My gosh. Welcome, everyone. Wow. We are a lot bigger than last year. I can already see so many familiar faces. It's so great to see all of you. It's just absolutely amazing to be here. It is an absolute pleasure to be here with you today. I was thinking about the group that we're going to have in this room, and together we're co-inventing the future of communication, the future of storytelling, really the next generation of how we come together. Right now, in the world, at this moment in time, this, I believe, is where it's at. This is the group. This is the group of creators, of inventors. There is no other collection of creators in the world today. It is right here. Woo. We are reimagining the way people come together. We've been doing it for over 15 years. When we started together, when Roblox first shipped in that very blocky world, all the way to where we are today, we are reimagining the way people stay connected, the way we tell stories, the way we learn, the way we work, the way we play, the way we consume entertainment. We have a long history of trying to reimagine things. We've used a lot of technology to think about this, and sometimes we lose track of how exciting it is to be right in the middle of it. We are right in the middle of two big technology waves. The first is how we storytell. We started storytelling a long time ago around the campfire, and ever since then, whether it's been cave art or paintings, we've been trying to figure out how to tell stories to more people, how to make them more vivid, how to tell them to a million people at the same time. We have continuously figured out ways to do this. We figured out how to write down our language. We figured out how to move to books. We invented photography to capture images so we did not have to paint them all the time. About 100 or 100+ years ago, we merged images and audio, and we created the movie format that is still big today. There is something about movies that we are all very aware of, and that is we cannot jump into the movie, and we cannot do the movie with our friends. You are all telling stories on Roblox that we together are interacting with. First off, on storytelling, we're participating in the next evolution of how we share and tell stories, and it's interactive, and it's something we do together. There's another long history. How do we communicate at a distance? How do we send a message across town? How do we send a message to another country? How do we stay in touch with loved ones when we can't be together at the same spot? The mail system served us for a long time, but that ran out of steam. The Pony Express took 10 days, I believe, to go from San Francisco to the East Coast. We invented the telegraph. We can't use the telegraph. Once again, our desire to communicate pushed forward. Along came the telephone. Now we could all use this with audio. That was pretty good. We could not see each other's face, so we are all familiar. We moved to video, and we have been doing this a lot for the last two years. There is a next generation to this, and there is a next generation where instead of being in different places where we can be together, instead of being isolated, we can do things together. We can play hide and go seek. We can walk down our graduation aisle. We can do things together. That happens to be the same discipline we have been working on for storytelling. Both storytelling and communication, we together are inventing the future of both of those. It is a future we have been working on for 15 years. There's a big future ahead of us, and we're all here together to talk about not just the last 15 years, but what is the next 15 years. With all of that, let's take a little stock of how we're growing around the world, how we're aging up, how we're starting to explore new areas of the platform, how our economy's growing, just how we've all been doing together. First off, wow, we've just been growing. I want to highlight it is all ages. It is around the world. It's more than just play. In addition, we're seeing more time on the platform as well. This is growing up real quickly. For those of you that I think there might be some of you that were in the Exploratorium when we did that very first BLOXcon over 10 years ago, when we said the target is 400 million hours, and we were at about 40 million hours. We have blown past that target. We have hit this 4 billion hour target. This is half an hour for everyone every month on the planet. There is a long way to go here. As people start using this type of technology to communicate every day and tell stories, this number is going to be much, much bigger. We are growing around the world. Let's take a look at our global user community. Of course, the U.S. is big. There are over 11 million DAUs there. If we look at Brazil and the Philippines, there are over 3 million daily active users. If we look in Russia, if we look at Mexico, there's over 2 million daily active users. If we look at Indonesia, there's over 1 million. We are growing in every country: Thailand, Vietnam, everywhere. A lot of growth there. This growth is powered by our technology and your creativity. Together, there's some amazing stories that are still happening. Let's take a look at Brazil, for example. If we dive in in Brazil, we've seen over the last three years, I think, over 5x growth. This may be four years, which is absolutely amazing. Let's look at a couple really interesting countries. We're going to look at India and Japan. India, very interesting. A billion people, very difficult infrastructure, very difficult bandwidth, very difficult form factor for devices. Japan, very interesting because it's one of the largest economies for game in the world right now. Both of these are growing amazingly. This is 15x growth in four years in both of these. We are really excited in India. We are not talking a light version of Roblox or a special version. This is the exact same Roblox version we have been using for 15 years. We will never bifurcate the product. We will make our core product faster and better. In Japan, through all of our awesome efforts of translation quality and all of that that we will talk about, we are seeing this growth as well. Our global developer community just keeps growing. We have, I want to call out, I think today, hopefully, maybe in the audience, we have Argentina, maybe Croatia, Finland, the Philippines, Turkey, Venezuela, and beyond, all in the how right now. Let us give it up for all the travelers. You'll notice I didn't mention the obvious ones: Australia, New Zealand, the U.K., Canada, Europe. That's just par for the course. We see that all the time. It's really growing. Welcome to all of you as well. Let's take a look, just a little dive at a few areas I want to dive into. One is growing around the world. Another is all ages on the platform. Another is our economy. Another is going to be just the expansive use of our platform. The work we've done on translation technology is absolutely amazing, and it's consistent with the vision that you build once on any device in any language. This started with a dream 10 years ago where we imagined someone who is very young, first experiencing game development, pushes the button. They're on a phone. They're on a tablet. They're in the U.S. They're in Germany. They're in Japan, everywhere. This has really come to hold with our machine translation technology. Thank you, all of you who've adopted this. Let's take a look at a few countries and see what's happened here. I want to highlight specifically Japan and reference it to the prior slides we looked at. Now that we've started really focusing on the quality of machine translation, quality of search and discovery, we've seen 8% player growth in Japan. This is part of the engine that's been driving our growth there. Our translation technology is rapidly becoming best in class. We benchmark it against other cloud translations. It's reaching that critical mass where it's getting better every day based on feedback. Now, in addition to translation technology, what else is powering our growth around the world? A lot of you who are creators and developers know that underneath all of the fun, exciting content, performance is super important. I want to talk a bit about our global infrastructure. Let's take a look at what we've been doing. We have been on a vision ever since we really started to build our own worldwide human experience grid for performance and for cost. For performance because we're trying to connect people in different countries everywhere, and for cost because we try to run a really lean company. The leaner we run our business, the more efficient our global infrastructure, the larger our DevEx system is. We want to take every dollar that we save and move it to the developer community. That's really the way we think about and run our business. For those of you that are into it, we'll show you the march of our infra over the last three years. Some fun stats. What's 3 terabits per second of bandwidth mean? That's amazing. That's more than a lot of cloud company in certain configurations. What it highlights is under the covers of this, not just where we are, but a future. A lot of Roblox users don't know it, but when they log into Roblox in some country in some part of the world, they're connecting to our backbone. They're going over our bandwidth to our servers to give amazing performance. There's a lot of room here that we're going to expand beyond to really build what we think is the world's best human experience grid. India, I mentioned before. Interesting that we went up front. We said, "Let's build an edge data center in India." Whoa, India's growing very, very quickly. It highlights how important performance is, and we continue to build this out. Let's jump to our creator community a bit. I want to focus a bit on economics and some of the stuff we talked about doing. For those of you that were with us three or four years ago, which is generally the more we can enrich the economy of your community, the faster we're going to grow. Behind the scenes, when we're having our executive staff meetings and our board meetings, we're talking about you quite a bit. We're talking about moving money to the creator community, about running an efficient company, about expanding the economy. We talk about you a lot, and we know you power things. How is this doing? Let's take a little look. First off, how many creators are earning Robux every month with engagement-based payouts? This has just gone crazy. We have creators ask, "Is this all going to the top devs? Are the top devs getting bigger? Are the devs that might have an experience by playtime ranked at 1,000 or 10,000? Are they getting screwed?" All that. What's going on? We want to show you some of the numbers. Let's take a look. This is where we were in 2019. This is before engagement-based payouts. This is roughly what the yearly take would be of a creator on our platform, number 10, 100, or 1,000. We'll fast forward to 2022, and you're going to see something really interesting. What we hope to do, which is get the creators at rank 1,000 over $30,000 a year, we did it. We had an offsite about this three years ago. We said, "We want a $30,000 for these creators." It exactly happened. Let's take a look at the growth rate that we've seen. That 200% is a 3x, not a 2x for the top 10 devs. That 869 is almost a 10x for those 1,000 creators. The bigger we make our economy, the more we think all of this is going to go up. I want to highlight there are many more than 1,000 people in the world making a living on Roblox right now. Some of our top creators have more than 100 people in their studio. When you start multiplying that together, there are many more people we believe working on the platform in the ecosystem than are in our company itself, which is exactly where we want to be. I want to show an interesting thing about the UGC catalog. Once again, we talked a while ago where we want to move this all to you: avatars, clothing, everything. We do not want to be in that creation business. This is the very final part that is going to come to fruition. This creator base has grown. There are really two opportunities here. We will talk about this in a little bit. One is all of you that are UGC creators. This is going to keep growing. We are going to talk more and more about creation in experience itself. You are going to see us pushing and pushing and pushing where if you have a particular style of avatar that you use in your experience, your experience will become an avatar editor. Your experience will become an avatar creator. We want your experience to persistently save an avatar that can be used throughout Roblox. This goes for avatars and clothing. Our hope is you will help participate with us in moving this asset creator, expanding both our current asset creators. If we go forward a bit, we really want every person on Roblox, with your help, to be an asset creator. We think ultimately we want hundreds and hundreds of millions of creators working within your experiences to create things. We are working hard to expand the economy. We are working hard to expand our vision. I want to talk a bit about what we believe is happening right now. What we are all doing goes beyond making games. We are making social experiences. We are making other things. Let us take a look at where we think this is going. Beyond social experience creators, one of the ones we wanted and we hoped about when we started the company is actually educational developers. I want to talk about this because it is so special. Think of all the people who make educational software, curriculum, textbooks, who build things within schools. A lot of this is going to move to 3D