A common vision promoted by crypto folks is that "we have to decentralize everything and put it on the blockchain", leading to weird memes like "web3". In reality, public blockchains are only suitable for very specific use cases when both of the following criteria are met:
- Peer-to-peer: software that anyone can run, and information is passed directly between them
- Strict global consensus: all individuals must agree on the exact same outcome
Therefore, blockchain applications are a subset of "web2", not something new, nor something that can or should "replace" "web2".
First, not everything needs to be decentralized. Most services work best in a centralized fashion. Furthermore, running peer-to-peer is sufficient for decentralization in the vast majority of cases where decentralization is indeed beneficial. In other cases, local consensus is all you need; global consensus is not only unnecessary, but wasteful.
An example where decentralization is very beneficial and can be achieved without a blockchain is the Alliance for Open Media. Background: A private company called MPEG-LA requires that all media content (music, video, etc.) on the Internet be licensed. While open source alternatives do exist, they have never been adopted due to lack of infrastructure. So a large consortium of companies and independent researchers collaborated to create an open standard for all. The result was the development of AV1 and AVIF (emerging open-source, royalty-free video compression formats)—great public goods. No global consensus is required, local consensus is sufficient for this federation, which is basically a real world decentralized organization.
So far, there are only two applications that need to meet both of the above mentioned criteria: currency and identity. Now, arguably, there are many currency and identity applications that only require local consensus, but this is already achieved through traditional systems - although they can be enhanced with technologies such as smart contracts. The appeal to public blockchains will become apparent when global consensus and access is truly achieved.
Of course, money can come in different forms - non-sovereign currencies, representations of real world assets, governance/ownership tokens, collectibles, loyalty rewards, social rewards, etc. Identity is fine too.
Combining money and identity could have some interesting applications.
In most cases, though, putting most things on the blockchain is wasteful. But there is also the possibility of a hybrid solution.
Let's take a "metaverse world" or just an open source massively multiplayer game as an example: you have a 3D real world that people play in. It's worth noting that this isn't a traditional game made by the likes of Blizzard or Bungie. In terms of game development in the traditional model, you only need local consensus. So, a consortium of game studios, engines, and service providers can build standards and run them. Portability outside of this environment is largely useless anyway.
So, we're looking at a whole new genre of game that's developed open source, with lots of forks, etc. Here, 3D content can be rendered, zero-knowledge proofs generated, physics engine calculations, etc., and most/all data can be done peer-to-peer; specific types of data use solutions such as IPFS. Anti-cheat software is all that is needed to ensure local consensus, without going through the blockchain.
However, each user can have a global identity, you can choose to grant governance rights to players, share revenue with developers, and there may be some elements in the game that require strict global consensus.
Let's take the hybrid model a step further: you have a centralized company that wants to run a fork of the aforementioned decentralized game, but still participate in the wider ecosystem.
So, the game is mostly done locally, in the cloud, or via peer-to-peer; there are elements that require a blockchain-type solution. Companies can run an application-specific volition where all data is held centrally, but specific higher value transactions and assets can choose to publish data in a committee or even on the L1 data layer, for full rollup guarantees - e.g., collectibles in the ERC-721 standard. Most critically, even if all data is kept centrally, the worst-case scenario is that assets are frozen, not seized or stolen.
This way you can have a hybrid solution that really maximizes the benefits of each component and smallest use case in the most efficient way.
Are entirely new experiences possible beyond currency and identity? Of course, I haven't seen any new experiences so far, and it's a waste to use a public blockchain for other things. As mentioned, what I can see is an innovative combination of the two that enables new use cases, even if the product is mostly centralized, or decentralized through non-blockchain means.
Weiliang
Miyuki
Weatherly