 Thank you, Marcine. So I'm here to try to make people aware of a project that's been going on at Microsoft for a little over a year. We're trying to take tools from this community and bring them to the real world and impact real elections and in a real way have much better integrity and verifiability in elections. So what we're doing in particular is we are building a free open source toolkit. It's going to be released next month, actually, or at least the first version next month, on GitHub. And it will allow vendors to take this code and build cryptographically verifiable end to end verifiable elections. We are releasing a detailed specification that will allow anyone to look at exactly what's going on, do black box verification again of the artifacts of an election, make sure it's all right. And we've already built one prototype to show how this can be done. We may build one or two more. One thing I want to make very clear, this is something that Microsoft is not selling in any way. This is free to anybody. We're not building election systems. We're trying to work with the existing community of vendors and whatnot and say, here's something that you could really use. Please use it. We'll show you how. We have many partners in this project. Galois in Portland is doing most of the development work. Columbia University is an important partner. And most of the major vendors in the US have signed up as partners and agreed to look at this and try it out. And not just the US. We're talking to vendors worldwide. There does seem to be a lot of interest, so we're very encouraged about this. Just to give a sense of some of what's going on, showing what's in here, we're using threshold algomal encryption, doing homomorphic tallying of votes. Yes, I know homomorphic was a bad word to put in a talk if you want to get a paper accepted, but it's easier in the REM session, right? We have things that are full of interactive proofs. Basically, most of the work is proving that encrypted votes are well-formed, that there are ball zeros and ones, the right number of zeros and ones. So we've got Schnor proofs of possession of secret keys corresponding to the public keys. We've got Sean Pedersen proofs of correct decryption. We've got CDS proofs to do disjunctions so that you can prove that something is a zero or one with that proving which. The at Shamir is everywhere in this. Lots of tools that are very familiar. Nothing that new, but it's taking this and making it production quality, getting it really used. I want to just say a little bit about some of the design principles that we're using here. First of all, we decided to build a very simple API that turned out to be surprisingly clean, not a lot of configuration. We try to make it so that it's not a lot to screw up. We also have the principle that we're writing this code once. But we expect, we hope, that there will be hundreds, maybe thousands of people who write election verifiers, or the little election verifier apps. And I'll say a little bit more about how. So some of the design decisions that this triggers is we're using integer El Gamal, not elliptic curves, because we want first year programming students to be able to do this from first principles, very easily build a verifier. Also, just as an example, we could say we've got some property of a divided by b mod n that you want to verify. No, too hard, then you have to implement the extended euclidean algorithm. We'll provide c, you check that a equals bc mod n, and then you can check that c has that property just to make it as simple as we can for people to implement. It's a little thing, but that's the kind of thing we're working on. So the first version of this code is going to be released next month on GitHub. Here's the URL. Feel free to come talk to me about it. We're trying to get the spec out as early as next week if I can get off the stage and get this finished. Send comments to me. We want as much open review as possible. And as projects for your students, writing a little election verifier might be a one or two week class project, or maybe a little independent study for the early ones. It might be a little harder. And the documentation isn't quite there yet. We're trying to get it there. But this is a nice little crypto project for people to do. Hope you're interested. Hope we'll be using this.