Second and third impressions on Web3

There is a heated discussion among the crypto community about Signal’s Founder Moxie Marlinspike “My first impressions of web3” article. Conversation has been surging, especially on Reddit.

Interestingly, Moxie just resigned from Signal a few hours ago.

Moxie is definitely right on highlighting the “commitment issue” on NFTs — the on-chain ERC721 or ERC1155 should always include an hash/commitment to the off-chain stored data of the NFT, otherwise you can continuously change the NFT details in a centralized flavour without any consensus mechanism. His points on decentralized wallets (Metamask) relying on centralized parties (OpenSea) to determine if you really have an NFT are on point as well.

There are still some major issues in the article… maybe because it’s “first impressions of web3”? Moxie is an outstanding computer scientist and cryptographer, so I would definitely be curious to read his second and third impressions on Web3, after he ventures deeper :)

• The commonly shared thesis on Web1 and Web2 is not actually “Web1 was decentralized, Web2 is centralized, Web3 will be decentralized again”. The Web2 label became popular at the beginning of the century to distinguish from the read-only and informative Web1; represented by the days of the very first 1994 Yahoo portal and the birth of ecommerces, versus the new social web consisting of blogospheres, social networks and messaging. TLDR: Both Web1 and Web2 were centralized. To summarise “Web1 is Information, Web2 is People, Web3 is Money”, where basically Web3 is the revolution that adds programmable finance to information streams (Web1) and reputation/people (Web2)
• Saying that decentralized technologies evolves lower than centralized ones is a faulty oversimplifcaiton. Using Vitalik’s lexicon from his popular “The Meaning of Decentralization” — which outlines the differences between “Architectural”, “Political” and “Logical” (de)centralization, the pace of change of a project depends on its political decentralization. A great example would be Uniswap changing faster than traditional IT BigTech companies, showing the difference between efficient political decentralization and inefficient political centralization
• Stating that decentralization is only for servers is simply untrue. The entire blockchain space is about going beyond the traditional client-server architecture. We could mention dozens of examples. IPFS is a good one, being a network of nodes (clients or servers? Well, both at the same time! Maybe clients for some aspects/services, and servers for others) that work together to offer decentralized data storage
• While you cannot run a full-node mining instance of Ethereum on an iPhone, that doesn’t mean the only alternative is to “interact with the blockchain via a node that’s running remotely on a server somewhere”. Light nodes can perfectly verify blocks even on an iPhone. Not being able to mine on any low-end device doesn’t mean that it’s not really decentralized
• Moxie also completely ignores the focal point of trusted vs trustless systems . Ok, sometimes we rely on centralized server providers (Infura) to relay our transactions. But the point is that in the centralized space the service provider can do anything on its own with our money, even without our approval (trusted). In the case of service providers like Infura, it can’t do anything without our signed transaction. It’s pretty banal, I know, but Moxie evades this point in his article

To summarise, I’m really curious to read Moxie’s subsequent views on Web3 as he delves deeper into the space. Looking forward to it!