4 ms·

No, I said that reducing the attack surface to a subset of the problems you normally have is good and makes elections cheaper. That’s what cryptographic protoco

by EGreg 2y ago

No, I said that reducing the attack surface to a subset of the problems you normally have is good and makes elections cheaper. That’s what cryptographic protocols, including blockchain, do in general. They replace the need to trust corruptible middlemen, with a protocol that is infeasible or extremely hard to subvert, and which leaves traces of the subversion. Crypto is used all the time such as when you use cryptographic hashes to detect tampering, or merkle proofs to prove something was included correctly in a larger part. You then replied essentially: “well since you still have some problems, you can’t trust the election… the paper way is the only right way”. Some people might be wilfully misunderstanding because it’s “cool to rag on blockchain” or whatever. People who always repeat a refrain like “this is simply the only right way to do things” are trying to convince not by arguments but by pushing a dogma. And most skeptics of technologies have been wrong, including skeptics of airplanes, computers, etc. Estonia for example is already doing secure elections online for years, explain that https://e-estonia.com/how-did-estonia-carry-out-the-worlds-first-mostly-online-national-elections/ https://e-estonia.com/how-did-estonia-carry-out-the-worlds-f... The hand recount took too long and the Supreme Court stepped in and “just picked a winner”. Which later counts showed to have been the wrong result. Citing the machine counting alongside it doesnt really help your case because the machine counting was all kinds of ad-hoc and hybrid things (including the dreaded silly “butterfly ballots”) which is exactly what people advocate for, when they try to argue for avoiding a fully consistent and uniform electronic system. They want all the little variations and manual counting “so no one can hack the whole thing”. So yes it’s a perfectly valid argument to point out that delays caused by this led to the wrong outcome (and had consequences like ignoring Bin Laden, allowing 9/11, the invasion of Iraq, clamping down on civil liberties in USA, raiding Social Security etc.) All the problems you cited above are present in the current system — including having to prove how you voted to challenge the results. Except in the current system there are far more problems, including not even being able to physically show up at the polling place (because it is too far), or proving that the poll workers corrupted your vote, added extra ballots, literally anything. Out of sight out of mind I guess. And across the world, elections are done even worse. Consider the recent election of Lukashenko in Belarus. People in districts where he got 80% were trying to ask around who voted for him and complained that very few had said they did. It’s all arguments based on hearsay. That is the flip side of not being able to prove how you voted. In fact if they wanted to know how you voted, in your manual system, they could just take a camera outside the booth and look at timing to know when you voted. Or just put a camera in the booth. But in fact it’s far worse than that, the voter databases include driver’s licenses and addresses and social security numbers, in most US states, AND party affiliation is 94% correlated to how you vote so all this paper ballot “security theater” to prevent “being ABLE to prove how you voted” gets you nowhere: https://ballotpedia.org/Availability_of_state_voter_files https://ballotpedia.org/Availability_of_state_voter_files And oh yeah… in the system I described you can anonymously challenge the results because you have cryptographic signatures but your own private key came out of a mixer, so you don’t need to identify yourself to prove your vote didnt match what’s in the system. Enough complaints and we ALL know which districts were corrupt, and very quickly.

