Glen Weyl and Vitalik recently published a paper introducing what they're calling DeSoc – Decentralized Society. The paper tries to encourage the web 3 community to move beyond DeFi applications, painting a vision of blockchains as a programmable substrate for not just financial, but social value.

The technological primitive played with in the article is that of a "Soulbound NFT" – A non-transferable, non-fungible token. These tokens could be used to represent trust or credentials such as a university education. Combined with smart contracts, this would enable complex reputation logic and access to communities that cannot be bought or sold.

If you're a fan of self-verified and content-addressable data like me, you might ask yourself – why do these attestations have to be on-chain at all? Why can't you host them on IPFS?

When you have a blockchain hammer, everything looks like a token nail.

To be fair; Glen mentioned in a recent tweet that the point of the paper is the vision for what web 3 could look like as it grows out of its DeGen infancy. The "Soulbound NFT" is as much a narrative device as a fundamental building block. And I applaud them in their effort to steer the web 3 community in a more pro-social direction, regardless if they turn out to be successful or not.

However, mixing blockchains and social capital might lead us into thorny situations. Blockchain's core innovation is enforcing global consensus about the state of a database without a central authority. This is useful when building a monetary system. Trade would be impossible if people couldn't agree how much money everyone has, and assuming anyone is a self-interested asshole looking to backstab everyone else is a good idea when creating something that needs to be resilient to adverse attacks.

Not as much when it comes to sociality. Do we really want a global reputation score? Do you really want one globally enforced truth when it comes to credibility? Or does this smell social credit score to you too?

RadicalXChange are certainly against social credit scores, being a strong proponent of pluralism. Yet, in expanding the scope of blockchains beyond financial applications, they're pushing for a single, global, open and shared database of everything. Reputation and trust is context-dependent. People who are reputable within a certain scene are not automatically reputable in all scenes. By storing trust and reputation in this global and open database, the nuance is lost and subcultures are flattened. The more you condense and centralize reputation and trust, the more you end up with mediocrity devoid of that sought after pluralism.

In machine-learning terms, you're underfitting your model. And in the same way an underfitted model fails to pick up on nuance, an underfitted reputation system leads to mediocrity at best, and authoritarianism at worst.

I do think we're suffering from a crisis of legitimacy globally, and better reputation systems, broadly defined, is part of the solution. I am thus for the vision of decentralized society laid out by Glen and Vitalik. But it seems to me that in writing the paper, they started with a blockchain and retrofitted it to a use case.

Reputation is complex and multi-dimensional. Collapsing it into tokens that are storable on-chain might replicate the problems with our current reputation systems. An optimist may look at the phenomena and conclude that the renewed interest in reputation systems and webs of trust will lead to interesting experimentation taking place in forgotten areas, in the same way that DAOs have led to a renewed interest in alternative forms of governance. Or, web 3 will end up uprooting, simplifying and incentivizing social interactions until we've successfully wrapped the world in chains.

Time will tell.

Glen Weyl and Vitalik recently published a paper introducing what they're calling DeSoc – Decentralized Society. The paper tries to encourage the web 3 community to move beyond DeFi applications, painting a vision of blockchains as a programmable substrate for not just financial, but social value.

The technological primitive played with in the article is that of a "Soulbound NFT" – A non-transferable, non-fungible token. These tokens could be used to represent trust or credentials such as a university education. Combined with smart contracts, this would enable complex reputation logic and access to communities that cannot be bought or sold.

If you're a fan of self-verified and content-addressable data like me, you might ask yourself – why do these attestations have to be on-chain at all? Why can't you host them on IPFS?

When you have a blockchain hammer, everything looks like a token nail.

To be fair; Glen mentioned in a recent tweet that the point of the paper is the vision for what web 3 could look like as it grows out of its DeGen infancy. The "Soulbound NFT" is as much a narrative device as a fundamental building block. And I applaud them in their effort to steer the web 3 community in a more pro-social direction, regardless if they turn out to be successful or not.

However, mixing blockchains and social capital might lead us into thorny situations. Blockchain's core innovation is enforcing global consensus about the state of a database without a central authority. This is useful when building a monetary system. Trade would be impossible if people couldn't agree how much money everyone has, and assuming anyone is a self-interested asshole looking to backstab everyone else is a good idea when creating something that needs to be resilient to adverse attacks.

Not as much when it comes to sociality. Do we really want a global reputation score? Do you really want one globally enforced truth when it comes to credibility? Or does this smell social credit score to you too?

RadicalXChange are certainly against social credit scores, being a strong proponent of pluralism. Yet, in expanding the scope of blockchains beyond financial applications, they're pushing for a single, global, open and shared database of everything. Reputation and trust is context-dependent. People who are reputable within a certain scene are not automatically reputable in all scenes. By storing trust and reputation in this global and open database, the nuance is lost and subcultures are flattened. The more you condense and centralize reputation and trust, the more you end up with mediocrity devoid of that sought after pluralism.

In machine-learning terms, you're underfitting your model. And in the same way an underfitted model fails to pick up on nuance, an underfitted reputation system leads to mediocrity at best, and authoritarianism at worst.

I do think we're suffering from a crisis of legitimacy globally, and better reputation systems, broadly defined, is part of the solution. I am thus for the vision of decentralized society laid out by Glen and Vitalik. But it seems to me that in writing the paper, they started with a blockchain and retrofitted it to a use case.

Reputation is complex and multi-dimensional. Collapsing it into tokens that are storable on-chain might replicate the problems with our current reputation systems. An optimist may look at the phenomena and conclude that the renewed interest in reputation systems and webs of trust will lead to interesting experimentation taking place in forgotten areas, in the same way that DAOs have led to a renewed interest in alternative forms of governance. Or, web 3 will end up uprooting, simplifying and incentivizing social interactions until we've successfully wrapped the world in chains.

Time will tell.