4 ms·

> Well that's a shitty thing to do, and a limitation of the technology, but it doesn't make all uses of that technology Yes, it's a shitty thing to do. And yes

> Well that's a shitty thing to do, and a limitation of the technology, but it doesn't make all uses of that technology Yes, it's a shitty thing to do. And yes, it's a limitation of technology. I never said anything about "all uses". All I asked is, quote: How do handle domain squatters and legal disputes? > It's the case where no money changes hands that I'm interested in. If we have technology that can function without custodial trust, then I can pay my subscription fees by helping host the site that I'm subscribing to. So, no money is changing hands but at the same time you "can pay your subscription fees". So money is changing hands. > I wanna turn everybody's pile of phones with bad batteries or busted screens into a distributed datacenter. And to do that you: a) find enough phones with bad batteries and busted screens because there aren't that many: people tend to repair them or change them b) motivate people to run your software on them So with b), people are going to run your software just because they, ahem, trust you?

__MatrixMan__ 4y ago
And my answer is that you don't. If you want content moderation (via a judge upholding the law, or a registrar admin who would prevent squatting) then use a stack that respects those authorities. If you chose to participate in one that doesn't, then you don't get those protections. Engineering is full of trade-offs. -- Well if they're paying (in terms of the use of their hardware, electricity, and internet connection) to keep the distributed service running so that they can use it, then presumably they trust its author (and its dependencies) to some degree, right? Otherwise they wouldn't want to use it. Trust is a necessary ingredient to most cool things, and as long as it's explicit trust I got no problem with it. People trust that folding@home or bitcoin mining is not going to open their network up to attack or some other malicious thing, so I don't see how this would be any different. The question is not do they trust me (an author of a dependency of the service that they want to use). That happens all the time. The question is can that service's author offload "hosting" to the users without requiring that they trust each other? For certain very restrictive applications, the answer is clearly yes. IPFS users host each others' content all the time. But can "hosting" here be safely expanded to a wider and more useful set of actions? Because if it is, I think that at the very least we could all deal with less downtime, pay less for remote compute, and see fewer ads.
- dmitriid 4y ago
  > And my answer is that you don't. If you want content moderation DNS isn't content. And protection against scammers, squatters etc. isn't moderation. It's funny how even the most primitive societies had rules, and laws, and ways to upholding them. But the crypto utopia where "NFTs make sense for DNS" is invariably "you don't get these protections" and "engineering is full of trade offs" > Trust is a necessary ingredient to most cool things, and as long as it's explicit trust I got no problem with it. If there's trust you have exactly zero need for the crap that is blockchains etc. Because our society is built on trust. > For certain very restrictive applications, the answer is clearly yes. IPFS users host each others' content all the time. IPFS is really, really small. But when police comes knocking at your door for storing someone else's child porn, you will quickly learn that blindly storing someone else's content will not help you. > But can "hosting" here be safely expanded to a wider and more useful set of actions? Of course not. Because you keep assuming that trust is a technical/engineering problem. It's not.
  - __MatrixMan__ 4y ago
    If I used to trust my friend, and I gave him key to my house, and later I decide that I don't trust him anymore, I change the locks: a technology solution to a trust problem. If people want to revoke trust in the bank or the law or AWS, they're probably going to need technology to shore up the gaps made by the loss of trust. I'm not assuming that you can derive trust from technology alone, but I do think that you can use it to get more out of whatever trust you have.
    - dmitriid 4y ago
      > If people want to revoke trust in the bank or the law or AWS, they're probably going to need technology to shore up the gaps made by the loss of trust. So. Back to the question about NFTs and DNS, which is literally about trust: How do you handle domain squatters and legal disputes? Oh wait. I know the answer to that: "Technology limitation, you don't get protections" and something about "engineering trade-offs".

__MatrixMan__ 4y ago

And my answer is that you don't. If you want content moderation (via a judge upholding the law, or a registrar admin who would prevent squatting) then use a stack that respects those authorities. If you chose to participate in one that doesn't, then you don't get those protections. Engineering is full of trade-offs. -- Well if they're paying (in terms of the use of their hardware, electricity, and internet connection) to keep the distributed service running so that they can use it, then presumably they trust its author (and its dependencies) to some degree, right? Otherwise they wouldn't want to use it. Trust is a necessary ingredient to most cool things, and as long as it's explicit trust I got no problem with it. People trust that folding@home or bitcoin mining is not going to open their network up to attack or some other malicious thing, so I don't see how this would be any different. The question is not do they trust me (an author of a dependency of the service that they want to use). That happens all the time. The question is can that service's author offload "hosting" to the users without requiring that they trust each other? For certain very restrictive applications, the answer is clearly yes. IPFS users host each others' content all the time. But can "hosting" here be safely expanded to a wider and more useful set of actions? Because if it is, I think that at the very least we could all deal with less downtime, pay less for remote compute, and see fewer ads.

dmitriid 4y ago
> And my answer is that you don't. If you want content moderation DNS isn't content. And protection against scammers, squatters etc. isn't moderation. It's funny how even the most primitive societies had rules, and laws, and ways to upholding them. But the crypto utopia where "NFTs make sense for DNS" is invariably "you don't get these protections" and "engineering is full of trade offs" > Trust is a necessary ingredient to most cool things, and as long as it's explicit trust I got no problem with it. If there's trust you have exactly zero need for the crap that is blockchains etc. Because our society is built on trust. > For certain very restrictive applications, the answer is clearly yes. IPFS users host each others' content all the time. IPFS is really, really small. But when police comes knocking at your door for storing someone else's child porn, you will quickly learn that blindly storing someone else's content will not help you. > But can "hosting" here be safely expanded to a wider and more useful set of actions? Of course not. Because you keep assuming that trust is a technical/engineering problem. It's not.
- __MatrixMan__ 4y ago
  If I used to trust my friend, and I gave him key to my house, and later I decide that I don't trust him anymore, I change the locks: a technology solution to a trust problem. If people want to revoke trust in the bank or the law or AWS, they're probably going to need technology to shore up the gaps made by the loss of trust. I'm not assuming that you can derive trust from technology alone, but I do think that you can use it to get more out of whatever trust you have.
  - dmitriid 4y ago
    > If people want to revoke trust in the bank or the law or AWS, they're probably going to need technology to shore up the gaps made by the loss of trust. So. Back to the question about NFTs and DNS, which is literally about trust: How do you handle domain squatters and legal disputes? Oh wait. I know the answer to that: "Technology limitation, you don't get protections" and something about "engineering trade-offs".