co-experience. There are so many things we cannot do in the real world. There are so many places we cannot go. Many of these experiences are so much richer than a video or a textbook. When I say travel to ancient Rome with your friends, this really is going to happen. People are going to be developers for this type of stuff. Of course, we are also going to see brand creators and music creators. We are already starting to see a whole flock of creators. Some of you are here who are side by side our game and experience creators. You're supporting brands. They need help inventing this new medium. This is no different when the movie projector was created or internet video was created. Everyone's trying to figure out who are the experts in building this stuff. There is also a big emerging market for our developers working with both music production and with brands. I want to dive in on education just because it's so near and dear to my heart. When we started Roblox, the prior company we ran was called Knowledge Revolution. We were at that time making very primitive simulators for physics. One of the dreams we had for Roblox is, "Whoa, not everyone has a robot construction set. Not everyone can participate in first robotics. How do you get all the motors and servos and hardware? Lo and behold, Filament Games along with FIRST Robotics on top of Roblox is building our dream, really. I mean, this is a robot construction set where you can drive, where you can compete, where a lot of the things that have so far been isolated to a smaller group of people who have access to this, this can be available for everyone in the world. Every student in the world can participate in FIRST Robotics now that this is on Roblox. Another example I want to give a look at is where are people going to learn to code? Is it going to be 100% in Roblox Studio, or will there be other ways that are maybe a more gentle on-ramp? Just as we've talked about avatar creation in experience, clothing creation in experience, we also want to highlight our vision. We want to support coding in experience so that that Turtle graphics that we might have used in college or Robby the Robot or whatever, this is going to become a 3D immersive thing. People are going to be introduced to learning within experiences. We like it because someone other than us is building this polished educational experience. There's a huge, huge opportunity there. Okay. Let's talk about something that's near and dear to all of our hearts, which is Discovery. Discovery is so big, right? It's so important. Number one, as we have older people on the platform, as we have people all around the world, this has to get infinitely more responsive and infinitely more personalized so every single person finds what we think is the best experience for them. It is really exciting because unlike video, we have all these signals. Who are your friends? What are your friends doing? What have your friends done that can power discovery as if we have never seen it before? We have an amazing team working on this. There are a couple of ways we talk about it. One is, how much can we know about you the second you sign up for Roblox? It is called the cold start problem. How much can the very first thing you see be the right thing? The other thing is, what are all those real-time signals to help you find new things? I want to highlight two areas that we're really focused on. The first is this traditional notion that we've historically completely ignored: monetization. The question of should we offer some benefit to developers who have figured out, "How do I balance both engagement, retention, and monetization?" That's the highly difficult thing that some of you are doing. We feel you should get a little benefit for that. We want to be very careful because historically, we've been 100% doing discovery based on engagement and retention. We are going to nudge it just a little. This is called the efficient frontier. For those of you that are mathematicians, you can see that's a flat line, which basically means up till now we've ignored monetization. We're going to make a small tweak, and we've already shared this with you, where for two experiences that are exactly the same on retention, exactly the same on playtime, where one is really monetizing, which is very difficult with that same thing, we're going to give them a nudge. I'll give you an example of what this looks like. By the way, these are all fake experiences. Just heads up. We didn't put any of you up here. This is our vision of making up Roblox games. You will see very subtly, if we see a couple of experiences that are monetizing really, really well with that same retention and engagement, they're going to get a little nudge on our discovery sort. Visually, you can see what that's going to look like. The next thing I want to talk about is new content discovery. This is the dream of Roblox. This is when there were four of us in a room together. We talk about this all the time. Can we show up on Roblox? Even when we're really big, can someone show up, make something amazing, get to the top of the charts, get millions of players without doing anything except making a great experience? We already do this pretty well, but we're putting a lot of focus in this. As we get bigger, we think it's really exciting. We want to stay true to that vision. We're getting more and more sophisticated on the signals we have to do this. We think there's an amazing opportunity here. For those of you that are curious, what does this look like when our ML and our algorithms see, "Oh my gosh, the retention of this very small experience is off the charts. Everyone's coming to this. There may only be 100 players. Here's what we'd like to see happen." Those experiences as well get a chance to participate in search and discovery. In addition to all of the things we'll do on boosting and advertising, this is still foundationally how we thought about and ran the company. Okay. The next thing I want to talk a bit about is the economy. For those of us that watch the gaming space, this has been a dream for 15 years, really. It's been a dream in addition to traditional online advertising, whether it's text or video or web. There's been a dream that someday there will be an immersive, native, non-invasive, not jumping off platform, very civil, very safe next generation of advertising. We do not take this lightly, right? I do not know. Is there anyone in the room who remembers pre-roll video on Roblox? Okay. Yeah, right? For those of you that remember it, we did the right thing and we got rid of that, which is a great day in, yeah, that is a great day in Roblox history. Woo. This is the one we have thought about for a long time. The reason we like this one is the brands we work with are fun and exciting brands. They are brands that a lot of our users want to engage with and interact with. We are creating a system where if you want, you can put units in your experience. These can be portals. They can be billboards. They can be pictures. It's up to you. If you have that in your experience, ads will dynamically run on it. High-quality brand ads in addition to engagement-based payouts, in addition to Robux. This is a third leg of the economy. We think this is purely additive. Purely additive. We have over 4 billion hours of engagement a month. When people run those numbers, that is a huge opportunity for us. At the same time, we can do it gently and slowly and with quality. This is what it looks like. A lot of our brand partners want to reliably bring new users to their experience. We want to make sure once you visit, you come back into your experience and can return home. More to come on this. We're going to be playing around with this with early versions as early as this year, which I'm super excited about. We have a history, whether it's international translation, whether it's our virtual economy. When we do these things, at the time, they don't seem as big as they are. In two to three years, this is going to be another way to help you fund your studios. Another one I just want to talk about. Let's take a look at our Domino Crown. What is unique about this? There's actually some stuff that's unique about this that is a legacy of old Roblox because it's limited. It's very rare. It's very expensive. We have had parents call me and send me. People are trying to figure out how to sometimes even buy these things, and they're very, very expensive. We want to get out of that game and figure out a way for you to interact with our catalog and our economy to make limiteds and to do it in a good way, in a thoughtful way. Hopefully soon, this is what our catalog starts looking like. These will be something we will not make. These will be something you will make. We want the day to come when the valuable items on our platform are not made by us. Okay. We talked about economy. We talked about around the world. Another interesting thing that happens when I talk to investors or employees is, "Yeah, but what about that aging up thing? When are we going to age up?" It's funny because we have been having that same conversation for five years. I typically tell them, just like right now, this is arguably the largest gathering of creative people in the world. We're right in the middle of aging up as well. We're not starting. We're in the middle. Let's start with our brands, and let's look at how we've grown over the last three years with our brands. The growth here is amazing. What's interesting here, this is bringing a lot of older traffic to our platform. A couple of other things is you can see some of the newer brands are highlighting the opportunity in music, are highlighting the opportunity in fashion, are highlighting the opportunity in beauty, and also highlighting not the old way where you have a big event, but highlighting a future where these will be permanent presences on the platform. They will be bringing organic traffic to their experience. They will also be figuring out ways to reliably get a million or two million visitors a month. They will most likely be asking for your help with immersive ads to do that. There is going to be a wonderful complement between these brands, permanent experiences, and the ad units that you put in your experience. Our answer to when are we aging up is this graph, which is, I know we are in the middle of it. The majority of people on our platform are 13 and up right now. There is a lot of headroom in the 17 and up segment. A lot of headroom. There are a lot more 17-60-year-olds in the U.S. than 9-12-year-olds. There is a lot. That segment is growing very, very quickly. For those of us that like to extrapolate, it is a huge opportunity. Part of aging up is to complement what we've done. This is both in voice and in text as well, with more freedom. We know we need this, right? My daughters are not going to use this product, right? They're very sophisticated, savvy users of social media. They're between the ages of 18 and 22. This doesn't cut it. We are aware of this. As we introduce spatial voice, consistent with our very huge focus on civility and safety, this is where we're going to this future, which is more freedom. Woo. All right. The other thing that we have coming on the platform that I just want to dive into is experience guidelines. This is going live on Monday. Really excited. This is like translation. It's something we talked about for four or five years. We were in the middle of it, and it was a lot of hard work. Now we look backwards at it. We go, "Oh yeah, of course, Japan is growing." We look back. We're in the early phases. We're not looking back. In three or four years, we're going to look back, and we're going to say, "Of course, there's some content that is great for all ages, and parents can control that." Of course, there's 13 and up experiences on Roblox. Maybe someday there's 17 and up experiences. Oh, excuse me. I think I'm getting an avatar communication thing from Kiran and Mahesh. Let's see what's going on here. Sorry about that. Hey, Kiran and Mahesh, I'm joining our experience. Let's see. I may be flooded with my Wi-Fi, so let's see if I can get in. Give me one second. If we do not have enough, if I do not have enough Wi-Fi, I will jump on video instead. Let's see. Okay. Cool. Hey, what's going on? Guys, what are you? Kiran, Mahesh, what are you doing interrupting my main stage? Hey, Dave, nice to see you here. Meet Mahesh in his true goblin self. Hey, guys. Look at this. I can finally move my ears. Guys, this is crazy. Roblox is the only product that I leave alerts on. I'm on the main stage and in the middle of RDC. Here you guys are. What's going on? Yep. This is an exciting way to chat with your voice and avatar on Roblox. I can't wait to see Mahesh's reaction when I beat him in Jailbreak. You wish, Kiran. There are dozens of avatar types to choose from, and you can even customize your accessories. Actually, I think I need to change my beard right now. Oh, that looks perfect. All right. Okay. Guys, I guess thank you for interrupting me because it's great to see our vision of where this is going, which is many times a day we're going to communicate on Roblox. Thank you, Kiran, and thank you, Mahesh. Great job. Bye, everybody. Bye, Dave. Sorry. Sorry about that. I left the alert on for Roblox, so sorry about that. Okay. When I was growing up, this is what the specs for a great game engine looked like: Z80, 6502, raster graphics or vector graphics, actually. I just want to highlight how far we've come. I want to highlight to you our huge focus on performance, our huge focus on scale, our huge focus on reality, our huge focus not just on visual simulation, but on acoustic simulation as well. You're going to hear a little vision