In addition to calling Craig Wright a fraud, Vitalik Buterin addressed serious matters in Ethereum scaling at the Deconomy Summit in Seoul, Korea April 4th.
— Blockchainlife (@Blockchainlife) April 4, 2018
Buterin has been repeatedly criticized for maintaining some central authority over Ethereum, a network built for decentralization and enormously expensive to run. Many Ethereum developers are reportedly resigned to working on a semi-centralized network as a matter of short-term necessity until the complex programmable blockchain is built out. The plan for Ethereum, they claim, is one of gradual decentralization.
Buterin commenced his “Blockchains and Trade Offs” address with a reference to costs.
“Blockchains…are a far less efficient computer and database than technologies that have existed for over forty years… The Ethereum blockchain has about a 1.04 million X factor (less) efficiency compared to (the) Amazon EC2 (Compute Cloud).”
“Clearly…efficiency is not what blockchains are built for.”
Buterin ran through the motivating virtues of blockchain often absent in central computing, virtues that justify costs to some users: censorship resistance, fraud resistance, transparency, robustness and permissionless interoperability:
“One of the reasons why…some developers dislike…platforms…like Facebook and Twitter is they often have very capricious and sometimes violent opinions about what kinds of products (are) OK to go on top of their platform. Sometimes there are…developers that build whole business on top of Facebook and Twitter and then Facebook just decided, ‘Ah, we’re just gonna shut that guy off and…build it ourselves’…People’s entire businesses were just destroyed overnight because of a one-person dealer decision made by, you know, King Zuckerberg or whatever other King of the Internet at that particular time.”
According to Buterin, with permissionless interoperability one can design a system that interoperates with apps built on Ethereum and with Ethereum itself, “…regardless of whether or not those other operations actively cooperate.”
“So basically all these things are different kinds of monopoly resistance.”
Added features like privacy, according to Buterin increase costs significantly. These persistent decentralized blockchain network costs mean that, even today, nine years in, the two main use cases are still, “…basically cryptocurrency speculation and Silk Road.”
Ethereum is exploring different mechanisms to take down costs, mainly moving contracts off chain and resolving them on chain using “Plasma,” “state channels” and other tech to make it only 1000x less efficient than centralized payment channels.”
“We don’t need to use blockchain for everything…Not all transactions will have to be processed by every node in the network.”
Outlined benefits of these mechanisms are that off-chain “gaming” (task admin and negotiation) between users will increase scaling. The gaming can also be private, “If done right. If done wrong you could end up leaking a lot of data to other people.”
“Purely off-chain activity does not need to be seen by anyone except participants in the game.” However, if users sought “strong guarantees” of privacy by incorporating zero-knowledge proofs, “…your thousand factor efficiency would go up to 100 thousand factor efficiency.”
Another potential cost of this system, according to Buterin, is that off-chain interactions are not immutable until resolved. Users can therefore be “griefed” or “mildly attacked by each other…by locking assets up for some time.”
Off chain gaming using state channels, Buterin explained, is good for, “repeated interactions with some set of participants,” but Plasma Cash would be a better choice in systems where participants revolve in and out of agreements, “perhaps inside some sub-ecosystem.”
Channels, plasma and sharding (on chain), said Buterin, will help to resolve the “complex, multidimensional…trade off curve” currently hobbling Ethereum and other chains. Which mechanism a client chooses, “is determined by which of the core properties you want.”
Buterin concluded by saying that if decentralized network users are willing to sacrifice, “specific guarantees in limited ways, we could create systems…even within a factor of ten efficiency of centralized services…and it could make sense for even centralized services to start adopting these things.”