4 ms·

> And yes, I know that I don't need a blockchain to do good things, but I do need a blockchain to retire on having done good things. And more importantly, to li

> And yes, I know that I don't need a blockchain to do good things, but I do need a blockchain to retire on having done good things. And more importantly, to live in a world that's full of people doing the same. I don't think a case has been made for that yet. For example: > We can implement webs of trust on chain and incentivize trusted individuals to validate whatever relationships between the chain and the real world that the NFT's represent. This sounds amazing! Except it falls into a recursive problem. You need to trust the knowledge about the real world so you validate it but you also need to trust the validators so you need to validate them somehow and that reoccurs. You could argue that there is a social proof in terms of users vouching for one another a 'web of trust' but those are easily gamed. You exchange explicit centralization for implicit centralization so not only have the issues of a centralized institution anyway but all the extant issues we see in crypto land in terms of fraud and so on. It's also veering into dystopian territory in terms of keeping an immutable reputation score about people which is deeply disturbing.

__MatrixMan__ 5y ago
Are they easily gamed? I give my friends the code to my garage, nobody has gamed the web of trust and made off with my tools yet. If we want tight knit communities that we can trust, then we have to practice it. If we want do have the same benefits that you get from a trusted group of friends, but with thousands or millions of users, we'll need algorithmic help scaling it. Re: reputation, I agree that some kind of trustworthiness score is a dystopian outcome, provided that it's a single number that the network has universal consensus on. But I think it's a lot less onerous if the score is calculated via graph traversal. Don't ask "how trustworthy is user A," instead ask, "given users A and B, who does A trust that trusts B (and what kinds of trust have they configured)?" I'm not angling for a credit score, I'm angling to replace implicitly trusted institutions (like banks) with explicitly trusted third parties (like mutual friends, or people that users A and B both recognize as an expert).
- meheleventyone 5y ago
  > Are they easily gamed? Yes, like a Ponzi scheme or MLM for example. That's literally a web of trust setup to defraud people and one that often relies on pre-existing trusted relationships to spread. Almost all crimes involving social engineering like fraud are going to involve exploiting webs of trust. And whilst your friends may never steal your tools other people do end up being robbed by their friends. > Don't ask "how trustworthy is user A," instead ask, "given users A and B, who does A trust that trusts B (and what kinds of trust have they configured)?" I'm not angling for a credit score, I'm angling to replace implicitly trusted institutions (like banks) with explicitly trusted third parties (like mutual friends, or people that users A and B both recognize as an expert). Right and people won't create fake webs of trust with this to engage in scamming like they do in the real world already? No one would try to ruin your reputation this way and so forth. It feels like proponents of these social credit schemes do literally zero adversarial thinking. Similar to the way no one seemed to have thought about the rampant bot fraud now occurring on NFT marketplaces which are probably a scheme to setup false identities more than offload low value NFTs. But theft, fraud and automating it seem to not have occurred to the designers.
  - __MatrixMan__ 5y ago
    > No one would try to ruin your reputation this way and so forth The problems that you're describing only come up if there's a centralized oracle for a globally consistent reputation score. But the only angle where we need global consistency is re: who has what token. Reputation can be ad-hoc and only locally consistent. And since you're trading goods for NFT's you're probably co-located with all of the right people anyway. After all, an NFT for filling a pothole is worth more to the people who had to deal with that pothole than it is to people far away. If somebody is out to ruin your reputation, then they're unlikely to be involved in the transitive trust pathways that are used to validate tokens, so unless you're trying to do business with them I'm not sure why it would matter. Yeah, there's always the possibility that somebody you trust has decided to trust a bad actor, but someone's going to know about the deception because the alleged impact is nowhere to be found, so get together as a community and figure out who the liar is and revoke their trust.
    - 5y ago

__MatrixMan__ 5y ago

Are they easily gamed? I give my friends the code to my garage, nobody has gamed the web of trust and made off with my tools yet. If we want tight knit communities that we can trust, then we have to practice it. If we want do have the same benefits that you get from a trusted group of friends, but with thousands or millions of users, we'll need algorithmic help scaling it. Re: reputation, I agree that some kind of trustworthiness score is a dystopian outcome, provided that it's a single number that the network has universal consensus on. But I think it's a lot less onerous if the score is calculated via graph traversal. Don't ask "how trustworthy is user A," instead ask, "given users A and B, who does A trust that trusts B (and what kinds of trust have they configured)?" I'm not angling for a credit score, I'm angling to replace implicitly trusted institutions (like banks) with explicitly trusted third parties (like mutual friends, or people that users A and B both recognize as an expert).

meheleventyone 5y ago
> Are they easily gamed? Yes, like a Ponzi scheme or MLM for example. That's literally a web of trust setup to defraud people and one that often relies on pre-existing trusted relationships to spread. Almost all crimes involving social engineering like fraud are going to involve exploiting webs of trust. And whilst your friends may never steal your tools other people do end up being robbed by their friends. > Don't ask "how trustworthy is user A," instead ask, "given users A and B, who does A trust that trusts B (and what kinds of trust have they configured)?" I'm not angling for a credit score, I'm angling to replace implicitly trusted institutions (like banks) with explicitly trusted third parties (like mutual friends, or people that users A and B both recognize as an expert). Right and people won't create fake webs of trust with this to engage in scamming like they do in the real world already? No one would try to ruin your reputation this way and so forth. It feels like proponents of these social credit schemes do literally zero adversarial thinking. Similar to the way no one seemed to have thought about the rampant bot fraud now occurring on NFT marketplaces which are probably a scheme to setup false identities more than offload low value NFTs. But theft, fraud and automating it seem to not have occurred to the designers.
- __MatrixMan__ 5y ago
  > No one would try to ruin your reputation this way and so forth The problems that you're describing only come up if there's a centralized oracle for a globally consistent reputation score. But the only angle where we need global consistency is re: who has what token. Reputation can be ad-hoc and only locally consistent. And since you're trading goods for NFT's you're probably co-located with all of the right people anyway. After all, an NFT for filling a pothole is worth more to the people who had to deal with that pothole than it is to people far away. If somebody is out to ruin your reputation, then they're unlikely to be involved in the transitive trust pathways that are used to validate tokens, so unless you're trying to do business with them I'm not sure why it would matter. Yeah, there's always the possibility that somebody you trust has decided to trust a bad actor, but someone's going to know about the deception because the alleged impact is nowhere to be found, so get together as a community and figure out who the liar is and revoke their trust.
  - 5y ago