later from both Dan and Manuel. You're also going to hear from Morgan and Arseny. I do want to assure you this is the spec we're working on, okay? That's our game engine spec. It's 50,000 players. It's photorealistic. It models sound as well as light. We're excited to jump in and handle this. We have the team to do this. You're going to see early signs of it. We're relentlessly working on perf. A couple of quick examples. Parallel Luau is absolutely crazy. We have big servers in the cloud. We want more and more of your scripts, not just to run 5x parallel, but 10x, 100x. We have exciting demos with hundreds of NPCs that we think are going to be hyperscalable on a server. Everything in our stack we're working on to make fast our real-time geometry, join times. You're going to hear more about it. I do want to highlight we think part of our growth here is the perf and scale of our engine and infra. The other thing that we're really imagining reinventing is how we do creation. Just as we are rethinking how we all come together to communicate and to storytell, we are also rethinking how we create. I want to share a few of the ideas we have: Studio getting more awesome all the time, handling larger and larger teams, building a plugin and a cloud architecture. In addition to all of that, what I mentioned earlier is supporting your ability to make creation experiences as well. Just as we have essentially crowdsourced with self-serve creation, we want to take that up another notch and build these types of experiences. The vision is Roblox would never build this. One of you is going to build the project runway experience on Roblox. We are going to come in. We are going to use sewing machines. We are going to go buy the fabric. We are going to hang it on mannequins. We're going to wear it, and we're going to keep it. It's going to be persistent. This is a way that more than we even do, our users start to build creation with your participation. This is one of our big visions and dreams. It goes beyond creation into how we're going to shop. We would like all of the shopping to happen in your experience as well. That said, when we come back to Roblox Studio, the vision is how does a 1,000-person team work in Roblox Studio? How do we collaborate and power that together? This gets into some of the things that we've made enormous progress on. The first is our cloud platform. More and more, everything on Roblox will work in the cloud. You'll be able to access it. Studio will be able to access it. We will be able to have larger and larger teams. We have a vision internally someday for those of you that are using Roblox Persistence, we would like to be using the same Roblox persistent to power Roblox itself. Now, that's kind of wacky, right? You're using it. We're using it. Isn't that kind of weird or dangerous? It's actually super hyper-eating your own dog food. We need scale. We need reliability. We need millions of people to be able to use this. That combination of scale, reliability, and ease of use is really difficult to solve, but we like solving difficult problems. Someday we hope to say that. I'm really pleased to say that packages are getting to the point of being awesome. We've done a lot of work on this. We have an amazing team working on this. That vision of 1,000 people with an architecture for working on different pieces, for having artists working in a different place than coders, for protecting your code from the artists at the same time is also super important. Finally, where does Roblox run? How do we access it? I want to give a huge shout-out to our Roblox app team that has been creating our universal app. This is a huge move for us. This is the move from web launching a client to an app that is more consistent: PC, Mac, iOS, phone, tablet, Android, ultimately console, TV, and whatever. The exciting thing about the universal app is it joins a family of compatible technology. Across the board, we are seeing improvements in every metric with our universal app over our web launch to a client. You can also see a bit of a hint here that we're building an architecture where ultimately on other platforms, newer VR platforms, other consoles, potentially TVs, we have a great way of making Roblox work across all of that. One final point on our universal app thing and team, I'm going to show you one metric that you probably don't see in the press very often, which is, yeah, who talks about app launch times? That's pretty boring, right? That's not like all these pretty pictures and new features and whatever. Inside Roblox, this is what we focus on. We know raw performance, simple quality drives growth just as much as new functionality. It is consistent with that earlier game engine spec happening in a second. That's super difficult, okay? Super, super difficult. Okay. As we get close to the end here, let's go to the famous part of the segment. All right. Woo. For those of you that were here four years ago, we made a bold set of 10 predictions. We watch them every year. This is good practice. It's kind of like in a board meeting, looking at your plan and seeing how we did. Let's take a look at how our predictions have done. Here was our launch. Here is how we progressed through 2020 and 2021. I want to then go through them today just to do a little check-in. Checkmark is good. It means we already did it. No country more than 50% devs did that last year. Million concurrency did that two years ago. Major brand did that last year. This one is right on the edge, the national movie one. We believe one of our studios may be in conversation for a national movie, which is exactly what we wanted when we said this. We never want there to be a Roblox movie. We want there to be a creator movie. We do not think it is going to launch this year. Stay tuned, and we will see what happens. More than 50 million people code, yeah. Once again, with some of what I showed you, that is going to get much larger. 1,000-person, yes, yes, yes. The toy thing, we already said we do not think it is happening. 50 million a year already did it. This year, we are seeing 100-person plus studios. Cha-ching, welcome. The next one, I should not even talk about this next one. We should just skip the next one, okay? That is a no. Given all of our brand engagement, the movie, we still think yes. Then finally, this is arguably the densest collection of Tesla owners in the world. This could be called a Tesla user group. I'm not giving up hope for that one. In closing, community's healthy. We're working with you to reinvent storytelling and communication. We're growing all ages. We're growing around the world. The things people do on our platform are growing. We talk a lot about expanding the economy to support you. We have an amazing show today. It's such a privilege to hang out with all of you. Thank you for coming, and enjoy the show. Please welcome Roblox Chief Product Officer, Manuel Bronstein. Hello, RDC. It's always amazing to hear from Builderman himself when he talks about the vision for Roblox. Last night, as I was preparing for my slides, I decided to check out DOORS. Oh my God, that thing's scary. I think I woke up my wife twice as I was playing the game. If LSPLASH is here in the audience, can you please meet me afterwards? I'm stuck in door 39, and I need some help. I'm Manuel, also known as Aelarix. I'm thrilled to be here with you today. This is a community of trailblazers and innovators. It's always a joy to be with you. I want to start by just saying thank you. We wouldn't be where we are today if it weren't for all of you. It's a community unlike any other. Creators small, creators who are big, creators who are just getting started, creators who are mentoring others, people who are hobbyists, people who are professionals. Look around you. Really, you're surrounded by genius. It is really amazing for us to come to work every day knowing that we have such a passionate community. Truly, truly, thank you so much. This is an impressive slide. It is full of stats and milestones that we have all collectively achieved together. I am very certain that each and every one of you hit a personal milestone in the last 12 months. I want to take a moment to really recognize and celebrate the accomplishments that you have all brought to the platform. Let's give it a round of applause to all the amazing things that you have done. I also love seeing you take advantage of the new technologies and the new features that we bring in the platform. Seeing more close to a million experiences enabling spatial voice, or seeing people find their next partner or collaborator using the Talent Hub, it's just amazing. I want to give you a very big stat. In the first half of 2022, we have collectively achieved more than 82 billion visits to Roblox experiences. That is more than 500 times the annual visitors to all Disney parks in the world. It is such a big number. It gives me a ton of joy to see some experiences that were created just this year hitting impressive milestones. These five experiences collectively have had more than a billion visits in the first half of the year: Sonic Simulator, Barry's Prison Run, Please Donate, Weapon Fighting Simulator, and Starving Artist. Let's give it a round of applause. I want to share a fun anecdote of Please Donate. I went the other day to check it out. It's a pretty amazing use of spatial voice. One user came to me and said, "I'm going to sing for you if you can donate." The singing wasn't great, but I ended up donating anyways. He got my money. Similarly, we're seeing other impressive stats for creators who are building items for Marketplace. More than 792 million items sold so far in the first half of the year. Here are some of the top items: Black Messy Hairstyle by DD GameDev, Shadow Head by Maplestick, Mask of Madness, the Red One by Jor-El, Rainbow Flame Aura by Nezco, and Elf Ears with White Pearson. I think I have two or three of those in my inventory. It's really, really amazing to see creators breaking through this year. I'm sure that we're going to see a lot more. This year, we also have for the first time at RDC an invaluable part of our community. These are the Roblox video creators. Collectively, they have exceeded more than 64 billion views of Roblox content in the first half of the year. Their fun videos highlight the experiences, the items, the updates, the events that you all create. I want to thank them for helping Roblox content reach new audiences every day. It's pretty amazing to have them here. Welcome to RDC. Last but not least, I want to call out a special group of creators. These are people who go above and beyond to give back to the community. These are creators that help organize events. These are members of the developer council. These are level-up panelists. You all are contributing not only your time, but your knowledge to make our community better. I want to thank you for that as well. Yes. One of the most rewarding parts of working at Roblox is to have a devoted community of creators that are always pushing us to be better. We come to work to Roblox every day to build the most amazing platform for you. Sometimes, we may not get it perfect the first time. Your feedback, your candor, your help us actually make a better platform. I am very, very appreciative of the open communication channels that we have with the community. Let's take the example of our planned release for UGC materials. We decided in June 2021, we announced a plan to upgrade materials. The initial plan would fully replace existing materials and later introduce the custom ones. You gave us feedback. You told us that's not going to work for us. If you do it that way, many of us will need to rework our experiences. You also wanted more control. We took a step back. We paused and decided to work it out. In 2022, we announced a new plan, a plan that started with custom UGC materials with amazing material managers that would give you more control. We also decided to keep the old materials so you did not have to rework your experiences. Your feedback was amazing. You told us that we got it right. More exciting than getting the feedback is seeing the amazing creations that you are all bringing together with these new materials. Check it out. Some amazing environments and new worlds that are creating. I cannot wait to see what you all do with this technology. It's going to be fascinating. In that spirit, I want to share about a few things that we're working on to make the platform better. Let me start with the first goal. We want to build a platform that helps you realize your creative dreams. Our plan is to continue investing every day on the technology, on the tools, on the platform, until one day anybody on the planet, everybody on the planet can wake up and say, "Whatever I want to build, I can do that super easily on Roblox." You heard from Dave about the vision for the platform. You're going to hear later from Dan, our CTO, who's going to share a little bit of the tech and tools that we're bringing to market. I also want to share some of the product investments and the programs that we have to help us achieve this goal. Let's start with cloud and collaboration. This is a big one. We have a vision where several hundred-person teams come together to work in studios seamlessly, collaborating real-time in the cloud. They will be able to build experiences that are 100x, 10,000x larger and more complex. We also want to continue moving towards our open cloud, where all of our cloud services and the data and the APIs that power Studio and power your experiences are available for you from anywhere so that a next generation of toolchain and an ecosystem of rich tools can be built on the Roblox platform. I'm very excited about this. We