I’ve been interested in the blockchain space since high school and began doing more work in the space around two years ago. In this short time, the landscape of crypto has changed drastically — from the speculation of its value add to business operations and to the technology itself. As someone who’s gone through the whole cycle of being a believer and a skeptic, I wanted to articulate my current stance on Web3, which is this:
The ethos surrounding decentralization is plainly wrong.
Background
Blockchain technologies are built around three pillars: decentralization, transparency, and immutability. These are the same that gave birth to the idea of Web3, which is an internet paradigm that allows users to not only read and create but also own their data. I believe that all three of these pillars are fundamentally flawed as net negatives but for the most part, the buzz around decentralization is one of the biggest drivers of people’s initial interest, which is why I’m focusing on this. The issue with decentralization is two-fold: it’s difficult to ensure as a protocol and offers tradeoffs with overrated upside to users. I believe that this is a fundamental flaw for almost all Web3 technologies, which is why we see little mainstream adoption today.
Difficulty
First, true decentralization requires that no single entity can manipulate the system at any point in its end-to-end flow. This is a strong and extremely restrictive guarantee since a consequence of this is that a system cannot interact with non-decentralized components — anything not on the blockchain. The internet today, or Web2, took off because it enabled the rapid, global exchange of culture and ideas. It’s nearly impossible to imagine how any notion of a new internet could get started as a silo, without being integrated with the data we use daily.
Even if blockchain does have utility without off-chain data, there still exist decentralization issues within its workflow. Any network that sells itself on security is defined by its weakest link. As blockchain workflows have become more complex to accommodate smart contracts or improve scalability, it has given rise to more vulnerabilities that are at risk of contradicting the decentralization of the chain.
Maximum Extractable Value
One example of this is Maximum Extractable Value, or MEV, where miners can re-order the transactions in a block to make money off front-running. It takes advantage of a single point in the process where a miner has the sole discretion to decide which transactions to include in a block and can front-run or sandwich user transactions. To date, over a billion dollars have been generated by MEV strategies [1], with one user profiting $34M in just three months [2]. Yes, one miner cannot profitably corrupt the security of the chain, but the vulnerability of user transactions to such attacks contradicts the ethos of decentralization.
Layer 2 Scaling
A second, broader concern about centralization lies in the Layer 2 community. L2s are smaller chains designed to offload some transactions from an L1 (e.g. Ethereum) to improve scalability. These transactions are run on the L2 and their outputs are bridged back to the L1. By nature, L2s offer weaker security guarantees because they’re smaller in size, more fractured market-share wise, and rely on the L1 in exchange for lower gas costs.
In L2s, transactions are bundled into batches and executed, and only the final state is recorded on the L1. Because L2s are designed purely for scalability, instead of mining and bundling transactions into a block centralized entities called sequencers decide which transactions are included in a batch. As a result, they have sole discretion over MEV, censorship, and network availability. Note that the challenge with having decentralized sequencers, or allowing anyone to submit batches, is that batches that aren’t selected have wasted computation and electing a batch is less efficient.
In April of this year, over 82% of all Ethereum transactions were settled on L2s [3]. Given this volume that L2s take on today, these centralized, L2-controlling entities play a huge role in the network. They are principally no different from today’s internet conglomerates. This begs the question of why users are even willing to pay gas fees to use a system just as centralized as Web2 and as an existential implication, if decentralization is even a meaningful value add.
Effects on User Experience
This is the truth underlying any decentralized system, which is that it will have important tradeoffs compared to a centralized counterpart. Centralized systems by default provide a consensus contract that all users follow to get the current state. Without a centralized facilitator to decide on consensus, the users themselves interact to keep the system consistent. As a result, a single direction of communication must necessarily be replaced by multiple rounds of communication between nodes and possibly financial incentives for doing this work. This is a commonly studied tradeoff in distributed systems when decentralized systems are used to prevent single points of failure.
Hence, blockchains can compensate by either increasing user costs and latency or loosening decentralization and network security. Because the latter is the fundamental value add that the industry is built on, the user experience becomes at risk.
This is an ongoing back-and-forth that developers in the space have to grapple with. In fact, most of the issues I pointed out above can be mitigated by replacing a single point of failure with a quorum-based layer. For example, Chainlink offers access to off-chain data through its oracle network that aggregates the results of off-chain queries from its system of nodes. However, this comes at a rising cost in transaction latency and gas for the user.
Therefore, the big question when valuing the utility of the blockchain space circles back to how much users are willing to trade-off between cost, latency, and having the benefit of decentralization. Unfortunately, I believe this willingness is low.
Value Proposition (?)
From a financial use-case perspective, the banking system in the US is heavily regulated and consumers have legal protections against fraud. It’s far from perfect, but it’s something that the vast majority of Americans can and have reasonably relied on for their needs, making decentralization a limited value add. Furthermore, the promise of decentralization is not at all an immediate guarantee of security. With $19B lost to crypto hacks to date, there are legitimate concerns about storing funds on-chain. The transparent and immutable nature of blockchain is actually more susceptible to costly hacks due to exposed backend code.
From a social case, the idea of internet companies collecting user data sounds pervasive, but does it sound pervasive enough that users will be willing to pay for interactions with an alternative that also operates a little slower? I would bet on no. Yes, corporations may own and sell our data exploitatively, but our data isn’t being exposed to public APIs for anyone to access (which is a concern for on-chain activity). Additionally, we still reap some benefits of consumerism like recommended content or friends at no cost. As it is, I believe the current system is not threatened by an unproven alternative.
Potential Upside
With that said, I do think there are two general areas where decentralized applications can provide value. First, when the central entity is extremely unreliable or already expensive. Examples of this would be banking in regions with political instability and conducting cross-border payments. These applications are already popular in areas like South America today. I could see this being an important use case of crypto that remains indefinitely.
The second is in cases where the central entity is logistically challenging or risky to set up. For example, a company could release a token to immediately form a two-sided market between producers and consumers instead of raising capital to bootstrap initial adoption themselves. This is different from paying in cash because users can individually value the token based on their long/short belief in the product. I’m mainly thinking about decentralized physical infrastructure networks here (e.g. Helium, Hivemapper), but different use cases may exist based on these same principles.
Closing
In short, crypto and Web3 are inherently limited by the challenges of decentralization. In seven years, over $90 billion has been invested as venture capital into the industry, and there has been no significant, mainstream user application to lend credibility to claims that Web3 will ultimately replace our current notion of the internet. I do buy that crypto has the potential to become (or remain) a useful application in the two areas above, but I don’t think it will be anything more than just a tool, albeit an important one.
The bottom line is that when an alternative technology introduces cost and efficiency tradeoffs to consumers like Web3 does, its edge against the prior becomes narrowed. Therefore, and given how the space has unfolded to date, I think it’s unlikely that this paradigm will materialize.
Sources
[1] Forbes
[4] Coindesk
[5] The Block