4 ms·

I was looking for this recently but as a Web3 service (so emailing other Ethereum addresses) and found one that let you host an email account but you had to sta

by exdsq 4y ago

I was looking for this recently but as a Web3 service (so emailing other Ethereum addresses) and found one that let you host an email account but you had to stake X amount of money to use it to avoid spam (spam gets too expensive /awkward to do when you have to hold $100 for a week before you can use it). Didn't sign up but seemed promising that there could be some cool self-hosting options in the future that somewhat mitigate spam and avoids the whole current email issue. Obvious giant caveat that you can only email other people who have Ethereum wallets but still, thought it was pretty cool to avoid the blacklist problem.

dane-pgp 4y ago
I think this could actually be a viable system for stopping spam if it was retrofitted onto the existing email system. Rather than each account putting up the bond, though, it should be done at the level of domain names, and (to encourage adoption) all existing mail-sending domains should be grandfathered in (allowing them to use a $0 bond instead). The only question is how do you stop spammers walking away with their stake and creating another domain? There needs to be a way to slash someone's bond, which requires some sort of consensus. With DKIM signatures, it's fortunately possible to cryptographically prove that a given email was sent by a given domain, but ultimately you need to give potential censorship powers to some entity. My suggestion for doing that is getting the ITU to vote on a set of maybe 7 organisations (e.g. mail providers, universities, non-profits) who share spam reports with each other and can slash a bond if at least 5 of them agree. Of course individual mail providers would be able to override these decisions and continue to accept emails from the blacklisted domains (and they could obviously continue to use other forms of spam filtering), but ideally the bond slashing mechanism would only end up being used once as spammers tested it and then gave up.
- exdsq 4y ago
  > The only question is how do you stop spammers walking away with their stake and creating another domain? This is why I think it makes sense for an email-by-email basis if you also enforce a deposit and withdrawal delay. It's the simplest solution that I think could work. Not perfectly, but who's willing to spend $100 per email account with a 2-week delay between addresses? Add a public blacklist where you can vote for people as spammers and implement on your email services if you so choose, and I think you have a pretty good system. You may well be right that there's an elegant domain-based system with slashing and consensus but I haven't thought about it long enough to think of one myself! And yes you could do this off of Web3 but then you'd need an escrow account and a centralised party to hold the funds which isn't as decentralised as I was hoping for (at least with Web3 the stake remains in my wallet), but definitely possible! Edit: Maybe you could do it at an account level - then if you get blacklisted you'd have to open a new wallet meaning not only the 2-week delay and $100 cost but an additional network fee. That means it's unsustainable in the longterm.
  - dane-pgp 4y ago
    > And yes you could do this off of Web3 but then you'd need an escrow account and a centralised party to hold the funds which isn't as decentralised as I was hoping for Right. In fact I was already thinking that this was a (rare) situation where a smart contract on a blockchain would make sense, because you want transparency and consensus and low-throughput financial transactions between countries that don't necessarily trust each other (or even have banking connections to each other). I wouldn't necessarily call that Web3, especially as there wouldn't need to be any web servers involved, but I suppose the system would affect people's webmail, so I can't really object to the label.

by exdsq 4y ago

dane-pgp 4y ago

I think this could actually be a viable system for stopping spam if it was retrofitted onto the existing email system. Rather than each account putting up the bond, though, it should be done at the level of domain names, and (to encourage adoption) all existing mail-sending domains should be grandfathered in (allowing them to use a $0 bond instead). The only question is how do you stop spammers walking away with their stake and creating another domain? There needs to be a way to slash someone's bond, which requires some sort of consensus. With DKIM signatures, it's fortunately possible to cryptographically prove that a given email was sent by a given domain, but ultimately you need to give potential censorship powers to some entity. My suggestion for doing that is getting the ITU to vote on a set of maybe 7 organisations (e.g. mail providers, universities, non-profits) who share spam reports with each other and can slash a bond if at least 5 of them agree. Of course individual mail providers would be able to override these decisions and continue to accept emails from the blacklisted domains (and they could obviously continue to use other forms of spam filtering), but ideally the bond slashing mechanism would only end up being used once as spammers tested it and then gave up.

exdsq 4y ago
> The only question is how do you stop spammers walking away with their stake and creating another domain? This is why I think it makes sense for an email-by-email basis if you also enforce a deposit and withdrawal delay. It's the simplest solution that I think could work. Not perfectly, but who's willing to spend $100 per email account with a 2-week delay between addresses? Add a public blacklist where you can vote for people as spammers and implement on your email services if you so choose, and I think you have a pretty good system. You may well be right that there's an elegant domain-based system with slashing and consensus but I haven't thought about it long enough to think of one myself! And yes you could do this off of Web3 but then you'd need an escrow account and a centralised party to hold the funds which isn't as decentralised as I was hoping for (at least with Web3 the stake remains in my wallet), but definitely possible! Edit: Maybe you could do it at an account level - then if you get blacklisted you'd have to open a new wallet meaning not only the 2-week delay and $100 cost but an additional network fee. That means it's unsustainable in the longterm.
- dane-pgp 4y ago
  > And yes you could do this off of Web3 but then you'd need an escrow account and a centralised party to hold the funds which isn't as decentralised as I was hoping for Right. In fact I was already thinking that this was a (rare) situation where a smart contract on a blockchain would make sense, because you want transparency and consensus and low-throughput financial transactions between countries that don't necessarily trust each other (or even have banking connections to each other). I wouldn't necessarily call that Web3, especially as there wouldn't need to be any web servers involved, but I suppose the system would affect people's webmail, so I can't really object to the label.