 Hi everyone, welcome to the last session of today. I promise I'm going to make it quick so we can all head to the party afterwards. So I'm Bernard Kolobara. I'm the co-founder and CEO of Lunatic. At Lunatic, we are building a web assembly platform inspired by Erlang Principles. This means massive concurrency, soft thrill time, and full tolerance. To wrap up today, I would like to share with you a bit of our story why we chose web assembly for Orstec and especially how it helps us around the full tolerance part of our platform. I started my career in IT support, and a big part of my day-to-day job was just to go around and turn things off and on again. It became somewhat of a joke in the IT world, but believe me, it's like a really effective method in fixing issues. It comes from the fact that long-running applications tend to get stuck in an unpredicted state, so the best way to recover from this error is just to revert to a well-known fresh state. And also it's the state that got the most testing out. Because developers, when they write some application, they write some code, then test it, and then write more code, test it again. So this fresh start after the application just starts, it's usually really well tested. And Erlang, the programming language, takes these principles and applies it to software development. And this approach has proven to create some of the most resilient applications out there. It works in Erlang because the whole programming language was designed around these concepts of small processes that don't share any state with each other. This means if one of them fails, we can safely restart it while the rest of the system keeps running. This is also known as the let it crash philosophy in the Erlang world. And with Lunatic, we are trying to bring this approach to all programming languages using WebAssembly. To show you how we do it, let's, for example, take Rust. Rust application usually share a whole memory space. So there's only one memory space in the Rust applications. And in this example, all threads access one shared state. This means if one of them misbehaves, corrupts the state, it's kind of corrupted for everybody else. So if we just shut off one thread, restart it, this would actually not really help us because the thread would be respawned inside of the same memory space. This means that we need to take a different approach. This is also true for async tasks. In async Rust, all the tasks run in one memory space and can reference the same memory locations. So we never really have the guarantee if we restart it that it will start from a fresh, completely new state. So what we do with Lunatic is we introduce a new concurrency model to Rust, turn each task into a WebAssembly instance, so each of the tasks gets a linear memory on its own. This means that it's safe to restart each individual task and we can guarantee that the memory is gonna be completely fresh and the state is gonna be new. And to be able to do this, it feels a bit like a superpower, right? Because the Rust language was not even designed to be able to support this. Just by applying WebAssembly to it, we get all this awesome property. And I feel like this is a common theme with WebAssembly. As we saw today, we take it, apply it to a specific domain and we gain all these new superpowers. It allows us to run high performance C, C++ application inside of the browser. We can nowadays run Photoshop, Figma, Google Earth, all at a high frame rate inside of the browser. And it's also bringing compute closer to our data in the database and also compute closer to the users on the edge and it's powering a lot of new innovation in the cloud native space. And Bailey showed also today in the keynote many more use cases for WebAssembly. Every time I talk to somebody new that uses WebAssembly, they kinda show a slightly different variation how they are using WebAssembly to solve their problems. And that's why I'm really grateful that we have events such as today where we can all meet, share our experiences, how we use WebAssembly and learn from each other. And I would like to finish today by thanking everybody who made this event possible. I would like, first of all, to thank all the organizers and program committee, Matt Butcher, Taylor Dolezal, Divya Mohan, Liam Randall, Ralph Scrillacci, Lindsay Gendro and Sherry Twenbly. I would also like to thank the sponsors, Microsoft Azure, Suborbital and Profian. And of course, I would like to thank all the speakers. We had so many amazing talks today and everybody attending the conference in person or watching the event online. Thank you so much.