 This is a classic dilemma between, do you want decentralized security with all of the delays and all of the messy that that creates? Or do you want a quick answer, even if it's the wrong one? Everything is better with blockchain, right? I'm Adam B. Levine, and this is Speaking of Bitcoin. For today's session, I'm joined by the other host of the show, Stephanie Murphy. Hi there. And Andreas M. Antonopoulos. Hello. Jonathan Mohan is out today. On today's show, we're talking about voting. How does it work today? Would adding blockchain to the mix make it better or worse? Long-time listener Jaws wrote in with some questions, and it's a discussion you won't want to miss. I think we do all this live, so I actually have no idea what we're about to say. So, getting to the email from Jaws, so he says, Hi, Adam, I'm interested in a real discussion of blockchain voting. How that could play out, what the potential issues would be, etc. I've been fascinated with the idea of voting via blockchain since getting into the space many years ago and the recent presidential election, which, no, I don't think was fraudulent, but it has certainly raised voting security conversations in general in the public space. I think about blockchain voting from time to time, and obviously the auditing and honesty truth aspects are an amazing appeal, definitely more so than the various vendor-based closed source voting machines. But the main concerns I think about are, one, how to onboard all citizens? How to essentially get everybody a private key? Do you do a national ID that also happens to be embedded with some sort of secure NFC private keys that is used to vote and then can scan the card or a public key QR on it on a home computer or phone to verify the vote? The challenge is making it usable to a large population of non-tech savvy foes. And then the other point is, what blockchain would this run on that would be most secure? Would the government run it, but anyone can run a node to help? Could that be trusted? Wouldn't a new blockchain with low incentive put up by the government be easily attacked by other stakes? Should it run on Ethereum? Could the Ethereum network sustain 150 million votes over a day or a week's time? How would foreign governments attempt to attack the Ethereum network to disrupt this? And how successful could they be? Wouldn't it be more palpable on a sharded ETH2 network? Anyway, still loving the show so much. Thanks again for your hard work, Jaws. All right, so that's a long email. A lot of questions in there. We're not going to get to all of them today. Well, I think we can simplify it. You said at the beginning of the show, Adam, that is everything better with a blockchain? Well, if you ask certain people, they would say, yes, but I don't think that I agree. So I'm going to say that this listener, Jaws, we love you, but you're overthinking this because this blockchain technology is not ready for prime time with voting for national elections. It's probably going to be used in voting in corporate situations. I'm sure various iterations of it, but for voting in national elections, no, it's not happening yet. I think the best bet is to go back in time actually rather than forward and go back to paper as much as possible and little Scantron machines. So when you start to think about voting, I think that there's all of this space to improve the way that democracy works in so far as getting more people to participate, being more inclusive about this. And so I think that people often come to technological solutions like blockchain on one side. But I mean, I recall back during the primaries this last year, there was the Iowa caucuses, right? And the Iowa caucuses are like this really, really old school way of doing politics where it's not even like a vote where everybody goes to. It's like people gather in a room. Yeah, they take a straw poll, right? Yeah. What is a straw poll for people who have never heard that, maybe international? So a straw poll is like a sentiment poll, right? And basically, instead of there being one vote where you vote and then that's the person you voted for, there's opportunities to discuss and to caucus, so to speak, to find the people who you agree with mostly and then to vote with them. And the part that's relevant, I'm not an expert on caucuses, I'm living California most of my life and I've never actually participated in one. But the important part is, is that it's a very manual person-to-person, in-person, at like community center type of thing. And what happened during the primaries last year is that an application was rolled out that was used to collect the reporting. And the application had some problems with it, but there was mass chaos throughout the system. And the mass chaos throughout the system had little to do with the software and it had more to do with the fact that you were asking literally tens of thousands of people to use an application when most of them were like in their late 60s, right? And you were asking them to change a process that they were very used to to a new process that introduces new challenges and new advantages, but which they weren't familiar with, in which they hadn't been very well trained on. The relevant part here is that the problems that blockchain is going to have when you try and put it into these situations, it's almost not about the technology like most of the things we're talking about, it's about those processes, right? And there's another part too, which is that blockchains are really good at knowing about stuff that happened on the blockchain, right? So if I send a transaction from me to you, Stephanie, then the blockchain is really good at knowing that I sent a transaction, because the thing that we were doing is encapsulated on and happens on the blockchain. But a vote isn't something that naturally happens on the blockchain. And so as Jaws correctly points out in his questions, there's all this opportunity for someone to get in the way. And really the thing about voting that it comes back to is how do you tell who is a person and who is somebody who is pretending to be a person such that they can get more than one vote, right? And this is a problem that actually we're very familiar with in