vitalik should be taken seriously. Dan’s concerns with the inability to put human behaviour and unpredictable social dynamics into cryptoeconomic assumptions, is well reasoned. Both Ethereum and EOS carry tremendous potential to decentralize organizations and should be watched closely. With Pool of Stake, users will potentially have the chance to stake on both protocols. It is conceivable that both projects will simply cater to different use cases. Regardless, at Pool of Stake we are an eager supporter of both projects.