It's interesting to think about an extension of this system that allows for decentralized sign-ups. For example, instead of each new member needing to make a transaction with a central authority, existing members could spend or stake a certain amount of their own reputation to invite another person onto the network. You'd end up with an invitation tree, similar to the Lobsters site, where it might even be possible to propagate reputations back to the person who invited them.
Agreed, we have considered a system called "proof of trust" that is invite-based, and we're currently working on a decentralized group chat system on Freenet that will be invite based. I think you really need to offer a combination of approaches since invites work well in some contexts but not others.
I see a problem with the "feedback" and how fast it will get it's value. Creating a 100$-ghostkey and using it to send 1mil emails to a lot of people who accept emails marked with at least 10$-ghostkeys would allow the emails to go through as long as the "feedback"-key is "high" enough in reputation. The feedback will not build/decline fast enough (depending on the initial value of rep and the steps in decline and build-up) and therefore will not be good enough to fight the loss of 100$ for the initial ghostkey. If I'm missing something, please educate me.
That's a good point - however one advantage of Freenet is that the inbox doesn't just check the reputation of the ghost key once - if the reputation drops rapidly after the email gets into your inbox it will be removed, to stay in your inbox its *current* reputation is what matters, not just the reputation at the time it was sent. Additional safeguards could be implemented that would limit the rate at which emails could be sent under a particular ghost key which would also thwart this type of attack.
if an email is removed after the fact when reputation drops, would that mean I've already seen the potentially bad email? What if I read an email and want to keep it but reputation drops, will it be removed?
@@nightshade427 It depends on how the email system is implemented, but typically as soon as you read an email it will be transferred to your email client so it can't disappear. There will only be a short period of time during which the email might temporarily be in your inbox before the ghost key reputation is tarnished.
The action described (auto-gen key pair from encoded-keys) ending-ish at 7:12 ... Can this action be ran locally even if only with a hash manipulation formula to generate a potential expected result or an acceptable hash result (even if an empty result is given when opened)?
What if, to mess with me, a bunch of people decide to provide negative feedback on my Ghostkey. Don't we need a way to ensure that this feedback is real and not malicious?
Similar to eBay, people will only be able to give feedback on your ghostkey if they've transacted with you in some way - random people can't provide feedback.
So...if this takes over...do we loose the free as in free beer internet? It sounds like you will need to buy into the internet? Am I missing the point? Could this not also be used to trace a user back to a donation in a way that a TLA could demand credit card transactions and identify the person?
You won't need a ghost key to use Freenet, it's just an option available to people building services on Freenet if they need to mitigate spam or other types of abuse.
1) if ghost keys are actually decentralized, i should be able to start up a service to generate them myself, right? or are they generated centralized on your server. 2) using these ghost keys sounds like opting-in to "3rd party tracking cookies". a fresh key might not link someone's identity, but unless they are blinded again on every use, can't they be used to build a profile on you?
1) Ghost key creation is centralized but anonymous, verification and other uses are all decentralized. 2) You are correct that if you used the same ghost key on a bunch of different services then in theory someone could tie these different usages together, although they wouldn't be able to tie them back to your donation. In principle a zero-knowledge proof algorithm could be used to prevent this and we may support this in the future.
@@IanClarkeSanity Any centralization of a social credit score type system is a terrible idea. No thanks! I do not see the technical requirement for centralized registration or for money to be involved at all. We have spent the last 10 years figuring out alternatives to currency based trust systems and none of them rely on centralization. Frankly, the worst of the spammers have a budget and the premise that increasing the cost of spamming would eliminate the business model is entirely untrue, it may even worsen the issue because there would be less competition for those who could pull it off and users would be more likely to read the emails that do get through making it an even more lucrative model. It may price out some but it would equally incentivise others. At the end of the day, no matter your reputation I may view you differently than someone else and no amount of positive or negative interactions with others will determine what your reputation is to me. Threat actors love to game these types of systems by building up networks of high reputation accounts over years and using them to socially engineer people i.e. xz and wikipedia. It is concerning to me that you'd even suggest centralization without any technical requirement on something as sensitive as a social credit scoring system.