cryptocurrency. It's called a civil attack. It's kind of like a third degree problem, because the first problem, the one the blockchain solve is how do you ensure that participants in transactions or the consensus mechanism are not easily civil attacked? But when it comes to voting, the vote isn't available to everyone on the planet. So simply being human doesn't make you eligible to vote. So eligibility is a form of attestation. And attestation is something that is associated with identity. So this is a second layer problem. This is not just civil attacks. This is about identity and identity because you have to first attest that specific people have eligibility to vote. And in the US specifically, eligibility to vote isn't enough because we don't have automatic registration. So you have to have eligibility and be registered. You have to in advance express the desire to vote and be registered as an eligible voter. But that's not where it stops. That's the second layer problem, which is identity, which we have not solved with blockchains yet. It's a very difficult problem. It's one of the holy grails of computer science. And it's not something that has been solved yet with blockchains. Then there's the third problem, which is great. You've got identity registration eligibility sorted out. But when you actually cast a vote, you now need to strip that identity and yet ensure that no one gets to double spend their vote. So you need to verify eligibility, but then you have to preserve the anonymity of the vote, which is a critical component of the franchise. Without anonymity, the vote becomes something that is sold, extorted, all kinds of shenanigans start once you remove anonymity. Remember, the only difference between a North Korean election and an American election is the fact that in North Korea, you have to vote in front of the official. And that somehow magically ensures that all of the voters vote for Kim Young-un, at least all of the voters that are still alive after the election. So anonymity is a pretty critical concept. And that's the third tier. So blockchains can help if you can solve first, transactional civil resistance, and second, identity attestation and eligibility, and third, anonymization after eligibility. We are very, very far from solving these problems. So the United States has an electoral system. I've heard it described as decentralized, but they don't mean like decentralized, like voting on a blockchain or something like that. They mean decentralized in that every locality, every county has sort of its own process, and every state has their own rules. And it's really like this patchwork hodgepodge of votes, but somehow they all have to vote in a common election for the president of the whole United States. So do you guys think that represents a form of decentralization? It's a very good point, which is that the federalist system that decentralizes voting, most people don't understand this, but voting in the United States is managed by the state exclusively without any question. There is no federal enumerated authority to do anything other than count the results. The federal government has no role in certifying, it has no role in verifying, it cannot even challenge, has no standing to challenge, the way elections are run. There were some brief periods where there were some implications for that, specifically the Civil Rights Act. But most of those provisions were actually repealed recently. And so the states get to run their own elections. Now this is incredibly frustrating to people who want to see consistency, homogeneity, predictability, auditability, and also because many states attempt to do all kinds of shenanigans when it comes to the voting process. But it's the other side of the coin, which is that for all of those reasons that are incredibly frustrating, this system is very, very resistant and robust to centralized manipulation. It's actually very difficult to manipulate an election across the entire country because you have to do it in 50 different states. And even within those states, the process is decentralized in voting precincts that have actual voting centers, which have independent monitoring, et cetera, et cetera. So there's layers and layers and layers of decentralization. And at each layer, there's layers and layers and layers of oversight and often competitive or contentious adversarial oversight by the opposing parties and campaigns. And this actually ends up being very difficult to compromise. So the things that are very frustrating are also the things that actually make the system robust. But they're also the things that make the system very difficult to change. So if we did see any kind of blockchain voting or any kind of technology change that's implemented in relation to voting, we would start to see it on a local level first and a patchwork of different systems. And that is what we have today. There's some states that rely really heavily on paper ballots. Other states, you know, kind of you go to a machine and you press a button and it spits out a QR code that a human can't even understand. And then you go to another machine to scan it. And so they do have different systems in different places. And like you said, Andres, that's frustrating to some, but also you could argue that it does add a little bit of robustness to the system because it is decentralized. A lot of robustness because failures, if they occur and when they occur, are localized. They're contained. They're not everywhere. There is no single point of failure in the electoral system. And what you didn't mention is there's also, I think, five states where voting is entirely done by mail or where vote by mail is the primary mechanism of voting. I think there's one state where it's entirely by mail. Oregon has an entirely by mail system, yeah. Yeah. And four more where it's the primary voting mechanism and you can opt out and go to a local precinct, but instead by default, it's assumed that you will vote by mail. And many other states have no questions asked, no excuse needed. Very simple mechanisms. California being one of them I voted in California under this mechanism always by mail. And it's very simple. You go to a website, you ask for your mail ballot, they