Decentralisation, crypto, and you
Why reducing decentralisation could reduce risks in DeFi and Web3, and widen their appeal beyond early adopters.
There’s an inherent trade-off between decentralisation and the ability to coordinate things: the fewer participants, the easier it is to agree on anything. This is why democracies elect representatives to form a government instead of deciding everything in referendums: to have everyone agree on every detail in laws and budgets would be impossible.
Decentralised systems have no single entity in charge who could reverse mistakes, both in good and in bad. What you win in avoiding dictators, you lose in the ability to recover from catastrophic failures. This applies to both social and software systems. Successful decentralised algorithms tend to be simple and lightweight, as the simpler they are, the easier it is to prove the system functions correctly under all conditions. Take Border Gateway Protocol in IP routing or Apache ZooKeeper’s distributed synchronisation services. These algorithms display a form of emergent complexity, where individual nodes running simple algorithms form a complex and fault-tolerant distributed system. This is not dissimilar to market economies, where the simple actions of buyers and sellers form an infinitely complex global system of factories, shopping malls and supply chains linking it all together. Chaotic as they may seem, distributed systems are self-healing — up to a point. Since their networks have no single points of failure, distributed systems can remain operational even when individual nodes go down. However, they are susceptible to system-level errors, which are much harder to remedy since no single entity can take action to protect the system as a whole.
Everything said before applies to cryptocurrencies. The more decentralised a network becomes, the harder it is to agree on major changes. In 2016, the Ethereum blockchain was hard forked after the original “The DAO” was hacked for millions of ETH. Deemed too big to fail, the Ethereum mainnet was rolled back to bail out The DAO and recover the stolen assets, while the unforked chain became known as Ethereum Classic. Such a hard fork would not be possible today — the network is too decentralised, and getting everyone to agree on a fork over a single hack would be essentially impossible.
In 2016, the founders of Ethereum had the ability to correct catastrophic errors that could have triggered a system-wide collapse. Today, we are on our own. What are the implications of this?
Why Ethereum and Bitcoin are different
Bitcoin is a network protocol. The ownership of bitcoins, and transactions between the owners, is tracked simultaneously by nodes participating in a global peer-to-peer bitcoin network. The software running in each node is simple: verifying transactions and the hashes of newly mined blocks, playing a part in achieving consensus on blockchain state, thus helping to secure the network. Bitcoin also exhibits emergence, as simple nodes give rise to a secure decentralised currency you can use to order pizza.
Ethereum is more than that. It is a general-purpose computing platform. Ethereum’s smart contracts extend its functionality far beyond Bitcoin’s, from decentralised exchanges (DEX) to custom tokens tracking ownership of anything imaginable, from company shares to images of monkeys. This sophistication makes Ethereum a marvel of 21st century engineering, but it also makes the protocol more complex. And in the world of decentralised applications, complexity brings danger.
What are the risks? Every time you interact with a smart contract, you have to place trust on not just the Ethereum blockchain, but on the developer of the contract, as well as any software libraries they used in the development process. It is smart contracts where the biggest crypto hacks of today originate from. And since smart contracts are built to be callable by other smart contracts, a vulnerability in a low-level contract could bring down an ecosystem of DeFi apps like a collapsing Jenga tower.
I don’t think Ethereum and other smart contract blockchains can brush aside risks associated with smart contracts and immutable transactions. These networks might one day be critical infrastructure powering Web3 apps we all use. Bitcoin can get away with inherent risks of losing your coins by “only” being a protocol for digital money. Bitcoin is optional. Users should be aware of the risks. And even then, the Bitcoin codebase is relatively simple, stable, and battle-hardened. Anyone can audit the source code. While smart contracts published to the Ethereum blockchain are public, there’s so many of them that no one person can understand every contract that plays a part in the DeFi and Web3 ecosystems. And since contracts can be written and deployed by anyone, there’s no strict vetting process from the open source community to catch programming mistakes before its too late.
If Web3 truly takes off, our entire lives could be built on blockchains. They need to be protected, not by wishful thinking but by proactively planning recovery mechanisms for everything that could go wrong. Today’s blockchains weren’t designed for this. Here’s some ideas that could help.
On recovery mechanisms
For organisations, storing their digital assets in multi-signature wallets is a good start. These wallets require a quorum of trusted users to sign any withdrawals. As long as each key is stored separately, funds stored in the wallet are safe even if one or two keys were to be compromised by hackers. This also prevents the owners from being locked out of their assets should one of the keys get lost.
Even with multi-sig wallets, I am personally sceptical of critical systems that don’t provide an escape hatch for humans to interfere in case of errors. Not everyone is comfortable with managing their own wallet or being exposed to smart contract risk in DeFi. On the other hand, few people want their crypto locked into centralised companies’ vaults either: from Celsius to Anchor, centralised lending platforms seem to be no safer than decentralised smart contracts like The DAO were. Today, the decision is binary: either your keys or not.
Perhaps in the future we will see decentralised applications built on top of blockchains that provide some safety and recovery features for users who want them. An example of this could be a DAO where transactions above a certain amount need to be approved by a certain share of members, or a Proof of Stake network where transactions can be rolled back by an appeal process within a multi-day settlement period. The latter could strike a compromise between usability and security: should your wallet get hacked, you have 24 hours to appeal to real people until the transaction settles. It is perhaps surprising how the often derided T+n settlement periods of traditional credit cards and payment networks have real benefits and could one day reappear in the world of crypto!
One day, systems like these could bring similar levels of customer protection that exist in traditional finance and payment systems into crypto. Of course, there’s downsides as well: the network’s trustees have power to revert user transactions. Therefore it’s important that the identity of the trustees is transparent, and that they remain accountable through a governance structure such as a DAO, as well as through the conventional legal system. Adding a settlement process also introduces potential losses to merchants using the underlying tokens as currency, which means they might be better suited for use cases that demand less timely movement of assets, such as tracking the ownership of company shares.
I believe this kind of safety features could form a basis for mass-market adoption of crypto. While early adopters may feel comfortable regularly moving their life savings as tokens between lending pools to maximise yield, ordinary people will demand the same level of security and deposit insurance their High Street bank provides. None of that will be possible without some form of human intervention. Moving away from full decentralisation will not be for everyone, and crypto maximalists should keep using their networks as they are today. But for the rest of us, a little more protection could make DeFi feel less scary and more accommodating, and bring more people into a better, fairer financial system.
If you enjoyed the writing, do consider following me on Medium. Every blog post is a product of 10+ hours of writing and editing. If you want more, check out my thoughts on how the Creator Economy could revolutionise education, or a longer read on the benefits blockchain brings to interoperability.
The author is a software engineer working in crypto. All opinions expressed in this blog post are solely my own, and do not express the views or opinions of my employer.
New to trading? Try crypto trading bots or copy trading