Interesting project but I think you may face the same issue as PGP did which is adoption. Also would be nice to have some code examples regarding integration. How do I fetch a public key of someone I know (that also have a GhostKey)?
So if i understand this properly this is essentially tor + blockchain so i'd hope that something like monero can be used to maintain anonymity. I see these keys as having the same issue with theft though that many bitcoin wallets have and if someone was able to either steal your key or spoof it, there would be no way to fix the bad reputation that you could of potentially received or that bad actors could use the method for recovery. Additionally i feel like theres nothing stopping goverment actors from utilizing the payment model to tip the scales in their favor if this got enough popularity.
That’s not possible with Ghost Keys. Only people to whom you’ve given a feedback token can affect your reputation, whether positively or negatively. So a botnet wouldn't be able to manipulate your reputation in this way.
I think the network of trust would mean that this concentrated bubble of trusting bots would not be generally trusted by others due to lack a close connection with the rest of the global trust network. Only a guess though.
Essentially someone could but it drives up the resource requirements so that botnet attacks are less likely. Essentially what he said is that if someone wants to pay a bunch of money they could but they couldn't do it as easy as people do now. Every system is vulnerable to a botnet with high enough complexity just as no system is 100% secure so long as its on the internet.
It's similar in intent, but with key differences: Ghost Keys are anonymous, decentralized, and general-purpose. Unlike something like Twitter verification, Ghost Keys aren't tied to any single platform-they can be used across many systems, offering a flexible solution for verification without sacrificing privacy.
When an email can generate millions in revenue, spending a few thousands to keep it alive long enough doesnt seem like muchbof a deterrent. The fee will just become the cost of doing business for bad actors.
@@IanClarkeSanity Big disagree. I don't believe most spammers have a single cent per email type budget. You're going to have to show the data that suggests that a $1 increase in budget would break spammers. Botfarms are extremely profitable businesses with clients that often have large budgets as well.
How convenient ! The only way to use freenet is to make a donation to the freenet foundation lol, by the way if you're paying to get something in return that's not a donation but a purchase. Also, a credit card payment deletes your privacy since it contains your full name, you could atleast accept monero. I personally don't see a future where this project gets significant adoption if it requires a credit card payment. An alternative should be added, for example a cryptographic challenge like bitcoin, or something else that could help the network.
Thanks for your feedback. To clarify, Ghost Keys are not required to use Freenet-anyone can use the network freely. Some systems on Freenet may use Ghost Keys to mitigate spam and abuse, but this is optional and determined by those systems, not Freenet as a whole. Regarding donations, it’s important to note that providing a receipt or token-like a Ghost Key-doesn’t turn a donation into a purchase. Non-profits often give contributors a receipt, acknowledgment, or token of appreciation without changing the nature of the contribution. Ghost Keys function similarly-they're cryptographic receipts that acknowledge your donation while preserving your anonymity via the blind signature process. Although the credit card transaction isn’t anonymous, the Ghost Key ensures that your real-life identity isn’t tied to your keypair. As for alternative payment methods, we started with credit card payments to accommodate the largest number of potential users, but we’re open to exploring other options like cryptocurrency, including privacy-focused currencies, if there is enough demand.
@@IanClarkeSanity My message wasn't really about payment alternatives even if it would make more sense to be able to pay with privacy coins for this kind of project. What I was trying to say is that there should be no direct payment. Also if I'm giving you money to get something in return which I need to do something, that's exactly what I do when I pay for an online service. Maybe it's not a purchase in legal terms but basically that's what it is.
>a credit card payment deletes your privacy since it contains your full name Only if you give them your name. That is entirely up to you. Every time I am asked about "Name on Card" i just write some bogus stuff.
@@XDRosenheim you're threat model is different than others. Not wanting a company to snoop on you is one thing where wanting to prevent the government from snooping on you is another. Both should be considered. Payment is unnecessary anyway for this type of system.
So, a "decentralized" system that's predicated on a single central (proprietary) key authority, backed by a potentially unverified online purchase. Got it.