also have pretty amazing creator events. The cool thing about these events is that those are events created by the community for the community. Developers come together, and maybe they listen to a podcast, or they prepare a Q&A. We have seen clothing design workshops. We have seen game jams. Since inception, we have seen more than 300 events happening with more than 16,000 attendance. These have been happening all over the world in countries like Indonesia, the U.S., India, and the U.K. We want to build more on the program. What you're seeing on the slide is two big investments that we're going to make. The first one is we want to bring events to Guilded so that you can host an event for the community in your own Guilded channel. The second one, coming in 2023, is we want to give you space inside Roblox so that we can have events inside the platform. I can't wait to see what you guys do with it. It's going to be very exciting. The second program that is super, super powerful in helping the community is our Level Up program. This is a very cool program where experienced creators, experienced developers are willing to share with the rest of the community the things that they know best: their best practices, their knowledge, their tips and tricks. We have seen Level Ups for level design, for economy and monetization design, for live ops, and also for marketing. Since creation, we have had 25 hours of content across roundtables and interactive workshops from creators like MiniToons, The Gang, mista, KreekCraft, and Wonder Works Studios. Thank you so much for all that work. We want to make these events better next year. What are we going to do? We are going to do live Level Ups so that you can interact directly with the community and with the people who are presenting. We are also going to start doing short-form write-ups that we can share with the community on a more regular basis so you can have best practices at your fingertips in a real nice cadence so that you can learn more about the platform. We're also going to be using our Guilded platform to actually host events and have the community come together. Now, the game fund. All right. There's some amazing trailer going on in the background. I'll let you watch it. This program was created in 2021. Our goal was to provide monetary support to creators pushing the boundaries of what could be done in the Roblox platform, people who were pushing with ambitious visual design, innovative gameplay, or using the latest Roblox tech. We announced the fund with $25 million to support the pitches that were coming from the community and also developers who were completely new to the platform. We have seen an amazing collection of pitches. I mean, you're seeing the trailers here. These are some of the first few experiences that were greenlit in the program and are going to come to market soon. I'm very excited to see them. We decided to up our game. We're going to add another $10 million of funding to the game fund to get it to $35 million. Applications are open. If you have great pitches, please send them our way. I can't wait to see them. Yeah. Once you build something that you're proud of, once you realize your creative dreams, the next thing that we know you're thinking about is, how am I going to grow my audience? How am I going to build my fan base? How am I going to bring more users to this experience? I'd like to highlight some of the work that we're doing in this space. You heard from Dave talking about Discovery. I know this is a super hot topic with the community. I wanted to share more with you. Our goal is very simple. We want to match users with content they will love. We also want to help new and promising experiences to find an audience and grow. The key to achieving that with all the content that is created on the platform is investing in personalization. This is an area that is a long-term investment for us. We will keep investing in personalization for many, many, many years to come. People often ask me, so how do the recommendations work? What we're trying to do with the recommendations is that we're trying to look at signals that predict whether a user will value the experience that we recommended to them. You may ask the question, so how do you measure value? We look at retention, engagement, and monetization. Those signals help us understand if the experience is going to be a good recommendation for the user. When you're thinking about recommendations, your best bet to improve them is to work on retention, to work on engagement, and to work on monetization. We have other great tools for you. Once an experience gets discovered, our continued playing and favorites help people re-engage with that experience. We also have the friend list and the friend activity where people can find a new experience just by looking at what their friends are playing. We have a lot of things actually helping new experiences to get discovered. This is super exciting. We wanted to make sure that new experiences got discovered on the platform. How do we do that? When new experiences and promising experiences launch, we gradually start exposing them to people. As they show good results, we continue to expand. What this graph is showing is that out of our top 150 experiences, 7% of them were created just roughly in the last 90 days, which is showing you that new experiences are getting opportunities to get discovered on the platform and thriving. We want to keep investing here and continue to make it better. Last but not least, in the discovery section, we are working on improving discovery through sponsor experiences. Sponsor experiences were usually in the discover page. Very recently, we ran a test bringing sponsor experiences into the home page. This is where most of our traffic and our audience is. We saw an increase of 367% in ad conversion. I'm sure that some of you have seen the click-through rate improving in the home page in this experiment. We are going to keep iterating because we want to make this tool really useful for you so that you can get your content promoted and discovered on the platform. We also have these amazing Video Star programs. As I mentioned, it is a vibrant community of video creators that are invested in Roblox. They collectively average more than 10 billion views every month. That is huge. 10 billion views. It is a big, big, big number. When we talk to users, this is one of the most common ways in which people discover experiences, events, and updates on the platform. Some of the Roblox Video Stars have also become creators inside Roblox. They have created their collections of items. They have built their studios. I cannot wait to see what they do. We will continue to grow this program in two ways. One, we want to expand it to new geographies so that we can reach more people all over the world. Two, we want to expand it to other platforms like TikTok to reach our audience there as well and get more people coming into your experiences. Last but not least in this section is how do we think about communities? One of the things when I talk to you in the audience many times and ask questions, what's missing, I hear that one of the things that you want is to be able to engage and build your community base, your fan base. How do you communicate with them? How do you grow and nurture that community? I know that today you may have had to do that work outside of the Roblox platform. That's not ideal. With Guilded, we're working really hard to build the most amazing community platform for Roblox so that you can communicate with your audience, share when you have an update, share the news, link directly so that they can join the experience right away when you're actually announcing an update. You want to have channels to receive feedback, for people to submit bugs. You may want to have a calendar of events so that you can invite people to those different events. You may want to manage support issues. We're going to be working hand in hand with you, with the Roblox and Guilded team together. We're going to be reaching out for your feedback and your ideas as we continue to build this. Stay tuned. This is going to be a pretty amazing area of investment for us. You built an amazing experience. You grew your audience. Now you really want to grow your business. You decided that you want to expand your studio. You want to hire more people. You want to maybe build a second team that is building a second experience. We want to help you along that journey. The way we're going to do it is by giving you the insights, the information, and the tools so that you can build your business and better monetize. Let's start with the analytics. There has been tremendous progress on creator analytics. Our goal is to provide you with the insights to make decisions that improve the engagement, the retention, and the monetization of your experience. In Q1, we added the ability to break down retention by region, age group, and device so that you can see how your experience is working and understand if you have gaps in retention in any of those areas. Now, in the coming month, we want to show you a benchmark so that you can understand how your experience compares to other experiences with similar player traits and understand if you have gaps in retention. I believe that information is going to help you to improve and get more information about how you need to work to actually reach some impressive milestones in retention. Secondly, it's an area that I've also been hearing a lot about. Once you get retention, you want to grow. One of the things that I heard from the community is that you want to know where your traffic is coming from. We call that source of traffic. In the coming month, we're going to introduce what we're calling the acquisition dashboard. You're going to be able to see when people come to the experience which links, which areas of Roblox sent them or directed them to your experience. You're going to see if that traffic is organic. You're going to see if that traffic was through advertising or other sources. This is an area that we're going to continue investing in because we know that by having that information, you can improve your funnels and you can actually make your traffic flowing into your experience even better and grow with it. We also, in June, introduced the monetization dashboard. This monetization dashboard allows you to see where your revenue is coming from so that you can optimize your experience and continue to grow your business. In the next year, we want to introduce what we're calling the real-time dashboard. It's going to allow you to look at the health of your experience in real time. You're going to be able to see at frame rate, crash rate, CPU utilization. You're going to be able to make better decisions and choices of how you're building your experience. With all of this, you're going to have great information at your fingertips. This is an area that we want to continue to invest to give you even more insights and information as you grow. Now, let me talk a little bit about Marketplace. As Dave mentioned, we have a huge and growing transactional economy on Roblox. Just in June of this year, 2.7 million creators earned Robux on the platform. We asked ourselves the question, how do you make something that is good? How do you take it to the next level? Two answers. First, we want to open it up to more and more and more creators. We will not stop until every single user on the platform can eventually publish and sell on Roblox. That is our goal. The second one is we want to introduce true market dynamics into Marketplace. What that means is that we want to mimic how real-world economies work and bring that energy and that feature set into the platform. With that idea in mind, we are going to be releasing many cool things in the coming month. The first one is we want to introduce scarcity. The idea with this is that you're going to be able to create UGC, limited, and collectibles. Everybody on the platform will be able to do them. These are items that will accrue value over time. Now, once you have collectibles and limiteds, the next thing that you want to be able to do is be able to resell them and trade so that an item that is sold on the platform continues to accrue value over time. As subsequent transactions happen because people are buying and reselling, the creator of the item can continue earning. You will not only earn from the first time. You continue earning on the subsequent transactions. That is going to be very huge. I think it is going to drive the economy in pretty meaningful ways. The third one is one that I have been hearing also from you loud and clear. You want to have the ability to offer free promotional items. We want to bring that capability so that any Roblox creator can publish free promotional items. Last but not least, we are going to continue investing in sponsor items, which is very similar to sponsor experiences so that you can promote your items in the Marketplace. I think that the four things here, plus more creators coming in, is going to supercharge Marketplace. I cannot wait to see what you guys are going to do with it. As big and as impressive as our transactional economy is, we still want to invest in new sources of revenue for the creator community. As Dave mentioned, we are excited to work on this concept of immersive ads. These ads are native to 3D immersive experiences. They are fun. They should be interactive. They should be safe. They should not be obtrusive. They should respect people's privacy. We also want to give you control as developers where you place them, how you run them. With more than 58 million users coming every day and billions of hours spent on the platform, brand advertisers