Blockchain does not equal decentralization. It’s just a tool that can - with proper protocols - achieve that aim with specific scopes. Yet due to blocksize constraints, low throughput and latency, costly transaction trees for every operation, and the limitations on the actual amount of data that can be stored and processed onchain through technology like smart contracts, the effective applications of blockchains acting as decentralized software environments is exceedingly narrow. 

As an immutable ledger of who owns what token, such as in a peer-to-peer cash system (i.e Bitcoin), it is manageable and effective, but as a solution to wider real-world computer software utilities that centralization corrupts, it’s only a small piece of the puzzle.

As such, many popular use cases for blockchains are reliant on off-chain data and external services like oracles and indexers to deliver their function. There is nothing inherently wrong with that - except the fact it utterly betrays the principle of trustless, decentralized systems that are community owned, censorship resistant, and permissionless. 

Off-chain dependencies may be practical, but they do not result in decentralized services, they do not result in an Open Web. What’s wrong is that too much of the so-called blockchain ‘’industry’ (*shudder*) uses the appeal of those principles to draw investor and retail sentiment to their services. They market their protocols as upholding and expanding a new distributed internet, whilst their very existence undermines them. 

Small validator sets that centralize power, opaque or unfair economic distribution at inception, rent-seeking offchain dependencies, closed-sourced development, or entirely centralized hosting environments. It’s harder than ever for retail consumers to critically unpack these creeping forces of centralization that are piggybacking off the implicit potential of blockchain.

Validators and False Egalitarianism

Small validator sets are a massive and insurgent problem. In blockchain lore, the fabled 51% attack is almost a reality at birth for many Proof-of-Stake blockchains. If only 10 enterprise-users own 99% of a blockchain’s supply and those tokens confer governance rights or are used to validate the network, then those 10 actors can easily conspire to defraud it. Or, cloud server providers that host validator nodes can block them, as happened when Hetzner took 1000 validators offline and caused the blockchain to go down with it.

“VC-chains” like Solana are particularly vulnerable to this. The claim that such malfeasance would only destroy their holdings and so there is no incentive to conspire is scant consolation. If it’s possible at all, then at any future juncture it could become incentivized. An unforeseen attack vector - a highly efficient MEV-bot, a brand new flash-loan attack, anything at all - could be cause to take the chain ‘offline’ if the validators suddenly consider these activities to be acting against their interests. This is not the decentralized economic communities most people were - are - promised.

Ethereum now requires validators to hold 32 ETH, a large amount mainly attainable by large vested interests, and make new entrants difficult and centralizing power. Bitcoin has a problem with a few core mining groups with powerful specialized tech and unique energy-access dominating to an extent that, should even a few team up, they could break the network. These onchain oligarchies are to be feared, as they run counter to the cypherpunk ethos of democratically accessible and widely distributed pluralistic governance. 

Founders, insiders and early adopters often take outsized shares of token supplies. Public sales like ICO or IDOs are often laughably small amounts of the total token supply. Yet these chains then often purport to be DAOs wherein communities have an active say in the future of the protocol and that retail users are buying voting rights, when in truth it's a few whale-holders who are directing the fate of the chain, and always will be.

Off-Chain Dependencies

If occluded economic distribution is the soft misrepresentation of the blockchain industry, then offchain dependencies is the hard one - because it’s a lot harder for even technically-minded observers to spot the centralization, and take the fact a blockchain is being used means that there is distributed principles at play. 

Naming Services

Take the Ethereum Name Service, for example, an attempt at decentralizing domain registers, with Registrar contracts that govern those permissions. Firstly, economic distribution of voting power is an issue, and depressingly small groups hold outsized power even as the DAO expands (better than its original four of seven wallets deciding), but the more creeping problem is its string of off-chain dependencies. 

There is a reliance on IPFS metadata storage combined with third-party pinning services like Pinata for data availability. If a node that contains the data goes offline, the data becomes unavailable. If the pinning service chooses to act maliciously, it can censor the data. Of course, there is also centralized DNS for any Web2 domain integration. Although the resolver contracts may be onchain, the records pointed to are often not. 

