I saw a website. There was a web page for each question.

No.

Anyone who wants to, can submit ballot question language.

Here’s how I would imagine it. We would run a tournament of, let’s say, binary contests.

Let me just roughly sketch out for you. You could ask me, David, how would you foresee the random sample voting be used if say you had your way? It was miraculously the case that we decided to use it here in Taipei, and make the best use of it. ...

Yeah, I know it’s coming up and it’s a big deal. I think there is also the question of absentee voting.

Let’s switch gears, break the thread there. Let me talk about, briefly, an example way to use this system in Taipei.

This means that the choice of voters is locked in. As soon as that random experiment is made, which you could use blockchain for, or you could use a stock market closing price, or something, there is no way to change it, the possible attack that you were sketching doesn’t ...

You flip a coin basically and decide whether to say it’s heads you get to be a voter, or tails you get to be a voter. I have no way of knowing what you’re going to do, there is no way for me to change it once its covered by ...

No one’s supposed to see what the outcome is, and then the counter-party says, "If it’s heads, I’m going to be a voter." Then, only then, is it opened or revealed. Because of that mechanism, there is no way for me to cheat you out of having a 50/50 chance ...

This is an excellent observation, but the way we do it is secure it against that certainly. What happens is that who the voters are is locked in...Let me just explain the basic randomization of who the voters are process in outline, because that I think explains a lot of ...

You need to actually go and check that real people really did get the ballots. The way we do it is perfectly effective, but it does have the consequence that the simple way, the most secure way, is we just reveal who the voters are, but only afterwards.

No. It’s because the election officials could pretend that they did all this, if no one ever knew who the voters were, there might be no voters at all. [laughs] They could just say, "We mailed out a thousand ballots and here are the results," but actually, they never mailed ...

No.

Now, why do we need either of these or at one of these mechanisms? Do you have any answer to that?

Then we say, "Go talk to that voter, get this number from them, and make sure they’re happy and then bring that number back and when we see that number come back then we’ll know that you really did check on them."

The way we do that is not perfect, but it is by a kind of what you might call indirection in computer science. We say, "OK, we’re going to reveal the name of all voters, but each voters name is only revealed to a single randomly chosen person that’s specifically ...

The only bifurcation in the proposal, and it’s actually just a footnote, is that there is a version where...The basic version is where who the voters are is revealed after the vote in a kind of audit. The special version is different in that the voters are more or less ...

I’m sorry. You threw me for a loop there.

I wish I was half as smart as you.

The second thing is that the tally does, in fact, reflect accurately the votes cast. That second property is also unconditional, statistical. The third property is of voter secrecy, where how an individual voter voted will never be revealed. In that white paper, which I think you have in front ...

One thing we found helpful in explaining this was to say that it is equivalent to each voter being given the opportunity to flip, say, 20 coins. They would be allowed to vote in that election if all 20 were heads. We proved that the choice is unmanipulatable by even ...

One is that the voters are selected independently and uniformly, at random. We found that most people, even scientists, when they hear about this first, their biggest concern initially -- this is valuable information -- is that the selection itself would somehow be manipulated and that it wouldn’t be truly ...

Random sample voting proves to the public, irrefutably, because the proofs can’t be falsified or fabricated, even by someone with infinite computing power. You realize that they are statistically -- we say unconditionally -- convincing. That’s very strong, and that’s what you want in a situation where a state actor ...

You have some understanding of the random sample voting. Let me try to refresh your memory, just briefly, and then maybe we could talk about it more substantively.

Maybe we could talk about the random sample voting just generally for a moment.

You’re voted on by the national population?

Thank you. By the way, what is your tenure here? Is that going to be for quite some time?

Plausible deniability.

That’s interesting.

That’s interesting. I said I was big fan of this.

I see. That’s interesting.

[laughs] Help is quite a thing. I have to keep in mind that I’m on the record.

We had a system called Two Ballot, which hasn’t been published yet.

Oh, really?

Yeah, Nicholas.

Did you have a chance to maybe look at some of the things that I’ve been doing more recently?