and even experienced creators are looking for ways to reach this audience in novel ways. We have a great lineup of things that we're working on that we're going to be iterating with along with you, the community and brand marketers. Think about portals that you can navigate from one experience to the other. Think about the billboard that you just saw on my video where the video is promoting an item. By tapping on that billboard, I can choose the item that I want to buy, try it on, complete the transaction without leaving the experience. Think about a scenario where brands want to embed content inside experiences. We have a true mechanism to make that seamless and direct. It is going to be very, very amazing. It is going to be something that we are going to start, like usually, iteratively working our way up. I see a huge opportunity for all of you here. I am excited to see what we build. There is a lot to be excited about. You heard me chatting about the work that we are doing to help you realize your creative dreams. You heard me talking about the work that we are doing to help you grow your communities. You heard me talking about the work that we are doing to help you become a meaningful business. There is really no doubt in my mind that this world is ours for the making. We are going to make something amazing together. Thank you so much. Now I'm going to have Dan Sturman come talk to you. Please welcome Roblox Chief Technology Officer Daniel Sturman. Hello. Wow. This is just amazing. This is my second RDC. I was just blown away last year. That was like, I don't know, a third the size of what we have here. This just blows my mind. I'm Dan Sturman. I'm the Chief Technology Officer of Roblox. I'm also known as the TheKingsHappyFool. I have the very distinct pleasure of taking all of you through what our engineering team has been doing over the past year and where we're going in the next. We are over 1,000 strong now in engineering. It has really grown as a team. Part of what makes this my really distinct pleasure is all 1,000 of us engineers, we all get up every morning psyched for what we can do to make you more productive and what we will see you guys do with the crazy stuff we produce. I will call some of those out during the talk. Now, it is 1,000 folks. We have actually gotten a lot done. There is a lot in this presentation I have to take you through. In fact, my prep pretty much was three shots of espresso and a Red Bull chaser to get ready for this. I am going to go kind of fast. In some cases, I am just going to give you a word cloud with a bunch of things we are working on that I will not be able to go into today. Do not worry. They'll all be talked about in the sessions later on this week. With that, let's dive right into it. Last year, we talked about things at RDC that we'd be doing this year. I want to give you an update on how those went. Of course, layered clothing, one of our biggest releases of the year and really a game changer. It's a game changer from a tech point of view. No one else has really done this before. It's also a game changer for how you can be expressive on the platform and how you approach it as creators. We now have the ability that our users can take, mix, and match any items and layer them on their avatar. In fact, over half of our daily actives are using at least one layered clothing item. Just curious, how many folks here have equipped layered clothing stuff on their avatars, right? You saw my avatar at the beginning. Definitely, it's using that. It's really cool. It is also a game changer for creators because we're going to be introducing, as you see here, all sorts of cool, strange bodies. You can design an item once, and it can work on any avatar. You are not pigeonholed into some sub-market, but you can just make your items. They can be awesome and let our users be creative with it and figure out where they want to go. OK, facial animation. You saw some of this with Dave's awesome demo before. This is a game changer also in avatar expressiveness. You will hear about it a lot over the next two days. You will see a lot of things over the next year and more on how we are taking this elemental technology and applying it all over the place. The general idea is heads are equipped. We call them dynamic heads with the ability to do facial animation. It can be any head. We have introduced this in Studio so you can create and publish new heads to the platform. It will be our classic heads. It will be dinosaur heads. It will be alien heads. We also launched what we call auto skin transfer. This is stuff like the beard you see in there, it can be eyelashes, eyebrows. It can be hair. Though personally, I am a little less invested in the whole hair thing. I think a lot of you will be, right? Anyway, this is going to be an elemental tool that lets us do all sorts of expressiveness on the platform. Chat with spatial voice. You saw the demo before with just doing the whole thing with dynamic heads and so on. Spatial voice was introduced this year. We have launched it. We're really thrilled that so far we have 2.5 million 13+ verified users signed up for spatial voice. It has been incredibly popular. We have up to 50 player limits in experiences. We're looking to scale that to 100 and more over the course of the year. This is one of my favorite features. I'll explain why in a minute. This is about our new materials and the way you can use them and customize with them. The idea here is we introduced this concept of PBR materials, or physically based rendering materials. They just have, in a sense, layers within them so you can get light reflection much better, more realistic use cases with these materials. You can see, for example, from Infinity LT how much more realistic that looks. This is why I love it because I am a terrible, terrible artist. Absolutely horrible. I might be able to code, but I cannot do art. I was able to go into my own experience, apply these materials. All of a sudden, it looked like I had some skill. I just really loved that. What we also realized as we were rolling this out is all of you have put a lot of thought and time into what your experiences look like. We can't just roll out a bunch of new materials. We have to give you absolute control on how materials impact your experiences. We came up with this material manager, and they're able to do custom materials. Now you can keep classic Roblox materials. You can grab the new material set. You'll soon see new material packs on the Creator Marketplace. You'll be able to even download or create things from the web and bring in your own materials right into the manager. You have absolute control over how this works. It's pretty cool, right? OK, Dave mentioned this as well. Real-time CSG is one of the things we're really excited about long term. In fact, this is the start of something you'll be hearing about called in-experience creation. We have a bunch of features that are building up towards this. The idea behind in-experience creation is you do not have to be outside of Roblox or in Studio to create cool stuff. You can be doing this as, in a sense, a player in the game, but you are creating. Maybe you can take that onto the platform for something on your avatar or your friend's avatar. Real-time CSG, we went. With CSG v3, we are now 100 times faster than we can do. You can see this with this lathe example. This is the old world right here. It is clunky. It is slow. Now look how smooth this is. The lathe has been something internally we have been focused on forever. Yeah, no, it is very cool. It is very hard. Speed, performance is a feature. I talked about Project Harmony last year. We are continuing on that. This is going to be a long journey for us, but one we're super excited about. A lot of features up here. I'm not going to go into them all. Dave mentioned things like Parallel Luau and 10x faster raycasts. We're also doing spatial query filtering at about 3x the speed now. Humanoid ladder checks about 4x the speed. More than just faster performance, it's also getting dynamic. We want to do the right thing, whether you're on an old hand-me-down iPhone 6 or if you are on the latest, greatest GPU-enabled gaming PC and have the right thing happen between our servers, your experiences, and you, the creator, should not have to worry about it. That's our job to make that happen. Another multi-year journey. In fact, when I got involved with it, probably my first week at Roblox, is this idea of the Universal App. We have seen great results as now we have rolled this out to Windows and to Mac on top of, of course, iOS and Android. This is going to give us a bunch of new capabilities. First of all, it gives us one system to bring to any platform. From an internal engineering point of view, we can be much more efficient at tackling new platforms and then optimize them so we say, oh, this is the part that has to look different on Windows. We will be able to tweak that. This is the part on the phone. Meanwhile, bringing all the goodness for the things that are common very quickly to the platform. It is also our biggest dog food to date. We've built all of this in Lua. We're like, if our creators are living in Lua, we need to live in Lua. That's how we ensure it stays fast, effective, expressive, and we can continue to optimize and understand the world you're living in. OK, Studio. You all spend countless hours in Studio. It is important we continue our investment there. We've done this thing this year. We're calling the Studio design refresh. It starts with new fonts, new icons, basically a new look and feel that I think is really cool. It works for high-definition monitors, which I'm sure most of you have access to at this point, right? If you have that great display, why not use it, right? We want Studio to be able to support that. This is rolled out to about 30%-50% of you at this point. Has anyone seen it, using it? It's random who gets it at this point. Yes, there's some of you. That's rolled out. With it also is this whole idea of a complete rework of the docking system. It's much more flexible. It does the right thing. It's so much more intuitive. A responsive ribbon. The whole thing scales based on how you move the studio layout around. It doesn't do dumb things like clipping off important controls on the screen. The whole thing is going to be faster. It's more customizable. We think that makes you more productive. I hope you find the same thing. It'll be rolled out to everyone shortly. OK, that was the stuff we talked about last year. There's been a whole bunch of things that either were being sneaky and we didn't tell you about last year, which is really code for my engineers had no idea if they'd land it, right? Or we picked it up partway through the year. I want to take you through some of those items. A few, first of all, just graphic stuff out of our great engine team. I want to start with lighting occlusion. You can just see the effect here. When the moon is bright, you can see that glare reflecting off those leaves. When it gets clouded by the clouds, the right thing's happening. It gets much more muted. Same thing with the sunny and overcast example on the bottom, where on a sunny day, those shadows are sharp and precise. When it gets cloudy, it gets much more diffused. That is the effect we all want. Again, something the engine should do for you. You should not have to spend a lot of time tweaking that. Particles are really cool, whether it is for fire, smoke, arbitrary special effects. We think you can do a lot of cool things with particles. You can see what EmilyBendsSpace did here with fire, kind of going from the old world to what we are able to do with this awesome new, really hot fire in the new example. Also, particles can change as they go. Fire can turn to smoke. You can have bubbles that grow and pop. It is a lot more flexibility for those kind of special sort of tweaks within your experiences. This is something I am really excited about, both for what we did this year and the future of this. You'll be hearing a lot from all of us over the next few years, this whole idea that deep learning techniques can start to really accelerate creation on the platform. I mean, a lot of this is going on Roblox, deep in our technical teams and our research organization, but also in the world that we're in a mode where this tech is just coming to the point where it can do tremendous things. You can see right here, look how much easier it is to go and do a dance to animate an avatar. Or if you're like me, get one of your more coordinated friends to do a dance to animate your avatar, right? Then just transfer it directly. Of course, you can fix it up and tweak it as you go. Notice this is applying to all sorts of different avatar bodies and all sorts of different avatar faces. Same thing with faces, by the way. This is where those dynamic heads and facial animation are so important because now we have a common standard. We can build something like this where deep learning watches your face on camera and can put it on any head that's rolling out under this new system in Roblox. I'm really excited about these two. You can imagine putting together a concert in one of your experiences where you capture a singer dancing on stage just with these two tools. No fancy mocap suits, nothing like that. It's accessible to pretty much anyone. OK, so we've also heard you about how difficult it is to kind of tweak in the animation editor and that you get really jerky animations with kind of straight interpolation. Now, unfortunately, I say that. My video is a little jerky itself. Ignore that. If you went and played with this, these curves we've added into it make things much smoother. You can, again, control