In short, the hashes to certain data may be ‘fully’ decentralized, but the data itself is completely in thrall to traditional web architecture and its inherent points of failure. All of which, to repeat ourselves, is no great travesty - except it’s not a fully distributed system and, in current architecture, never will be.

NFTs

Or you could look at NFTs, once considered a form of true digital ownership that could revolutionize virtual economies. For what it’s worth, moonbois, a narrative that is almost certain to come back in force with attendant capital inflows - but we digress. Yet an NFT is just a receipt and a pointing mechanism to data hosted on very traditional services. If the data becomes unavailable, the NFT is just a worthless hash. The receipt may be maintained by distributed consensus, but the actual content is not.

Rollups

The issue becomes even more pernicious with rollups like Arbitrum designed to scale blockchain transaction effectiveness, bundling transactions, executing them offchain, then committing them to the mainnet. In these cases, centralized sequencers are responsible for ordering and batching transactions - often just single nodes. These nodes have all the power to censor and manipulate transactions to their advantage. What’s more, the fraud proofs that maintain transaction integrity are also operated by single validators. Rollups may make blockchain environments fast enough to handle some real-world utilities, but only by breaking the very principles underlying the value of these blockchains as they do.

Oracles

Many DeFi protocols are by now extremely reliant on oracles, particularly Chainlink, to the point that the ‘decentralized’ part of the finance is suspect, considering how much they need offchain data sources for accurate price feeds or to power their prediction markets. Oracles are, by their nature, an external off-chain trust layer - so we must be more circumspect with our criticism. There are consistent efforts to increase decentralization among oracles, yet the fact is oracles often rely on an excessively small active validator set for their data integrity, which introduces obvious risks as more and more DeFi protocols and prediction markets rely on one single oracle for their effective function. 

Add to that the interlinked nature of these protocols, and it doesn’t take a huge leap of imagination to see how one faulty or fraudulent price-feed upload could then trigger a cascade of smart contract operations that crash the entire market - such as Compound users being liquidated by a faulty DAI price feed. Furthermore, although Oracle’s like Chainlink aggregate data, the data sources themselves are all traditionally centralized operators (for example tracking the price of USD against GBP), merely shifting the centralization upstream.

Developer Centralization

Ultimately, core development teams often lead the direction of these new technologies, and those development teams take an outsized role in the direction of the chain. When Ethereum undertook the Merge, was that truly a democratic decision by a large community, or simply the decision of the Ethereum foundation and its core team of developers? There is a problem with crypto that begins with development teams and the power they wield to alter the protocol, especially newer chains with far less distributed voting power or collectivized cultural imprint.

If small teams of developers can alter the core function of the chain, what exactly are the participatory communities holding tokens for governance rights truly buying into, except disguised neo–startups company stocks masquerading behind the allure of cryptography? If developers are the only ones who understand the technicals anyway, then they are the gatekeepers to future progress. If the team also controls the voting power, then the so-called “immutable and eternal” protocol is really anything but.

An Antidote to Mania

Blockchains are fantastic tools - we use one - but they are only a small element of the march towards truly distributed systems. We’ve known since forever that blockchains do not scale well because they are unable to store and process large amounts of data. Usually viewed through the lens of transactions, we should view it through the lens of software utility in general. Blockchains are an anchor point for distributed systems, but they are not that in and of themselves. 

The key to a truly Open Web is not a blockchain, but the tools to decentralize data management and power local edge computing by having devices sync, query, process, store and execute utility functions without recourse to any centralized server, gateway or point of failure. A choir of computers not just maintaining a distributed ledger, but delivering real-world utility through community owned and managed software environments that truly uphold the potential trustless, permissionless, censorship-free and resilient end-user internet utilities. The operational substrate for truly distributed AI, cohesively functioning smart cities, resilient military systems, low-latency satellite arrays, and sovereign applications on end-user devices. That’s what we at Source Network are building. It’s not that we don’t love blockchain and crypto, it’s that it’s only the first step into a far wider world.