The authorization is centralized but anonymous - verification is entirely decentralized. All of the code is open source including that used by the key authority.
Ghost Keys and WebAuthn both use cryptographic keys for authentication, but Ghost Keys focus on anonymous, decentralized identity verification, allowing you to prove who you are without revealing your identity. In contrast, WebAuthn is designed for passwordless security in centralized systems, like websites, using hardware tokens or biometrics, but it doesn't prioritize anonymity in the same way. Ghost Keys are for privacy in decentralized networks, while WebAuthn secures access in centralized environments.
so a ghost keys is a universal identity on freenet plaform given to user by a central identity provider, but the identity once given can be used throughout the freenet platform? Centrally given but universally scoped? and webauthn is where each server/domain in a decentralized manner gives an identity to the user and that key is only good for that server/domain or wherever that server/domain specifies it can be used? Decentrally given but only valid within that scope?
@@nightshade427 yes, GKs are centrally allocated but anonymous - verification is entirely decentralized. I'm not that familiar with webauthn so can't comment on that in much detail.
@@nightshade427 Yes, ghost keys are authenticated centrally but anonymously, after that verification and use is entirely decentralized. I'm not that familiar with webauthn so can't really comment on that in any detail.
So Basically anyone who doesnt own a ghost key ... Doesnt have a reputation... Which basically makes him the same as a bot or a spammer cuz either they have negative reputation or non at all.... Which basically means that if freenet truly blooms into the decentralized network it is planned to be.... All people have to make a donation which is technically not a donation but a purchase.... Tf is wrong with people these days 😂
It's interesting to think about an extension of this system that allows for decentralized sign-ups. For example, instead of each new member needing to make a transaction with a central authority, existing members could spend or stake a certain amount of their own reputation to invite another person onto the network. You'd end up with an invitation tree, similar to the Lobsters site, where it might even be possible to propagate reputations back to the person who invited them.
I second this, it would be really useful for such a system
Agreed, we have considered a system called "proof of trust" that is invite-based, and we're currently working on a decentralized group chat system on Freenet that will be invite based. I think you really need to offer a combination of approaches since invites work well in some contexts but not others.
Thanks folks
I see a problem with the "feedback" and how fast it will get it's value. Creating a 100$-ghostkey and using it to send 1mil emails to a lot of people who accept emails marked with at least 10$-ghostkeys would allow the emails to go through as long as the "feedback"-key is "high" enough in reputation. The feedback will not build/decline fast enough (depending on the initial value of rep and the steps in decline and build-up) and therefore will not be good enough to fight the loss of 100$ for the initial ghostkey. If I'm missing something, please educate me.
That's a good point - however one advantage of Freenet is that the inbox doesn't just check the reputation of the ghost key once - if the reputation drops rapidly after the email gets into your inbox it will be removed, to stay in your inbox its *current* reputation is what matters, not just the reputation at the time it was sent.
Additional safeguards could be implemented that would limit the rate at which emails could be sent under a particular ghost key which would also thwart this type of attack.
if an email is removed after the fact when reputation drops, would that mean I've already seen the potentially bad email? What if I read an email and want to keep it but reputation drops, will it be removed?
@@nightshade427 It depends on how the email system is implemented, but typically as soon as you read an email it will be transferred to your email client so it can't disappear. There will only be a short period of time during which the email might temporarily be in your inbox before the ghost key reputation is tarnished.
@@nightshade427 don't know.
The action described (auto-gen key pair from encoded-keys) ending-ish at 7:12 ... Can this action be ran locally even if only with a hash manipulation formula to generate a potential expected result or an acceptable hash result (even if an empty result is given when opened)?
Sorry, I'm not following your question - are you referring to the unblinding process after the blinded public key has been signed by the server?
What if, to mess with me, a bunch of people decide to provide negative feedback on my Ghostkey. Don't we need a way to ensure that this feedback is real and not malicious?
Similar to eBay, people will only be able to give feedback on your ghostkey if they've transacted with you in some way - random people can't provide feedback.
@@IanClarkeSanity So the feedback key is unique for each transaction/person?
@@suchithsridhar Yes, exactly.