send it to you, you send it back, done. And can you track it? Like, can you go and check what the status is? Yes. You get a little barcode tear off that's part of your ballot that has a serial number. And you can go on a website and they will tell you if they have received and then if they have counted. It's almost like tracking a UPS package. You can say, okay, they've received it. All right, now it's actually being delivered and it has been counted. Yeah. So presumably if we were talking about like, how would that look on a blockchain, quote unquote, because that is like a buzzword, right? Put it on the blockchain. But under a blockchain based system, perhaps you would use your unique identifier or your private keys or your public keys to look up your voting transaction and, you know, what time it occurred and validate that it was actually you casting that ballot and that it's been counted and recorded in the blockchain and it happened at this time. Yeah. And interestingly enough, the whole idea of having an envelope that is sealed on the outside, the serial number on the inside, a homogeneous ballot where, you know, you've got to understand that when I send in my ballot and I've taken off the serial number off the envelope, there's a serial number on the envelope, but there is none on the ballot itself, meaning that they can track that my envelope was received and when they open the envelope and put it through the counting machine, they scan and track that my vote was counted. But of course, the ballot itself that can then be audited after the election is not related to a serial number and they can't see who voted. They only know that mine was counted and one of the ones that was counted voted a certain way, right? So in fact, in cryptography, if you want to teach someone cryptography, the idea of having a sealed envelope that you sign on the outside that leaves an imprint on the inside is one of the techniques we use to teach this concept. This is called a blind signature. And a similar concept that is very popular and has received a lot of research lately is the concept of zero knowledge proofs. A zero knowledge proof is where you can prove something is true without actually revealing what that something is. So this is a classic application of zero knowledge proofs. You could prove that your vote was counted, but no one can actually tell what your vote was, right? And so the election officials can prove to the voter that their voter was counted. You can even prove that it was added to the sums, but the election officials do not know what that vote was for. And most importantly also, the voter cannot demonstrate to someone else how they voted. And this is important because in the past, it was very common to buy someone's vote and then require proof that they voted the right way. Yeah, this could also be used for all different forms of retaliation. And so it is important to have some kind of privacy protections for the voter to keep their vote secret. And you know, that's why some states have banned these ballot selfies, you know, where you take a selfie with your filled out ballot, that doesn't obviously preserve privacy even though you're choosing to share it yourself. So that's a whole different question. But the point is, it is important to have some kind of privacy protections in there so that people can keep who they voted for secret, but there's still some kind of audit trail so that you can demonstrate and validate the election results later if you need to. I'd like to briefly dig into this a tiny bit more because I find this a fascinating topic. I voted in three different countries, and I've also ran an election in three different countries. I was a volunteer for the elections in Greece. I have also both voted and volunteered for elections in England. And I also have voted and volunteered. In fact, I was a precinct captain in California. So I ran a precinct and I did the audit. So I got to see kind of the mechanisms of three different countries and how they work. And I find fascinating this whole concept of the secret ballot, some brief history, until 1890, in the United States, ballots were not secret. And you're like, what? Really? Is that true? Yes, it is true. And in fact, until that time, a secret ballot was known as an Australian ballot, which blows my mind. And the reason is because Australians started using secret ballots and then that was adopted by Americans and first by Massachusetts. So then the secret ballot was known as a Massachusetts ballot. And only after 1890, did it become a secret ballot everywhere in the United States. And you see this in a lot of countries. Before that, various organizations would distribute pre-completed ballots that you would cut out of, say, your newspaper. So instead of endorsements, they would print out a ballot already voted. And you would cut it out of your favorite newspaper. And you would just drop it in the box and vote not just on a party line, but on a newspaper line effectively. And one of the things that created were situations where people were obviously pressured, bribed. But I think one of the most interesting things was the issue of the suffrage for women, the suffragettes. Because one of the things that happened when women were first given the vote in the United States was that they were pressured by their husbands to vote the way the husband wanted. And not only pressured, but they required proof. So this is a very interesting thing. Most people don't think about this, but the secrecy of the ballot isn't just because you might get chased by some opposing element down the street. In many cases, it's your own family. And I love this idea that someone can go into the voting booth screaming at the top of their lungs the husband's candidate name the entire time, and then drop a ballot for the opponent. And no one will know. I mean, that's the important part about that anonymity is that you have plausible deniability. And so it's actually something that used to exist within parts of the higher government. What was the guy's name, Stephanie? James D'Angelo? Yeah, that sounds familiar. When we were getting started with the show, he did a bunch of videos about Bitcoin. And then he left the Bitcoin space after