tsimionescu 2y ago
In paper based systems, you and other volunteers do most of the counting, alongside representatives of all parties. None of the problems I described exist in such a system: you can't add a million votes unless you convince a whole lot of volunteers and representatives of the parties that they are real votes. You can't put multiple votes in unless none of those same people see you. If one volunteer attempts to change a vote, another one will stop them. If you think your vote was miscounted, a re-count can be issued, with even more observers, and the exact same artifacts are available for all to see (how do you fix an error in the Merkle tree, even if everyone agrees it happened?). Even if you have an extremely corrupt county, that doesn't generally matter in the grand scheme of things; and its extremely unlikely, as any citizen in that county can stop the corruption by simply participating in the process themselves. Wide-scale voter fraud of this kind is simply impossible in a paper system. The only times it happens is like in Belarus, where it's not "an election", it's a public show that looks like an election, but where the result is pre-determined. The Merkle tree would show the same thing there: it's a mock election to make it look like a mock democracy. Lukashenko wouldn't have stopped leading the country even if miraculously the election would have shown he lost. Or, it can happen in other more complex and more discoverable ways, such as busing voters around to physically vote multiple times in multiple (preferably far away) polling places. As for Estonia, they'll come to regret this system sooner or later. It can work for a while, but there is no doubt that the system will get hacked, or the losing party will be able to convince enough people that it got hacked even if it didn't. Someone will accidentally publish private keys, like in this Colorado case. The system will go down on election day because of a bug. Who knows which one will be first, but it'll end their experiment. The rest of the world will continue with paper voting and not face such problems.
- EGreg 2y ago
  “In paper based systems, you and other volunteers” No, 99.99% of “you” go home and “trust the system” to some poll workers, many with major bias and incentives. Many of “you” don’t turn out to vote or are disenfranchised by simply living too far from the polling place or not being able to take time off work, when you could have just voted from your app. Certain parties even rely on suppressing turnout. (Can you guess which party does that in USA? Hint: it’s the one that closed 1600 polling stations right after the Voting Rights Act got neutered, and then got mad about mail-in ballots ruining their carefully laid disenfranchisement plans during the pandemic.). In fact, if you want the election to be “secured by multiple distrusting parties”, that is exactly what byzantine-fault-tolerant cryptographic protocols (which power many blockchains) are designed to do. (how do you fix an error in the Merkle tree, even if everyone agrees it happened?). Even if you have an extremely corrupt county, that doesn't generally matter in the grand scheme of things; and its extremely unlikely, as any citizen in that county can stop the corruption by simply participating in the process themselves. You are literally arguing from a double standard. In a paper election, somehow “any citizen” by themselves can stop the corruption… by simply participating in the process.” Yeah sure one guy exposes the entire corrupt county, with no ability to prove how anyone voted, why didn’t a single Belarussian and Venezuelan think of that? LOL” And on the other hand, when you have tons of anonymous irrefutable proofs by participants submitted publicly, you throw up your hands and say “what can we do to fix the merkle tree, even if we all knew it was corrupt?” The point of the trre is to catch errors, prove them and publish the proofs widely. As a society, you then have the proof nexessary to fix errors the same way you’d normally do it — by identifying the corrupt districts, and having a recount or revote just there. And bringing those responsible for tampering to justice. If you stop conflating all the things and unpack them, you’ll see that adding cryptography makes things strictly better: 1. You have more chances to catch if there have been extra votes cast because the private keys are coming from tokens handed out at registration. In a paper election you might have corruption at registration AND all manner of ballot stuffing later too. 2. Everyone can check their vote and report a discrepancy. Not just the volunteers at the polling places. And all because they can prove how they voted and do it anonymously! 3. Everyone can see exactly which districts are corrupt in giving out fake voter registrations, and where there’s smoke, there’s fire. They can do an audit and guess what, the cryptographic signatures are helpful for creating a PROVABLE trail that implicates the system. 4. The attack surface reduces to pretty much just the voter registration sybil attacks. Eliminating a whole class of problems on actual election day. 5. The results are reported to everyone reliably and quickly, or even in real-time (though the latter is “too good” because it might affect how later voters vote). There’s practically not a single problem that adding cryptography creates, which wasn’t already present in the paper system. And you know all this because if you honestly asked yourself whether dictators, who want sham elections, would want to do their next election with cryptographic signatures and merkle trees or not — what would be your answer? Be honest. And think about what that means for your argument.