them. It's all about control for the creator. You'll be able to take the curves, tweak them, get exactly the interpolation you want between points. It means you don't have to make a bazillion points in each of these animations. Another concern we've heard a lot from you, nothing's more frustrating than you've got a great experience. You made it successful. A bunch of annoying copycats come along and grab it and build these things that just start to steal your traffic. We have really been starting to focus on preventing experience cloning. We have a bunch of new models that are working really well in this domain. It is, again, going to be something we work on for quite a while, though. For example, here, you can have some clones. We now have a new pipeline. We have taken down 40,000 clones just since we deployed this stuff in late April. It is already engaged. We are going to keep ramping it up. We are going to keep getting new models. Our goal is to make this cloning a thing of the past. There is a lot of other stuff we have been doing. There is one of these word clouds. I wish I could go into all of them. You can find great Roblox devs around the place. This week, they'll be in talks and so on. They'll go into more details of these. OK, now I want to talk about a set of things that we continue to work on. What I mean by this is we've done some good progress this year. We view these all as big, multi-year journeys. They're kind of big things. We'll never be done with them. We're really excited. We want to give you an update on where we are on them. I mentioned the universal app before. It's something we promised we'd deliver this year. Now that we've done that, we're on to the next big phase. First, let's make it hum. Let's make it really perform. Real big improvements in app launch time, drops and crashes, particularly on the mobile devices where it is so tricky, doing things like seamless app background in iOS so you can respond to a text without losing Roblox. All incredible improvements. We are going to keep pushing on performance on the universal app. At the same time, we are also saying, hey, we have got this universal app. Any improvement we make here will go to everyone. New and experienced menus are being worked on as the next step, new ways to navigate in the app. Some will make more sense on different platforms. Setting ourselves up for more platforms. Here, I am not going to go into detail. You should know now that we have got the universal app, we are going to take it out for a drive. Platforms are much easier for us to port to. We're going to be looking particularly at more consoles, touching on things like VR. We now have the tool we need. It is not a massive rework of all of Roblox in order to get onto one of these new platforms. In Studio, Dave mentioned this before. Manuel mentioned it about collaboration being so key. Now collaboration is just one click away within Studio, and you're collaborating instantly. All new places by default are created as Team Create. We are moving in some directions that are going to make it really so much easier to collaborate. The first one is my favorite, which is a detailed version history. I'll admit, I even need this one. I'm just working by myself because I'll mess things up and have no idea how I did it. You can see what's changed, who's changed it, why is your experience looking the way it is, and then a bunch of tools to help you understand what your collaborators are doing on the platform. Where are they entering? What place are they in? What are they working on so you don't step on their toes? All that being visible kind of from a 3D point of view. I think this stuff's going to be really cool and enable really big teams. One thing you should know is we're going to continue—I don't have a lot of time here. We're going to continue a heavy investment in studio. We've got to make this tool the best in the world. We also already think it's pretty rocking. We know we can't stop. You guys are all a very demanding audience in a very good way. There is going to be a lot of features here. The ones in purple are ones we launched this year. Blue are things that we are working on for next year. The main thing you should take away is there is a big team behind this. We are really serious about it. OK, and packages. I told you last year under the section of things we did not deliver, hey, we kind of have to scrap packages and go back to the drawing board. That is exactly what we did. We basically threw out the back end, rebuilt it, rebuilt the user experience. It is now so much more stable and so much easier to use. No errors. Now we are able to do things like, say, hey, let us put packages into the UGC Marketplace so you can sell them to each other. We have the system we're looking for that allows us to share collaborative content across the platform, whether that's you guys on your own teams just sharing it with each other. I created the tree we're going to use. I'm going to tweak it as we go, whether it's sharing it with each other through the Marketplace or other mechanisms, or it's Roblox saying, hey, we have some example content we want to share with the community. How do we get it out there? We have a great way of putting it out there, all very consistent. As the next step on this, we're moving to this concept of package interfaces. These will be parameters or properties that can be attached to a package where if someone picks it up, they can just tweak the package by the parameters, the interface that the user intended, rather than having to dive into the guts of these things and fork them. A lot more flexibility in how packages can be used. I talked last year about Open Cloud, how important that was to us. Dave talked a lot about the cloud. We want to put everything there. We have made great progress on it this year. This year, we opened up the data store API. We added a cross-server messaging API. We have added group API key support. It is embedded now with OAuth 2.0 support for authenticating yourself through these APIs. All really cool stuff that opens up all sorts of things for you guys to build auxiliary tools. Next steps, where we're going this year, a really big one is the full asset API, including the ability to upload kind of from anywhere, any asset type. I'm really excited about that. We're also opening up memory store API. You can see uses for a back end system, helping coordinate your matchmaking, for example, which is a great use of memory stores. This is really cool. Back to the inexperience creation, inexperience authentication. A user will be able to authenticate themselves from a platform point of view while inside one of your experiences. It can just hit that API and do it. A lot of power in all of these things. We're just going to keep going. Docs. If you're anything like me, I can hold a thought in my head about 10 seconds, right? I'm always referring to docs whenever I'm coding, particularly when I'm coding on Roblox. We heard loud and clear from you that our docs were not where you wanted them to be. We've put a lot of investment here. It started with a revamped doc site. With that, behind the scenes, we've really scaled our internal docs team this year. There are a lot more people on that team. You need people to get this done well. We're now focused on things like cloud APIs, getting those well documented. They'll continue to go. Over the course of the year, you're going to see us translating to 15 languages, so they're available in the languages you're using. Beyond that, this is a really cool thing. We're going to open source all this so that we can take contributions to the docs and fix the docs directly from the community. Because I know so many of you said, I see you got this wrong. It could be so much better. Here's my idea. And there's no way for you to give that to us. So we're going to enable that. Yeah, right? Good docs make everything. We have this vision around a Creator Marketplace. The idea behind the Marketplace is that it is the way you can all share and, to some degree, profit from great creations beyond just your experiences with each other, building a community. That's not going to work if you guys do not trust this Marketplace. We're hard at work making the Marketplace much safer. We've taken down almost 50 million malicious or copycat assets just in 2022. We've gotten to the point now where spam has been reduced by 90%. Insertions are now almost—we're so close—99%, almost so close to 99% safe. That's up from 83% a year ago. We're going to keep pushing on all of these stats. We are not happy with these. We're going to keep going. It's a huge improvement. It's telling you where we're trying to go with this thing. Right? As some next steps, we're going to make it so you can know when a verified creator—and that's ID or phone verified creator—basically has created one of these. You kind of know who you're getting it from. It's at least someone who Roblox knows who they are. We're going to really go hard against code plagiarism over the course of the next year. All your users basically communicate with each other and love your experiences through chat in the experience. It's just a big part of being on Roblox. It's a social platform, right? One of the biggest challenges we had used to always be this trade-off of, well, we need to keep them safe. We need to do filtering. But then they get less expressive. This was no fun for anyone. We completely revisited our models and adopted new techniques here where we found we could both make chat safer and less restrictive. It's a win-win. This year in English, 27% less filtering while safety metrics went up at the same time. We have similar results for Spanish, Portuguese, French, and German. As a next step, we're going to push even harder on that. We think we have another easy 25%, again, reducing the amount of filtering while improving safety. We are really excited about where this could take the platform as a social platform. Look, we have heard loud and clear from you that it is just getting difficult to create a really competitive game on Roblox because of cheating. It is a focus for us. We have a good team on this. We have a long way to go. We are changing direction, though. In the past, we were all about what can we do to the client in order to make it harder to cheat. We are going to continue to do some of those things. We do not have control of the client. Once they have the code, they are making the calls. You can do anything to it if you are clever enough. We're really starting to turn our focus as well to server-side cheat prevention. We think this can really help an incredible amount. I almost put the email address of the manager for our anti-cheat team on the slide. I was told that was not cool. We're going to make sure we're talking to all of you about that. In fact, along those lines, one of the things we realized we need to do is this needs to be a partnership. We're going to be spinning up teams to build specific developer tools. We're going to be talking to a bunch of you about what tools you need to work with us so we can combat this together. Lastly, and this is a broad thing, safety-wise, we're really starting to go after alt accounts. If we find someone behaving badly and we ban them from the system and they can just pop in again, that does not do any of us any good, right? We are really getting better at understanding, yeah, you think it is all. We do not think it is all. We know who you are, right? We are going to be focused on that from a prevention point of view. OK, a few quick things. There is some stuff we always do behind the scenes. Dave touched on this, both from scale and reliability. You do not see these as features in the app. They are very important. We had a really big outage Halloween last year. I am sure you all remember. Another one on May the 4th. It is an area really focused. Our focus has been really, how do we avoid the biggest, most catastrophic issues? In the past, we centrally homed all our data in Chicago. That worked for us at Roblox to that scale. Since then, we have built a second full data center in Ashburn. We can manually failover from Chicago to Ashburn, allowing us to deal with anything catastrophic that really happens in Chicago. Really important step. We are going beyond that to a world that is what we call an active-active cellular data center, where both data centers are running at once. It is not a matter of failing over. We have capacity in both. We are breaking these data centers into smaller units we call cells that are completely isolated units so they can fail independently unless the whole building blows up or something along those lines. We have a lot of room there. Reliability is going to continue to be a focus of the team going forward. We're doing all that while we're massively scaling Roblox. Our edge capacity, all the PoPs, data centers, edge data centers out there have grown 72%. Our bandwidth in and out of Roblox is up to 3.1 Tbps. That's a 55% growth year over year. It's basically like we're a modest-sized telco company at this point. Our overall server capacity has grown by 113%. It's more than doubled. We're just going to continue to scale it so there's always the capacity you need. You never need to worry about the capacity being there. That took some time. That was what we did this year. I'm now going to jump into just some sneak previews of what's coming up next year. For most of it, I do not have the time to talk about it. Another word, cloud. I am going to dive into a few of these. Approved account authentication. Nothing is worse for a user than when they lose their account, it gets stolen, hijacked, or they cannot otherwise recover it. That is just an incredible