Vitalik Muses on How PoS Could Have Come Sooner
With The Merge Coming Ethereum's Chief Scientist and Muse Takes Stock
By: Samuel HaigDive
All eyes are on Ethereum’s coming transition to Proof of Stake (PoS) consensus. It’s a historic moment in the history of crypto.
Vitalik Buterin, Ethereum’s co-founder and muse, weighed in this week with his thoughts on what’s to come, and poignantly, the significance of decisions not made in the network’s evolution.
In an article titled ‘The Roads Not Taken’ and published on March 29, Buterin acknowledges that the move to PoS could have happened much earlier in the network’s lifespan.
The chain merge between Ethereum’s mainnet and the Eth2 beacon chain, dubbed The Merge, is expected to reduce new Ether emissions by 90% and cut the network’s energy consumption by more than 99% by doing away with Proof of Work (PoW) consensus.
Vitalik argues that Ethereum’s greatest challenges and criticisms have emerged from the difficulty of balancing two contradictory visions. To hear Buterin tell it, Ethereum must be “a highly performant and functional platform for building advanced applications” but also a “pure and simple blockchain.” In other words, complex and simple at the same time.
But how? In his article, Buterin shared his thoughts on key elements of Ethereum’s next phase and how the network got to where it is today.
Gasper is Coming
Ethereum’s shift to its ‘Gasper’ Proof of Stake (PoS) model and away from Proof of Work (PoW) will kick miners off the network. Instead, ETH hodlers that lock up coins and operate a node will be tasked with validating transactions moving forward.
Ethereum’s environmental footprint will dramatically shrink alongside its block rewards under PoS consensus. Coupled with the fee burn introduced with EIP-1559 last August, many analysts believe Ethereum will become deflationary once it moves to PoS — meaning that more Ether will be burned than created, and possibly paving the way for significant price gains.
But Gasper is simultaneously a “complex” and “very powerful system,” Buterin writes. He acknowledges that adopting simpler versions of PoS could have enabled “large improvement[s]” for the network’s energy consumption and supply issuance at a much earlier date.
“If we had been more modest at the beginning, we could have focused on achieving a more limited set of objectives first.”
He cites NXT PoS, which has existed since 2013 and could have been leveraged for Ethereum written into the code when the network was launched. “If we had been more modest at the beginning, we could have focused on achieving a more limited set of objectives first,” he said.
Buterin adds that switching to a simpler model for PoS at an earlier date could have reduced the scale of both the network’s ecological externalities “and anti-crypto mentality as a result of environmental damage “
However, Vitalik argues that Gasper is highly complex because “it tries to accomplish much more than [alternative algorithms] do,” enabling greater operational functionality for Ethereum moving forward. He asserts that adopting PoS consensus from launch would have been “a mistake.”
Sharding is Key
Sharding describes horizontally dividing up a database to improve the efficiency of how data is processed. In the context of Ethereum, sharding will reduce congestion and increase the number of transactions that it can process by sharing the computational load across a network of smaller chains, dubbed shards.
Currently, Ethereum’s Layer 2 networks are on track to become its shards following future upgrades that will be shipped after The Merge.
In contrast to PoS, the design for a sharded Ethereum has become increasingly simple as research progressed in recent years. Its current scaling roadmap will culminate when Ethereum opiates as a unified settlement layer where all shards execute in parallel within a single block, dubbed danksharding.
Buterin says more complex sharding designs have existed for several years but likens them to little more than “ideas and mathematical models.”
“Danksharding is a complete and almost-ready-for-implementation spec,” Buterin said. He adds that taking time to progressively simplify sharding “was absolutely the right move.”
Despite his preference for a less complex sharding model, Vitalik acknowledges that “more ambitious research also has a very important role to play” in identifying new directions for research that often yield “reasonably simple” innovations.
Supply Distribution and the Question of Liquidity
Ethereum became just the sixth ever initial coin offering in 2014 when it raised funds from the public in exchange for a share of its native tokens; they were distributed when the network went live the following year.
Ethereum’s issuance model meant that a significant portion of Ethereum’s supply would already be liquid when the network went live, unlike Bitcon’s model that will distribute the entire supply of BTC as block rewards to miners. Alongside the 60M Ether earmarked for ICO participants, 12M ETH was distributed to Ethereum’s roughly 100 foundation members and early contributors at launch — a move called a ‘pre-mine’.
Breakdown of Ether supply distribution. Source: Etherscan
The Ethereum co-founder acknowledges criticisms of the pre-mine. Buterin accepts that distributing three-quarters of pre-mined rewards before Ethereum was live “left far too little [ETH] for later contributors.” He also highlights that within six months of the network going live, “the need to sell to financially survive” resulted in the Ethereum Foundation retaining just a third or 1M Ether to pay contributors with moving forward.
“The problems were related: the desire to minimize perceptions of centralization contributed to a smaller pre-mine, and a smaller pre-mine was exhausted more quickly.”
Vitalik asserts there is no clear answer as to how the early distribution of Ether could have been better handled.
Buterin ponders the “‘DAO from day 1’ route popular among some DeFi projects today,” where a share of block rewards are temporarily transferred into a development fund over a specific period (for example, 2 ETH per block for two years). But he speculates that the model would have led to criticism that the network appeared overly centralized.
“Even if the dev fund had been fully credibly neutral, the people who yell about Ethereum’s pre-mine today may well have just started yelling twice as hard about the DAO fork instead.”
Looking Back at the Ethereum Virtual Machine
The Ethereum Virtual Machine (EVM) enables smart contract execution, which is the heart and soul of the network. Buterin notes that it launched in 2015 without many functions that were originally planned for it.
He argues that excluding most of the features that did not make the final cut “have proven to be very good decisions” that prevented security vulnerabilities and kept it simple and accessible.
Layer 2 Optimism to Remove Whitelist and Open Floodgates to Devs
“Sometimes, good ideas just take years to arrive at and there is no better way around that,” he said, noting that many of the functionalities left out have since been addressed with more elegant solutions.
Vitalik notes that EVM could have also been shipped in very different forms, comprising either a higher-level language catering to greater variables, or a copy of an existing virtual machine — such as the computational frameworks offered by WASM or LLVM — to enable greater composability with coding languages already used in computing.
Ethereum’s co-founder asserts that basing EVM off an existing virtual machine was proposed and rejected many times, concluding that “there probably was never a viable path for the EVM that’s radically different from what we have today.”
Buterin criticizes EVM as a higher-level language due to the added complexity its structure would have taken. But he concedes that an opportunity was missed for an approach offering “the best of both worlds.”
Citing EIP-2315, Ethereum’s chief scientist notes that “some EVM changes could have given us a lot of those benefits while keeping the basic EVM structure roughly as is.”