by EGreg 2y ago

tsimionescu 2y ago

In paper based systems, you and other volunteers do most of the counting, alongside representatives of all parties. None of the problems I described exist in such a system: you can't add a million votes unless you convince a whole lot of volunteers and representatives of the parties that they are real votes. You can't put multiple votes in unless none of those same people see you. If one volunteer attempts to change a vote, another one will stop them. If you think your vote was miscounted, a re-count can be issued, with even more observers, and the exact same artifacts are available for all to see (how do you fix an error in the Merkle tree, even if everyone agrees it happened?). Even if you have an extremely corrupt county, that doesn't generally matter in the grand scheme of things; and its extremely unlikely, as any citizen in that county can stop the corruption by simply participating in the process themselves. Wide-scale voter fraud of this kind is simply impossible in a paper system. The only times it happens is like in Belarus, where it's not "an election", it's a public show that looks like an election, but where the result is pre-determined. The Merkle tree would show the same thing there: it's a mock election to make it look like a mock democracy. Lukashenko wouldn't have stopped leading the country even if miraculously the election would have shown he lost. Or, it can happen in other more complex and more discoverable ways, such as busing voters around to physically vote multiple times in multiple (preferably far away) polling places. As for Estonia, they'll come to regret this system sooner or later. It can work for a while, but there is no doubt that the system will get hacked, or the losing party will be able to convince enough people that it got hacked even if it didn't. Someone will accidentally publish private keys, like in this Colorado case. The system will go down on election day because of a bug. Who knows which one will be first, but it'll end their experiment. The rest of the world will continue with paper voting and not face such problems.

EGreg 2y ago
“In paper based systems, you and other volunteers” No, 99.99% of “you” go home and “trust the system” to some poll workers, many with major bias and incentives. Many of “you” don’t turn out to vote or are disenfranchised by simply living too far from the polling place or not being able to take time off work, when you could have just voted from your app. Certain parties even rely on suppressing turnout. (Can you guess which party does that in USA? Hint: it’s the one that closed 1600 polling stations right after the Voting Rights Act got neutered, and then got mad about mail-in ballots ruining their carefully laid disenfranchisement plans during the pandemic.). In fact, if you want the election to be “secured by multiple distrusting parties”, that is exactly what byzantine-fault-tolerant cryptographic protocols (which power many blockchains) are designed to do. (how do you fix an error in the Merkle tree, even if everyone agrees it happened?). Even if you have an extremely corrupt county, that doesn't generally matter in the grand scheme of things; and its extremely unlikely, as any citizen in that county can stop the corruption by simply participating in the process themselves. You are literally arguing from a double standard. In a paper election, somehow “any citizen” by themselves can stop the corruption… by simply participating in the process.” Yeah sure one guy exposes the entire corrupt county, with no ability to prove how anyone voted, why didn’t a single Belarussian and Venezuelan think of that? LOL” And on the other hand, when you have tons of anonymous irrefutable proofs by participants submitted publicly, you throw up your hands and say “what can we do to fix the merkle tree, even if we all knew it was corrupt?” The point of the trre is to catch errors, prove them and publish the proofs widely. As a society, you then have the proof nexessary to fix errors the same way you’d normally do it — by identifying the corrupt districts, and having a recount or revote just there. And bringing those responsible for tampering to justice. If you stop conflating all the things and unpack them, you’ll see that adding cryptography makes things strictly better: 1. You have more chances to catch if there have been extra votes cast because the private keys are coming from tokens handed out at registration. In a paper election you might have corruption at registration AND all manner of ballot stuffing later too. 2. Everyone can check their vote and report a discrepancy. Not just the volunteers at the polling places. And all because they can prove how they voted and do it anonymously! 3. Everyone can see exactly which districts are corrupt in giving out fake voter registrations, and where there’s smoke, there’s fire. They can do an audit and guess what, the cryptographic signatures are helpful for creating a PROVABLE trail that implicates the system. 4. The attack surface reduces to pretty much just the voter registration sybil attacks. Eliminating a whole class of problems on actual election day. 5. The results are reported to everyone reliably and quickly, or even in real-time (though the latter is “too good” because it might affect how later voters vote). There’s practically not a single problem that adding cryptography creates, which wasn’t already present in the paper system. And you know all this because if you honestly asked yourself whether dictators, who want sham elections, would want to do their next election with cryptographic signatures and merkle trees or not — what would be your answer? Be honest. And think about what that means for your argument.