Interview with Emin Gun Sirer: There Are No Truly Decentralized CoinsBy Andy Hao
Emin Gün Sirer is a computer science professor at Cornell University. He is a well-known researcher and builder who has been involved with cryptocurrencies since even before Bitcoin came onto the scene. We caught up with him to hear his thoughts on the "sad" state of decentralization in the cryptocurrency world. You can see some of our own analysis here.
LongHash: How did you get into crypto?
I am a professor of computer science at Cornell who has been active in cryptocurrencies since 2002, before Satoshi came around. One of my claims to fame is having built a system called Karma. It involved the use of currencies that were minted in a decentralized fashion, using proof of work. That’s something that Tron is trying to do right now.
LongHash: You also do data analysis.
I work at finding and measuring the scalability of cryptocurrencies like Ethereum and others, as well as quantifying the level of decentralization in these systems. I’m really a builder. When I’m not building, I try to analyze stuff to find where the problems are.
LongHash: What have been some of your major findings?
These studies, which you can learn more about here, here and here, pointed out the sad state of decentralization. It turns out that crypto communities make a lot of noise about the value of decentralization, but when you look under the covers, the entire coin comes down to a very small number of participants. For instance, Bitcoin’s blockchain is constructed by 19 mining entities, that’s it. Ethereum’s blockchain is constructed by 11. These are tiny numbers. While it’s true that each and every one of these mining entities consists of multiple sub-players, the fact is that they have come together under a unified entity and are operating together as one big business unit. There is a narrative that mining pools are internally decentralized, that there is invisible decentralization in these systems.
That argument turns out to be complete bunk. It’s like the emperor’s new clothes: they claim that there is something there that no one can see or measure or touch. The bottom line is participants in a mining pool are typically in no position to question what a pool operator is doing. They are in no position to detect when a pool operator launches an attack. So this narrative that these entities are internally decentralized doesn’t hold water. They might not be incorporated, but they are very much one group of people operating for a common cause.
Are there any truly decentralized coins?
No, they are pathetic. All of the existing coins are far from the decentralized dream that has been sold to the masses. The protocols that we have are not very good at scaling to large numbers of participants, they have built in forces toward centralization.
LongHash: Does this problem have any effect on crypto prices?
No. Markets are highly manipulated, and we’ve seen time and time again that, even when there is the discovery of an obvious security problem in an altcoin, as was the case with Bitcoin Private, it has no impact on the underlying price. I’ve been in this space for a long time and I’ve seen many security flaws discovered, and many issues unearthed that you would think would be devastating for a coin.
The price often does not reflect what technical people will tell you about the strengths of a coin. Especially in the short term and in the early days of cryptocurrency, we have seen the price move in opposition to what rationality would have you believe. This might be because the masses discount security flaws, thinking they can be patched up. But that explanation fails for coins that believe that code is law. Those coins have less options available because they can’t change their fundamental properties and they can’t respond to security flaws.
LongHash: What is the cause of the crypto centralization problem?
The protocols that we have are not very good at scaling to large numbers of participants, they have built in forces toward centralization. In proof of work currencies, economies of scale, ability to acquire cheap electricity and access to supply chains mean that there will always be a few hardware manufacturers that dominate the mining industry. We’ve seen that Bitcoin mining tends toward centralization, and certain groups become more and more prominent. The only force aiding us is that these mining concerns operate in a competitive industry, and there’s high turnover. But right now, just a few players can easily launch 51% attacks and can censor transactions if compelled.
LongHash: So why aren’t these attacks happening more often?
Because Bitcoin is just not big enough in the grand scheme of things. Why would you bother?
Longhash: Billions of dollars pass through Bitcoin.
The attack just hasn’t happened yet.
LongHash: Other than the possibility of attack, why should we care about the centralization problem?
The root of all this is that there is a gap between the narrative behind the tech and what the tech is actually doing. Every time you have that gap, you are in an industry that can collapse overnight. We saw that with the internet bubble. The narrative was that there would be great riches due to disintermediation, and then people suddenly realized that a lot of these dotcoms were hokum. The narrative was that disintermediation will bring us so much money that we will get rich selling pet food over the internet, and then people realized that this kind of revenue is not achievable through this technology. The narrative was that eyeballs would bring cash, and then people realized: no.
LongHash: What are some solutions to crypto centralization?
We are seeing a new crop of protocols that actually live up to the dream. We’re going to see proof of stake protocols come out in 2019 that can achieve much higher transactions per second, much lower latencies, and allow thousands of people to participate.
LongHash: Why is proof of stake a better approach?
In proof of stake you have a number of participants who hold a coin and agree to participate in selecting transactions. The protocols can be much more efficient because the participants do not have to constantly carry out work. The fact that they hold coins is proof enough that they want the best outcome. In POS protocols, the participants are incentivized by having some stake in the proper functioning of the overall system.
In proof of work, participants are required to burn some energy, and they are incentivized to turn energy into value. The problem with POW is that we end up leaking value out of the ecosystems and into the hands of the power companies. That means that if the price of a coin goes below a certain level, it’s no longer feasible. POS is potentially much more energy efficient, faster, scalable and secure than POW.
LongHash: What are some positive indicators for a coin?
Over the long term, the one thing that is not manipulable is the size of a coin’s community. And I have seen time and time agin that the communities that grow and the communities that are healthy are those that are technically strong. Those with a large stake can always manipulate the price, they can even buy press to fluff up their investment. One thing you really cannot fake is a large community of people developing applications on top of the coin. That only happens if the infrastructure is sound. Over the long term, people vote with their feet and they find themselves using infrastructure that is sound.
LongHash: Have you seen examples of security flaws that have affected a coin’s price?
When the DAO hack was happening in 2016, the Ethereum community wanted to respond by enacting a soft fork to freeze the hackers' fork. Then I received a message out of the blue from a high school student with whom I corresponded when he was a junior. He had written to me saying, 'Hey professor, I’m excited about crypto, I want to apply to Cornell.' The summer of the DAO hack he wrote again and said he was looking through the proposed code for the soft fork and found a problem that had escaped the scrutiny of Vitalik Buterin and many other Ethereum developers. If you try to freeze the complete Ethereum chain, you end up opening up yourself to DOS attacks. This is a good thing, in the sense that you have an uncensorable system. So we wrote a blog post detailing this problem; as soon as the blog post went public, it went viral, and the Ethereum market cap went from 1 billion dollars to 900 million dollars. I felt bad about it at the time but Vitalik assured me we hadn’t caused a 10% drop, but retained 90% value.
LongHash: What is the moral of this story?
The moral is that there are communities that will respond to scientific findings. But many others don’t.