# Thoughts on DeSoc

I'm capturing my thoughts on [Decentralized Society: Finding Web3’s Soul](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4105763).

I can't help but consider privacy implications, though it says

> We initially assume publicity despite our deep interest in privacy because it is technically simpler to validate as a proof-of-concept

I'm also comparing it to [BrightID](https://www.brightid.org/whitepaper) and [BrightID soulbound tokens](https://github.com/BrightID/BrightID-Soulbound-NFT).

## Soulbound token provenance

> SBTs would enable us to trace the social provenance, giving us rich social context to the Soul that issued the work—their constellation of memberships, affiliations, credentials--and their social distance to the subject.

The authors criticize OpenSea as being too centralized, but OpenSea does offer a limited way to provide social proofs of creator authenticity. Traders can view other collections by the creator, and social links, and there is a centralized vetting process similar to Twitter's checkmark. I agree this could be improved with a decentralized process where creators can supply their own social proofs in the form of [Verifiable Credentials](https://www.w3.org/TR/vc-data-model/).

As far as provenance beyond the creator, this may provide additional information about a tokenized work of art's authenticity, but we need to allow for anonymous holders.

## Soul lending
> non-transferability prevents transferring or hiding outstanding loans,

The idea here seems to be to look at someone's (soulbound) credentials, and make a loan to the soul with those credentials. This seems like a better use-case for [Verifiable Credentials](https://www.w3.org/TR/vc-data-model/), where the subject can control how the credentials are shared rather than having them publicly viewable by everyone.

I agree that the current opaque system of assessing credit worthiness needs to be replaced.

## Recovery

> successful recovery nonetheless depends on curating and maintaining trusted relationships with a majority of guardians.

> A more robust solution is to tie Soul recovery to a Soul’s memberships across communities, not curating but instead drawing on a maximally broad set of real-time relationships for security.

I prefer the first approach of maintaining and curating your own set of trusted social relationships for recovery, but I welcome this experiment with a new kind of recovery (community recovery).

> Precise details to make this work will require experimentation. How guardians are chosen and how many guardians’ consent is required, for example, are key security parameters for further research.

I welcome it.

## DAO of souls

> checking for correlations between SBTs held by Souls who support a particular vote, and applying a lower vote weight to voters who are highly correlated.

This is potentially useful to get wider representation among voters with differing views. It has to be used carefully, though. If souls can choose which credentials to share (in order to increase the size of their [anonymity set](https://privacypatterns.org/patterns/Anonymity-set), and therefore their privacy), then this approach seems to do little to prevent collusion. Smart colluders could choose credentials that minimize their correlation.

> Ultimately, “markets” and “politics” are not separate design spaces;
SBTs can be a major part of a technological stack that enables the entire space between the
two categories to be explored

I completely agree and trying to treat them as separate has created plutocracy and other problems. I wrote an [article against plutocracy that also serves as a lead-in to attention streams](https://mirror.xyz/adamstallard.eth/pIy5KjNgjSTmI2M5qVipl9wYXz-0sgqkh-xU9KH6RnQ).

## Plural network goods

If the goal is also as quoted here

> favoring cooperation across differences simply means discounting cooperative rewards to similarly affiliated or correlated Souls-similarity measured by their shared SBTs.

then using SBTs to measure correlation is certainly interesting, but Souls might avoid correlation that decreases their voting power by choosing not to share certain credentials.

## Plural property

> Permissioning access to privately or publicly controlled resources (e.g., homes, cars, museums, parks, and virtual equivalents). Transferable NFTs fail to capture this use case well because often access rights are conditional and non-transferable: if I trust you to enter my backyard and use it as recreational space, that does not imply that I trust you to sub-license that permission to someone else.

Permissioning access to property seems more in the domain of OCAP (object capabilities), verifiable credentials, etc., that can more privately be issued to a DID ([decentralized identifier](https://www.w3.org/TR/did-core/)) subject. Verifiable credentials can be presented such that the verification is single use; i.e. the verifier can't credibly replay the presentation to someone else--the subject must authorize each verification. [Ethereum: when homomorphic encryption?](https://ethresear.ch/t/smart-contracts-from-fully-homomorphic-encryption/9465)

## Plural Sensemaking
This section holds the most promise. Plural sensemaking provides a strong reason to share ones SBTs (credentials), to be considered an expert to be included in future deliberative teams.

The critique of prediction markets is thought provoking. Predicting the future and uncovering ideas are best done as team sports. We need new ways for gathering the right deliberation teams that the rest of society can recognize as efficient and accurate.

I'd like to consider how [attention streams](https://mirror.xyz/adamstallard.eth/pIy5KjNgjSTmI2M5qVipl9wYXz-0sgqkh-xU9KH6RnQ) could be improved by adding provisions for discovering deliberation teams.

AI consuming large data feeds could certainly benefit from tailoring their output to different social groups based on SBTs. The participatory power of Souls involved could be increased as stated in the quote below.

> SBTs can also program bespoke governance rights to data creators, allowing them to form cooperatives that pool data and negotiate uses

## Privacy
> Rather than privacy-as-transferable-property-right, a more promising approach is to treat privacy as a programmable, loosely coupled bundle of rights to permission access, alter or profit from information.

It's true that in BrightID (and everwhere else), your private information is only as private as your counterparties allow it to be and privacy is social contract between pairs or groups of people.

There are ways, though, to share information such that it can be plausibly denied by the sharer if the recipient breaks their trust. I think this is a key to so-called self-sovereign identity, and one of the "rights to permission access to information" as phrased in the paper.

> Some SBTs may even permission access to data in a way where certain computations can be made, but the results cannot be proven to third parties.

Exactly.

> But VCs also have a key limitation: at least in their standardized form, VCs do not support most of the applications we have enumerated because of their unilateral privacy.

I don't understand this objection. Verifiable credentials don't preclude arrangements where multiple subjects need to authorize sharing, and there are recovery methods available.

## In relation to BrightID

> PoP protocols are limited to applications that treat all humans the same.

Agreed. I've known this which is why I created [attention streams](https://docs.google.com/document/d/1TKA-K8YadRdgz-Qek01TUcCkRaI9CKCXGtJ31AbVWIU/).

> An attacker can always recruit disinterested humans who are not yet participating to act as Sybils

Yes. I spoke about this in [Making and Breaking Plutocracy](https://mirror.xyz/adamstallard.eth/pIy5KjNgjSTmI2M5qVipl9wYXz-0sgqkh-xU9KH6RnQ) which is the intro post to attention streams.

## Conclusion

I like the concept of an SBT. I think soulbound tokens are useful, which is why we issued [some of the first ones](https://epor.io/tokens/0x8884c28d13648128a7191cb3c9af7a9e14fcbf71/8?network=xDai) and [created smart contracts](https://github.com/BrightID/BrightID-Soulbound-NFT) that others can use to create more. In our examples, "soulbound" means you can recover them (steal them back into a new address) using BrightID social recovery--this precludes selling them (even by selling a wallet private key). If metamask or other Ethereum wallets end up being our identity wallets, I think this is a good start.

Privacy comparable to that of an [identity wallet conforming to W3C standards](https://w3c-ccg.github.io/universal-wallet-interop-spec/) should be the goal, especially selectively sharing credentials as needed, instead of leaving them wide open for anyone to view or re-share.