he made this video about how the reason why Congress doesn't work anymore is because I can't remember the exact date, the 60s or the 70s. They made it so that all of the politicians who voted their votes were logged and you could see who voted for what. And his whole thing was that by removing that anonymity, it was cheered in many circles as a way to make sure that our government representatives are being held accountable. But what it did, in effect, was it removed the plausible deniability that the representatives had so that they could have lunch with lobbyists, right? And say, oh, yeah, sure, I'll vote for that thing. But then they actually, when they're making their voting decision, could vote their conscience. So again, the use of that plausible deniability, whether you're talking about voting so that your family isn't upset with you, but you still get to express your will as a voter, or it takes place in Congress, really, there's this valuable thing to it. And again, if you look at how the US government is set up, and this is a whole other episode that we'll have, decentralization is at the core of almost everything. Many of the changes that have been made over the last 100 years to the way that the government works have actually removed certain elements of that. It's still really decentralized. 100 senators right from 50 states, and you're talking about 400 and I think 35 representatives. Decentralization is messy. It is messy. It is inefficient. It is not the best way to do things. But it's hard to mess with. It's hard to go and compromise all of these different areas. There's just so many moving parts and they do things differently. Now, some of that difference and some of the value of decentralization can actually come out of it a little bit. If it's a decentralized system, but everyone does the same things the same way, right? We've seen a little bit of a drift towards that as voting machines have come in and other things like that. But again, as a system on the whole, it still is highly decentralized and therefore very robust against any sort of systemic disruption. Disruptions at a local level, those things happen. Systemic disruptions much harder to pull off in the decentralized system. Hey, folks. Adam E. Levine here for a quick update. Last week, along with the show, I listed our first NFT collectible in a no reserve auction and within a day we'd received a very generous bid for one ether. To be honest, it's really humbling to receive that sort of support from a member of the audience. And it's enough of a proof of concept that we're moving forward with some larger projects around tokenized collectibles, which I'm excited to share more about probably next month. In the meantime, check the show notes for this episode to find our second collectible token available in another no bid auction that, just like last time, will run from the release of this episode on Sunday, the 14th of March and end sometime between Saturday and Sunday of the following week. So if you missed out on last week's, here's another chance. And for folks who'd like to collect or patronize our show, but who can't justify the prices we've seen so far, have no fear. We're working on a less exclusive series that you won't have to compete for. But all that's in the future. For now, happy bidding to those of you who will be doing so, and thank you very much for listening. And now, back to the show. I'd like to relate this back to an argument or debate we've had about Bitcoin in the past, which is extremely relevant to this. And that is that you can have very decentralized robust transactions, or you can have cheap and fast transactions, but rarely can you have both. And one of the things that's happened, and I've certainly noticed it in the 20 years I've been in the United States with elections, has been that the influence of the TV real time bar chart pie chart reporting of election results has over the past 20 years steadily increased the emphasis on speed of closure, to the point where some people are like, what do you mean they're still counting on the fourth? When in fact, it's very, very common to still be counting ballots a week after the election. So if you want really, really fast elections that give you closure where you get a result as quickly as possible, North Korean elections achieve that. The result is already known before the election opens. It's a very efficient system. And if you want elections that cannot be compromised, that require manual auditing and recounts and have triggers that occur when things are close and they need further scrutiny, and they have layers and layers and layers of auditing and certification, well, that's going to take time to count. And it's perfectly okay to take that time. And so this is a classic dilemma between, do you want decentralized security with all of the delays and all of the messy that that creates? Or do you want a quick answer, even if it's the wrong one? And I think the emphasis here has to do with the difference between casting and counting. A lot of the efficiency technologies that have been introduced in American elections over the past decade have been about counting votes fast. And in the process of ensuring that votes can be counted fast, they've changed the way the votes are cast. Stephanie mentioned before the idea of a paper ballot that then is scanned by a scantronic machine. So for people who are outside the US who have never heard of this idea, this is also how the university entrance exams are scored. Yeah, I was going to say that if you're at least like kind of my age, this is how your teachers tabulated the results of your tests. Right. So the way this works is you have a card or single page sheet and it has some choices. And the choices are these oval bubbles that are a black line with a white interior. And you take a soft pen or pencil and you fill in the oval so that it's black. And once you've done that, the ballot or exam result or multiple choice test answer or whatever it is is fed through a machine. And all the machine does is it scans a barcode to know which ballot it is because there might be national, local, federal, there might be several ballots with different arrangements on them. And then once it's