Could this theoretically integrate with decentralized ID as it is currently being developed across standardization bodies?
Potentially, Ghost Keys are a very flexible approach designed to be integrated with other systems.
Freenet needs a killer app.
So...if this takes over...do we loose the free as in free beer internet? It sounds like you will need to buy into the internet? Am I missing the point?
Could this not also be used to trace a user back to a donation in a way that a TLA could demand credit card transactions and identify the person?
You won't need a ghost key to use Freenet, it's just an option available to people building services on Freenet if they need to mitigate spam or other types of abuse.
1) if ghost keys are actually decentralized, i should be able to start up a service to generate them myself, right? or are they generated centralized on your server.
2) using these ghost keys sounds like opting-in to "3rd party tracking cookies". a fresh key might not link someone's identity, but unless they are blinded again on every use, can't they be used to build a profile on you?
1) Ghost key creation is centralized but anonymous, verification and other uses are all decentralized.
2) You are correct that if you used the same ghost key on a bunch of different services then in theory someone could tie these different usages together, although they wouldn't be able to tie them back to your donation. In principle a zero-knowledge proof algorithm could be used to prevent this and we may support this in the future.
@@IanClarkeSanity Any centralization of a social credit score type system is a terrible idea. No thanks! I do not see the technical requirement for centralized registration or for money to be involved at all. We have spent the last 10 years figuring out alternatives to currency based trust systems and none of them rely on centralization. Frankly, the worst of the spammers have a budget and the premise that increasing the cost of spamming would eliminate the business model is entirely untrue, it may even worsen the issue because there would be less competition for those who could pull it off and users would be more likely to read the emails that do get through making it an even more lucrative model. It may price out some but it would equally incentivise others.
At the end of the day, no matter your reputation I may view you differently than someone else and no amount of positive or negative interactions with others will determine what your reputation is to me. Threat actors love to game these types of systems by building up networks of high reputation accounts over years and using them to socially engineer people i.e. xz and wikipedia. It is concerning to me that you'd even suggest centralization without any technical requirement on something as sensitive as a social credit scoring system.
Interesting project but I think you may face the same issue as PGP did which is adoption. Also would be nice to have some code examples regarding integration. How do I fetch a public key of someone I know (that also have a GhostKey)?
So if i understand this properly this is essentially tor + blockchain so i'd hope that something like monero can be used to maintain anonymity. I see these keys as having the same issue with theft though that many bitcoin wallets have and if someone was able to either steal your key or spoof it, there would be no way to fix the bad reputation that you could of potentially received or that bad actors could use the method for recovery. Additionally i feel like theres nothing stopping goverment actors from utilizing the payment model to tip the scales in their favor if this got enough popularity.
What would happen if a botnet created positive feeback for all of it's nodes and was used to create negative feedback for targeted accounts?
That’s not possible with Ghost Keys. Only people to whom you’ve given a feedback token can affect your reputation, whether positively or negatively. So a botnet wouldn't be able to manipulate your reputation in this way.
@@IanClarkeSanityyou didn’t address the first question
I think the network of trust would mean that this concentrated bubble of trusting bots would not be generally trusted by others due to lack a close connection with the rest of the global trust network. Only a guess though.
Essentially someone could but it drives up the resource requirements so that botnet attacks are less likely. Essentially what he said is that if someone wants to pay a bunch of money they could but they couldn't do it as easy as people do now. Every system is vulnerable to a botnet with high enough complexity just as no system is 100% secure so long as its on the internet.
So it's like twitter verification?
It's similar in intent, but with key differences: Ghost Keys are anonymous, decentralized, and general-purpose. Unlike something like Twitter verification, Ghost Keys aren't tied to any single platform-they can be used across many systems, offering a flexible solution for verification without sacrificing privacy.
When an email can generate millions in revenue, spending a few thousands to keep it alive long enough doesnt seem like muchbof a deterrent.
The fee will just become the cost of doing business for bad actors.
The revenue per email sent is a tiny fraction of a cent for most spammers, even a $1 dollar ghostkey breaks that business model immediately.