bummer. They have worked all this time in your experiences to build up their character or whatever it is, and then they lose it. We are starting this past summer with more prompts for either cell phone number or email so we can help them recover. When they do not have that stuff attached to their account, there is almost nothing we can do when they call us up in tears saying, I cannot get my account back. How do we prove it? How do we know? And so on. Moving beyond that, though, once we have those, we can actually move it. We could take passwords, which I personally hate, out of the system. You can just sign in with phone or with email. We think that's going to be a lot more secure than passwords. Passwords appear on the dark web. They're just kind of old school. Most users repeat the same password over and over again. We're trying to get away from that. Another thing that's been a big problem is this whole idea of web cookie theft. In early 2023, we're going to roll out hardware-backed unexfiltratable tokens that replace cookies. The idea here is all our devices have the ability to create and store safely these tokens. If you use that, it's not something a hacker can grab or even ask you to give them. It's so much more secure. This is a fun one. I just love this because you guys really taught us this one. Let me just set this up, a little story. It's Saturday morning. We're hitting a new peak load. All of engineers cheering, yeah. One of you goes and says, now is a great time to push a new update to my game to close the whole thing. We'll have a million players want to rejoin right away. At first, we just thought you hated us and you wanted to torture us as we get these thundering herds storming into our matchmaking system and our scheduling system. We said, wait, no. They're geniuses. What they've figured out is this is the best way to promote a new feature, a new capability. They're amping the crowd. This is wonderful. We got to support it. That is one of the things we're focused on this year. Before, the only way we could deal with this was this really haphazard matchmaking, putting people back in random instances as fast as we possibly could. We're now going to make it so we do an outstanding job of matchmaking. Our goal this year is to be able to do 1 million matches and deployments in under 10 seconds. We're going to do that by basically taking a snapshot of where you're at and using that for 60%-80% of the work, giving us plenty of capacity to handle the rest of that. Like I said, 1 million in 10 seconds is where we're aiming on this. Go and close all the games you want and restart them. We'll handle it. I know reviews for moderation can be a real pain for you. We're really working on both speeding it up and upping the quality. To give you an idea of the scale, each and every day, we're doing about 500,000 image moderations and about 15,000 audio. We're starting again to deploy new tech that's making this much more efficient. At this point, the new stuff is doing an auto review of about 1/3 of the images. We're starting to deploy this for audio as well. We expect to ramp this up and have a lot more be auto machine-based so we can go much faster. You're not waiting around for that image to be available to use in your experience. Some improvements in Creator Marketplace. I mentioned how we want to make it safe before. Then we want to go beyond that. Any verified creator will be able to sell plugins in the Marketplace. We want to go, again, to packages and models being available for sale so you can share work with each other, again, in a super safe way. We are going to allow rich media in the description so you can actually understand what the heck this thing is as you look around and have a great description for it. Also, something we are starting to experiment with, we are really excited about, is the idea of more flexible payment methods. A one-time purchase works for certain sorts of things. Let's say it is a great plugin that really improves productivity. That should be a subscription model, maybe a freemium model. We are going to enable these in the Marketplace as well. You guys have a lot of flexibility and a whole new way to share content with each other. Connecting with your users is also really important to all of you. I mean, they're your customers. They're the ones who make you guys successful. One of the best ways to do that, we think, is some sort of notification to them. We're going to be looking to build out the ability to have custom notifications even when someone isn't in your experience. Imagine, hey, your best friend just beat your high score. There's a new magic egg about to hatch. You can let them know about that even when they're not in your system, in your experience. Also, custom social invites. Imagine not a general Jeremy wants to be your friend, but Jeremy just knocked on your house's door. Do you want to join him? Kind of bringing people back in on an event basis. We think it will allow you guys to really take this and drive up engagement in all of your experiences. Good stuff. One graphic thing I want to call out, it's been a frequent dev complaint, is water sticking jello-like in our shorelines. You can just see the before and after. I do not think I have to say a whole lot how much better this gets. You heard Manuel and Dave talk about immersive ads. There is a lot of cool tech behind this. Of course, matchmaking and teleports got to get really fast. Understanding when someone looked at a billboard and how we count that as an ad is really important. There is a lot of stuff we are going to be building for this. We're very excited, as the others mentioned, about what you guys can do with this in your experiences in a uniquely Roblox way that just, again, drives engagement and spending on the platform. I want to end with chat. Dave did that awesome demo. I'm not doing a live demo. He's gutsier than I am, I guess. What I want to call out here is this is the culmination of a few years of really hard-hitting technology. We have layered clothing. We have facial animation. We have spatial voice, all coming together in a multi-year, in a sense, technical effort that I think you'll only ever find on Roblox. It's going to be really unique. It's really unique to this 3D platform. I'm incredibly excited to see what all of you do with these capabilities. With that, I wanted to thank you. I wanted to thank you for putting up with this incredibly long list of things I just ran through. I also wanted to thank you for being the incredible community you are. On behalf of all of engineering, you guys are the reason we get up every morning. You guys motivate us. We know that you'll hold us accountable. You'll also be appreciative when we do something great. You'll stick with us when we have a little slip like we did around Halloween last year. Again, we're extremely thankful. I'll be around the next two days. I hope to talk to as many as I possibly can. Please welcome back Roblox Founder and CEO David Baszucki and Microsoft's Chief Technology Officer Kevin Scott. All right, man. Welcome. All right, everyone. We have quite a treat for you. Kevin, thank you for joining us. Thanks for having me. Kevin really embodies a lot of the values and what we're doing here. There's the interesting coincidence in that Kevin has been at AdMob, has been at Google, has been at LinkedIn, has been at Microsoft. A fun fact for those of you that remember, we used to have that office in downtown San Mateo. We worked in the same space at one time. Yeah, it's really shocking. I had no idea Roblox was at 60 East 3rd Avenue in San Mateo. Yeah, small coincidence. We're going to talk about all things innovation and all things visionary. Maybe just to kick it off, Kevin, how did you get into this? How did you get into engineering? I think the same way that a lot of folks my age did. I'm 50 years old now. I started becoming interested in making things when I was 10, 11, 12 years old. The most interesting thing to make back in the early 1980s were programs on the personal computers that were emerging at the time. Interestingly, the thing that most captivated my imagination were video games. That's the thing I wanted to figure out. You're going to get spot. How do I go make my own video games? These are such fascinating things. It just went from there. I learned on Apple II. What did you learn on? I learned, so Apple IIs, TRS-80s were what the school had that I was at. The very first computer that I had for myself that I saved up for two summers to buy was a RadioShack Color Computer 2. Ooh. Yeah, so like a beige case with the chiclet keys that you hooked up to, like in my case, a 13-inch television. And you saved and loaded your programs on to and from cassette tapes. So super high tech. My very first place, the very first place I ever programmed a computer was in a teletype terminal in our middle school. The storage medium was a paper tape with holes in it. I hope that does not date me too much. It sounds like. OK, let's talk innovation, right? AdMob went on to be one of the leading mobile ad platforms in the world. You see what we are doing. We are trying to innovate there as well. I think AdMob was arguably hyper-innovative at the time. Was there a time when you realized what we are inventing here? There is something here that had never been done before. Yeah, I mean, at AdMob, our goal in life was not to stuff little rectangular ad units onto small screens. The problem that we were trying to solve is that, and this was in 2006, 2007, before the iPhone was even on the market and Android was a different thing than it is today. It was just really, really hard if you were a developer or content creator to get something done on a mobile device. You had to go cut a deal with a carrier, convince them to distribute you on their deck. They controlled everything about the experience and access to services on their phones. Even if you did find a way to get a carrier to give you distribution, you had no real way to monetize. You had to build all of it yourself. That was the problem we were trying to solve, how to give developers and content creators a way to get themselves distributed. Once they had an audience, how do you get paid? Yeah, I think we were, by far and away at the time, the biggest mobile ad network. I joined Google when we were 300 engineers. We were 7,500 engineers four years later. Google grew crazy fast. AdMob's the fastest growing thing I've ever been involved with. We were doubling every five months. When you joined Google, what was it like in those early days? Chaotic. Really? Fun. We had so many interesting problems. The environment was structured where you could just sort of gravitate to the thing that was most interesting. I had spent most of my 20s thinking that I was going to be a computer science professor. I took a leave of absence, which turned into dropping out of my PhD program to join Google because I wanted to do something that was going to matter to more than just 100 people. Once I got to Google, I just remembered why I had left academia. I just tried to find the things that were going to have the most impact, which were not necessarily always the sexiest things. They all turned out to be technically very interesting. When we talk about innovation, and a lot of the people here are creating studios, trying to do things that have never been done before, trying to have a, and I would say on our platform, we time and time again see things that have never been done before. Roblox, in a sense, is an innovation culture in that studio space as well as our own engineering space. Are there any tips that you might have for the studio creators around innovation culture? Yeah, I mean, for a whole lot of years, I've thought about innovation as just sort of where creativity meets user need. And as a consequence, if you want to build an environment that encourages innovation, you really want to think about both of those. How do you build an infrastructure, a culture, an ethos that lets you figure out how to discover that user need? Everybody does it differently. Some companies create user need by building the products that they can build and then marketing them super effectively. Some companies discover the user need by shipping things really quickly and getting a lot of data back from user interaction with them and then sort of iterating to great. Whatever it is, you have to have a hypothesis about how to discover that user need and then just sort of set your culture and your infrastructure up to go explore it. The same thing with the creativity side of things. Creativity, I think, I believe you and I chatted about this. I really believe it's like a muscle. It's not always that every creative thing you do is an innovative thing. Just trying to be in a process where you are creative every day, like you're trying to do something creative, sets you up for getting into that flow. I mean, I think about creativity from an engineer's perspective. The things that prevent me from getting into that creative flow are things that I like to eliminate, both in the tools that I build and the infrastructure platforms. You just want to get rid of the friction. For the creators out here, there's typically two sides to creation. One side is the crazy idea that no one's ever done before. The other one is metrics, iteration, metrics, iteration. There was a time about maybe nine years ago when a lot of game companies were thinking they were going to iterate their way to success. They were going to just throw something out there and then metric, iterate, metric, iterate. On the other side of the spectrum, there are a lot of creators who just have a huge vision. No one's ever done it before. The vision is so powerful, you