Web3 has quickly become one of the most exciting yet controversial ideas to emerge from blockchain. While the general concept of a truly decentralized, user-owned web is appealing, there remains a lack of concrete detail about how the paradigm would work in practice. Critics have also pointed to inconsistencies around the current implementation of early Web3 projects.
A recent blog post by the cryptographer Moxie Marlinspike, who created the instant-messaging app Signal, highlights some of these contradictions with forceful clarity. Marlinspike raises serious technical questions about the nascent Web3 industry that aren’t easily answered by blockchain platforms and ecosystems. But several of his observations align with the design of the Internet Computer blockchain, which indicates how emerging technology can help Web3 become a reality.
Web3’s Problems
Web3 has attracted plenty of media attention because, as one recent overview put it, it promises “to reduce [our] dependency on large tech companies like YouTube, Netflix, and Amazon.” At a time when Big Tech has become something of a bogeyman, this is an outcome few could disparage. But this independence from tech giants is based on decentralization — of both governance, in the sense of users exercising certain voting and participatory rights, and networking technology.
In his blog post, Marlinspike highlights various compromising flaws around Web3. Users, he points out, continue to rely on centralized servers to access blockchain platforms and protocols. A typical Web3 app might use a decentralized blockchain to store certain data, but its users still need to rely on centralized platforms and/or servers in order to interact with the app. “Almost all dApps use either Infura or Alchemy in order to interact with the blockchain,” Marlinspike writes. “In fact, even when you connect a wallet like MetaMask to a dApp, and the dApp interacts with the blockchain via your wallet, MetaMask is just making calls to Infura!”
In other words, centralized middlemen are still required for the “client/server interface” — though the backend is on a blockchain, the rest of the stack is hosted on centralized infrastructure. This is similarly true of non-fungible tokens (NFTs) — one of the key elements of Web3 — which Marlinspike writes do not actually store their associated digital assets or media on a blockchain. NFTs simply “contain a URL that points to the data,” with a centralized platform or server hosting the data itself.
These issues demonstrate how many supposed Web3 projects are only marginally decentralized at best, with centralized servers and platforms responsible for maintaining the core infrastructure on which they operate. But while few Web3 applications actually operate entirely “on-chain,” in the sense of their frontends and backends both being deployed on a blockchain rather than just the backend, the Internet Computer offers end-to-end decentralization. This means that in addition to applications having a blockchain backend, their entire frontend and user interfaces also live on the same decentralized infrastructure.
While Marlinspike rightly notes that people “don’t want to run their own servers, and never will,” the Internet Computer obviates the need for application developers to deploy their own infrastructure while eliminating dependencies on centralized hosting services. One of the key features that distinguishes the Internet Computer from other layer-one blockchains is that its protocol runs on nodes that are distributed across a network of independent data centers, none of which are cloud providers. The nodes form different subnet blockchains that interoperate and run smart contract software that can deliver web content to browsers, providing an alternative IT stack.
The Internet Computer’s end-to-end decentralization also bears on Marlinspike’s complaint regarding NFTs, which generally point to digital assets that are stored on centralized platforms and servers. In his blog post, Marlinspike reports that an NFT he created was taken down on OpenSea, and that when he went to view it on his MetaMask wallet, it had disappeared. This is because MetaMask, like many other third-party wallets, relies on OpenSea — a centralized marketplace platform — to identify its NFT data. Once OpenSea removed the listing, the NFT seemed to vanish into thin air, even though it still existed on the Ethereum blockchain.
The Internet Computer avoids this problem by storing the entire digital objects and not just NFT metadata directly on its blockchain network. Even if a marketplace delists or removes a particular NFT’s digital object from its platform, the NFT remains on-chain, available for anyone to retrieve.
AllSeeingSeneca, the lead artist of Bored Ape Yacht Club, recently demonstrated this by minting her new collection on Ethereum while hosting it on the Internet Computer, ensuring that it runs on the blockchain. As Rolling Stone reported: “Hosting the NFTs on the Internet Computer should ensure that the NFT artwork lives forever on a public blockchain without threat of a takedown or cloud-outage issues.”
Marlinspike also notes that decentralized protocols tend to develop and evolve more slowly than centralized platforms. “If something is truly decentralized, it becomes very difficult to change, and often remains stuck in time,” he writes. “A sure recipe for success has been to take a 90’s protocol that was stuck in time, centralize it, and iterate quickly.” An ability to be updated more quickly, he suggests, could allow centralized platforms to consolidate control of the emerging Web3 space.
The Internet Computer’s governance is controlled by a decentralized autonomous organization (DAO) called the Network Nervous System. The NNS community adopts or rejects proposals to upgrade the protocol, avoiding hard forks. Hundreds of such proposals have already been adopted, enabling the Internet Computer Protocol to rapidly evolve. For example, the community adopted a proposal to begin work on integrating the Internet Computer with Bitcoin’s network last September, and full implementation is coming this year.
Marlinspike describes Ethereum’s failure to specify how royalties are handled with ERC-721 (non-fungible) tokens meaning that OpenSea has taken the lead and developed its own, unilateral way of dealing with them. The Internet Computer’s adaptability gives it a better chance of avoiding such a scenario.
The Future
Of course, there’s no telling how technology like the Internet Computer will contribute to the maturation of Web3, to say nothing of whether innovations can manage to resolve all of the problems the new paradigm will encounter as it evolves. The central question, which Marlinspike leaves unexplored, is conceptual rather than technical: What exactly should Web3 look like?
Answering this means going beyond the vaguely stated principle of ‘decentralization’ and creating effective applications that genuinely address the issues that Web3 purports to resolve. No doubt this will take time, and multiple rounds of trial and error, but in a world where Big Tech still wields far too much power, it’s a question that necessity will force us to answer sooner or later.