scanned that, it looks down the columns and where it sees a black circle, it marks that column. Or if it's an empty circle, it marks the other result. And so that's how voting happens. And the brilliant thing about that is that it solves both of the problems. You can count really, really fast because you can have a machine count it. And then if there's any doubt about the count, you can audit the actual paper balance because they're still there. And these machines tend not to be connected to the internet either. If there's concerns about malicious actors being able to access the machines through the internet, which I've heard lots of people talk about recently, these Scantron machines are a technology that has been around before the internet really. And they don't need the internet to operate. And so it's like a good old-fashioned incandescent light bulb solution that works. Yeah. The way they look, in case you're wondering, is they look a bit like a photocopier that has a feeder at the top. So you put a stack of, say, 500 ballots in the top, and it goes flip, flip, flip, flip, flip, flip, flip, and makes a lot of worrying noise. And then at the end, it just shows you a count on a screen. Actually, the last time I voted in New Hampshire, they were using a system like this, and they would put your ballot through the machine. And then if there was any problem with it, it would beep or something, and they'd say, oh, there's a problem with it. But if it went through, you knew that your vote was counted. Yeah. Now, the beauty of this, of course, is that you can go back and manually go over them. This beauty was ruined by Florida Institute's mechanism where instead of filling in the oval with a pencil, people would punch a hole through a card. The hanging chads. And they created the hanging chads fiasco of the 2000 election. Well, I mean, to be fair, there was a problem. I remember you used to have to use a number two pencil to fill out these exams. Well, the point is, there could be problems with paper ballots. There could be problems with punch card ballots. There could be problems with any system. But I mean, it's great that we have the opportunity to try out different systems in different places and have this decentralized patchwork. And I think that, yeah, probably even though there's some concerns that I have about it, states and counties will move towards more electronic voting systems. Will they be blockchain-based? Some of them, yeah, eventually as we get farther into the future. And we'll see how it works out. I think that if blockchain voting is first refined and smart contracts that manage it are first refined in corporate settings, like not public kind of election settings, that would be helpful because the kinks will be worked out basically. And private companies, I think, would be good at refining these kinds of smart contracts. And we just have to try it and see if it works, I guess. We need to be very careful about separating the motivation to count fast and give a short attention span audience the closure they want with discarding the ability to even audit these results. It's one thing if you're filling in a paper ballot that gets scanned fast. It's a whole other thing if you're tapping on a touchscreen of a Windows machine that hasn't been patched that's connected to the internet and is counting. And there's absolutely no way for anyone during, before, or after to figure out if that thing actually counted the vote and if it counted it correctly. That's a problem. And my perspective with security has always been that the number one rule with security is keep it simple. The KISS principle, keep it simple, stupid. Nothing beats paper and pencil. I will say that when I'm talking about 24-word mnemonics and people will say, no, let's encrypt it and put it online with double encryption AES and no, no, steel, punch, paper, pencil, simple, offline, manual, auditable, transparent, replicable, resilient, robust. These are important principles and some things that a five-year-old can replicate. Or a 65-year-old. Or a 65-year-old. And these voting machines that are running Windows that arguably are 90s technology or maybe early 2000s technology are already opaque, unauditable, and not transparent. Adding blockchain to that mix, simply alienates a whole new group of the electorate that won't be able either to vote in the first place or if they do vote will have zero confidence in the outcome because they have no visibility into how it works. And I think that's the wrong direction to go in. In fact, as Stephanie said, keep it simple, keep it primitive. There's no reason to rush the count. There's no reason to make elections efficient. Efficiency isn't the parameter we should be optimizing in elections. Access, transparency, motivation, eligibility, and franchisement should be the goals we should be maximizing. You know, to me, I think one of the most iconic symbols of voting technology is indelible ink that people dip their finger in after they voted so they can't vote twice. So here's the deal. This is a longer conversation. This is the end of this episode, but it is not the end of the conversation. As Andreas started to get into there, we're going to talk on the next episode about the identity challenge, indelible ink, signature verification, you know, blockchain ID, all of these different ideas that all have their advantages and also have their disadvantages, too. The other thing we're going to talk about is what blockchain could something like this go on? And, you know, again, our listener Jaws who wrote in had some really interesting ideas around that. We're going to explore those ideas on the next episode. So this has been kind of like an introduction. And I think you got the broad answer, which is that if we just plugged blockchain into kind of the system as it stands right now, would that make things better or make things worse? Chances are pretty good, it would make things worse if you didn't change anything else because the problems that blockchain is trying to solve really don't address many of the issues that are underlying problems with accessibility or, you know, the ability to validate our election system as it stands right now. But that is the next episode.