get a topological leap. Is there a way to think about being aware of both sides of this? Is one better than the other? Do you need both? Are they both good? Look, I think in every endeavor you need both. There's no way to iterate yourself towards great if you don't have a good idea in there somewhere or have a good starting point or good sense for even how to pivot things once the data is telling you something. Because oftentimes, the data tells you very uninformative things. It's binary. It's like, OK, you try something. It's like didn't work. And you don't learn. There are ways to design your experiments where you get something more than just a binary back that helps point you in a direction. You have to have some real thing that you're passionate about that is your insight. I don't think you can fully mechanize that part of the process. The mechanical bits I always think about are just getting the friction out of the system. If you have an idea, it's one of the things that's great about your platform. You've built all of this infrastructure that tries to remove friction from people getting an idea in front of users. That's ultimately what it's all about. You can't dilute yourself about, oh, my idea is great. You sit on it forever, launch it to users, and then maybe it's a terrible idea. Just figuring out how to connect with that user need is really crucial. I have a little bit of a personal theory that every really big thing needs some element of topological innovation. I would say companies like Disney, for example, topologically innovated with theme parks. Other companies we can think of made innovations in that. Do you think that's a real theory? Do you think it applies to our developer community? Yeah, I think so. There are certainly ways that you can structure environments to encourage creativity in that topological way. One of the things I would highly, highly, highly recommend to everyone is on Disney+ right now, there is a documentary called "Lights and Magic" that is like a six-episode history of Industrial Light & Magic. It's phenomenal. At the very worst, you'll watch it and feel all kinds of wonderful nostalgia. If you really pay attention to what they did, it's basically a story of innovation. It's just staggering what they were able to do. Almost everything about modern cinema was produced because of the outrageous things or things that seemed outrageous that they were doing. There's been a bunch of people who've written about this over the years. So Arthur C. Clarke wrote this thing in the 1960s called the" Profiles of the Future." It is where his three laws, like he postulated, everybody's familiar with, is like any sufficiently advanced technology is indistinguishable from magic. The super insightful thing that he said in "Profiles of the Future" is that every innovative thing that we take for granted right now, somebody thought was impossible at some point. A large part of innovation is just sort of having the courage to go over that frontier of what's impossible to possible. It is scary. So many people will tell you that you're dumb or idiot. It's not going to work. I still get this all the time. I've been working on a whole bunch of AI things for the past four years. When I started four years ago, so many people, super smart people, were like, yeah, this is never going to work. This is impossible. If you cannot persevere through that, you just never create the things that are remarkable. Let's talk a little about your book, "Reprogram the American Dream." I want to share an experience that I had with my daughter, who's an artist. I do not know if you have all played with DALL-E or Craiyon or other tools like this. For those of you that are in the audience, creators or developers, it was really freaky. She gave me a cloud of about 13 words. Then she showed me an ultra-high-resolution image of forest, moss, mushroom, hobbit, cabin that looked as if it came straight out of a Tolkien movie or something like that. It freaked me out. Will our developers and creators have jobs in 10 years? Oh, you absolutely will have jobs. In fact, I think there are going to be more people consuming more of the content that you produce. You're going to be able to produce more and more delightful things for people using these tools. This was sort of the point of the book that I wrote, which is the AI systems that we're building right now are just tools. They're built by human beings. We may get to AGI someday, things that are just very generally look like human intelligence. What we're building today are things that are useful for solving really particular problems that can help you perform a task. They need direction. DALL-E. Do you think there'll be a future where a Roblox developer creating something what ends up being super important in addition to maybe the 3D model or the texture is the way they describe what they want? And will that become a new talent in addition to 3D modeling? Yeah, look, I think probably so. I think there were three really interesting things that happened in AI last year. One of them was DALL-E and these diffusion models that are now everybody's super interested in. One was Codex, which is a model built on top of OpenAI's GPT that has powered it's already powering GitHub Copilot, which is a coding assistant. Then, to give credit to Demis at DeepMind, the work that they did to produce a protein structure database for every known protein to biology. I think those are three extraordinary things. The first two are very specifically tools that let you create in new ways. I don't know how many of you use GitHub Copilot. Most of the people that I talk to who are daily users of Copilot are like, oh my god, this is great. I love this tool. Even when it's not perfect, it's just a useful part of their job. I don't know that it ever gets to the point where the thing is the whole programmer. I suspect not. How far could it go? We have brainstormed that. First off, we treat with a radical level of respect the code of all of the creators on the platform. We treat with a radical level of security that code. There is a possibility the cumulative knowledge in that code, if appropriately consumed, could hyperpower automatic code generation. How far do you think it could go? Or is there a natural limit to that? Look, I think it will be able to do, the AI models will be able to do a lot. I suspect it will be for certain types of code. The way that I use it is really helpful in writing boilerplate code, for instance. Things where no code that I write is on any critical path anywhere. I'm way too useless. OK, caveat. Caveat. This is not shipping in the next version of Windows. Yeah, correct. Because I'm an infrequent coder at this point, when I sit down to code, I spend a lot of time, go to Stack Overflow, read the documentation, fiddle around with a bunch of things. It's really great for things like that, where I can just sort of express a thing I'm trying to accomplish. I do not have to remember the Pandas invocation. Pandas is a Python data science library to transform a CSV file in a particular way. I do not know if any of you watch Westworld. I do not know if you do. There is a segment in Westworld that is about two or three minutes that is sharing a future audio creation experience. An operator is sitting at a console. You could imagine what it is. I forget it. It is like, okay, make a city street. We want a protagonist. No, slightly older, gray beard. Do this. Make it more rainy out. Do that. Now in 3D, we jump from DALL-E, right? We jump from pixels to 3D objects and meshes. It seems like someday that is going to be feasible. Probably. Yeah. You and I are old enough where we remember the days where you had to do a whole bunch of stuff with Assembly language. You probably wrote 50 different hash table implementations because your programming language back in the 1980s did not have a good standard library with a dictionary implementation. I think our programming tools, this is what we all want. We want them to become more and more powerful over time. The interesting thing that we all do as developers is we get to decide the problems that we are going to go solve, which ones are interesting and have value. Those are the choices that we get to make. How do we all choose to use these tools? Do they let us create more impact? Or are they doing something else? That is the thing I just strongly encourage everyone to think. It's like the tools are amazing. Get your hands on them. Try to figure out how you do something with them that moves the needle for your users, your stakeholders, the people you care about. Yeah, it's interesting to imagine we were assembly language programmers. And then we got Pascal and then C and then C++ and then reflective languages and C#. And so we climbed that. There is this fun thing to think about. There's still got to be parts where there's assembly language. Yeah, for sure. Within anything. I think there's always going to be a huge, huge value. Even before these AI coding tools came around, I have a podcast. For a while, everyone who came on the podcast, we were sort of lamenting the fact that too few developers actually appreciate the complete stack. It's not that you have to. You can operate at a high level of abstraction and do amazing things. Occasionally, the abstractions break. You need to dive one level down. You might have to dive two levels. The further down you can go when things break, the more power you have. It is the same with anything, right? I'm a woodworker. I have a CNC machine that I can program to cut things out, which is great. I also occasionally need to fiddle around with a hand plane to dial something in. I think you just, for me at least, it's fun to have as mastery over as many tools as you possibly can. Have the fully automated, powerful ones. Understand the old ones too. You can take delight in having mastery over all of them. Let's take an example of one of our creators who might have a specific avatar style in their experience. We're going to ultimately allow the users of that experience to create avatars. Imagine, pick your favorite avatar style, Disney, The Incredibles, Lego, Hello Kitty. I don't care what it is. I could imagine an avatar experience where I went into one of these. Rather than doing the traditional sliders, telling their experience, I'd like an avatar that looks a little like me. Maybe here's my picture. I'd like it in the style of your experience, Hello Kitty, Lego, or The Incredibles. I'm also really into rainbows. I want to be a cheerleader and whatever. Do you think it will someday see that combination of my picture, the style, and additional words where we would generate that? Yeah, look, so I think what you just described, the technology is already good enough to do it. So we're behind the scenes already. It's just a matter of deciding to go do it. Okay. That's super awesome. I've heard that before. I guess one other thing with your book. Are you an optimist or dystopian? I'm certainly not dystopian. The way that I always describe it is I'm a short-term pessimist, long-term optimist, which. Oh, I like that, right? Which I sort of think is it's sort of the flavor of engineer I am. I sort of look at the world. And I'm like, all this stuff is broken. This doesn't work. Why is that not more perfect than it is? My wife hates it because I walk around with a slightly jaundiced eye. The long-term optimism is I have faith that we can fix all of these things. I really do believe in the capacity of folks like the people who are here in this room to imagine a better way for all of this stuff and then to go create it. I'm super optimistic in that sense. Yeah, I'm an optimistic person too. Sometimes if we roll back the clock and look 50 or 100 years ago, we can note the things short-term we might have been very, very pessimistic about. Over the course of 50 or 100 years, we find out those problems are solved. That was nothing to worry about. Yeah, it was funny that while I was sitting in the green room, I was reading an article somebody posted to Hacker News about a 1983 BYTE magazine article about the C programming language. Somebody on the comment thread was saying, oh, it's hilarious looking at the internet archive of the magazine at the letters that people have written in. There was a letter in this issue, some grumpy engineer saying, oh my god, I can't believe anybody thinks that the mouse is a good idea. This is never going to work. It's just a foolish thing to even try. That is sort of part of, I think, it's part of the creative froth that we're in as technologists. Everybody's trying new things. There are a bunch of naysayers for stuff. Sometimes the naysayers are right. Sometimes they're wrong. That is what makes what we do fun, I think. I feel your longer-term optimism aligns with our value of take the long view. That is another trinket with that. Kevin, I want to thank you so much, inspiring as always, for being here with you. We really totally appreciate it. Oh, thank you so much. We are going to all go. I believe lunch is outside right now. We're going to all grab lunch and network and hang out and reconvene here, I believe, at 1:00 P.M. We'll be back at 1:00 P.M. Thanks again, Kevin. Thank you. OK, see you guys. Please scan the QR code you see on the screen and submit a question for the upcoming Builderman Q&A. Enjoy your lunch. Like a little kid, started up and driving. Come on, let's get ready. You know how to get me. Play it loud. It's my song. I know it drives you crazy. Come on, come on, let's ride away. Welcome, everyone. Wow, we are a lot bigger than last year. We're co-inventing the future of communication, storytelling, really the next generation of how we come together. Thank you for coming.