@@IanClarkeSanity Big disagree. I don't believe most spammers have a single cent per email type budget. You're going to have to show the data that suggests that a $1 increase in budget would break spammers. Botfarms are extremely profitable businesses with clients that often have large budgets as well.
How convenient ! The only way to use freenet is to make a donation to the freenet foundation lol, by the way if you're paying to get something in return that's not a donation but a purchase.
Also, a credit card payment deletes your privacy since it contains your full name, you could atleast accept monero.
I personally don't see a future where this project gets significant adoption if it requires a credit card payment.
An alternative should be added, for example a cryptographic challenge like bitcoin, or something else that could help the network.
Thanks for your feedback. To clarify, Ghost Keys are not required to use Freenet-anyone can use the network freely. Some systems on Freenet may use Ghost Keys to mitigate spam and abuse, but this is optional and determined by those systems, not Freenet as a whole.
Regarding donations, it’s important to note that providing a receipt or token-like a Ghost Key-doesn’t turn a donation into a purchase. Non-profits often give contributors a receipt, acknowledgment, or token of appreciation without changing the nature of the contribution. Ghost Keys function similarly-they're cryptographic receipts that acknowledge your donation while preserving your anonymity via the blind signature process. Although the credit card transaction isn’t anonymous, the Ghost Key ensures that your real-life identity isn’t tied to your keypair.
As for alternative payment methods, we started with credit card payments to accommodate the largest number of potential users, but we’re open to exploring other options like cryptocurrency, including privacy-focused currencies, if there is enough demand.
@@IanClarkeSanity My message wasn't really about payment alternatives even if it would make more sense to be able to pay with privacy coins for this kind of project.
What I was trying to say is that there should be no direct payment.
Also if I'm giving you money to get something in return which I need to do something, that's exactly what I do when I pay for an online service. Maybe it's not a purchase in legal terms but basically that's what it is.
>a credit card payment deletes your privacy since it contains your full name
Only if you give them your name. That is entirely up to you. Every time I am asked about "Name on Card" i just write some bogus stuff.
@@XDRosenheim you're threat model is different than others. Not wanting a company to snoop on you is one thing where wanting to prevent the government from snooping on you is another. Both should be considered. Payment is unnecessary anyway for this type of system.
So, a "decentralized" system that's predicated on a single central (proprietary) key authority, backed by a potentially unverified online purchase. Got it.
The authorization is centralized but anonymous - verification is entirely decentralized. All of the code is open source including that used by the key authority.
@@IanClarkeSanity please explain the technical requirement for the centralization?
sorta like webauthn?
Ghost Keys and WebAuthn both use cryptographic keys for authentication, but Ghost Keys focus on anonymous, decentralized identity verification, allowing you to prove who you are without revealing your identity. In contrast, WebAuthn is designed for passwordless security in centralized systems, like websites, using hardware tokens or biometrics, but it doesn't prioritize anonymity in the same way. Ghost Keys are for privacy in decentralized networks, while WebAuthn secures access in centralized environments.
so a ghost keys is a universal identity on freenet plaform given to user by a central identity provider, but the identity once given can be used throughout the freenet platform? Centrally given but universally scoped?
and webauthn is where each server/domain in a decentralized manner gives an identity to the user and that key is only good for that server/domain or wherever that server/domain specifies it can be used? Decentrally given but only valid within that scope?
@@nightshade427 yes, GKs are centrally allocated but anonymous - verification is entirely decentralized. I'm not that familiar with webauthn so can't comment on that in much detail.
@@nightshade427 Yes, ghost keys are authenticated centrally but anonymously, after that verification and use is entirely decentralized.
I'm not that familiar with webauthn so can't really comment on that in any detail.
So Basically anyone who doesnt own a ghost key ... Doesnt have a reputation... Which basically makes him the same as a bot or a spammer cuz either they have negative reputation or non at all.... Which basically means that if freenet truly blooms into the decentralized network it is planned to be.... All people have to make a donation which is technically not a donation but a purchase....
Tf is wrong with people these days 😂
Nostr
The Blockchain Commons is also doing interesting work like Gordian Envelopes.
How to corruption 101