WEBVTT - Snap CEO Evan Spiegel is Betting on Smart Glasses 0:00:00.960 --> 0:00:04.480 I'm Emily Chang, and this is the circuit. On this 0:00:04.559 --> 0:00:07.280 episode of the pod, I'm in Sunny, La to talk 0:00:07.320 --> 0:00:10.399 to Evan Spiegels. He's the co founder and CEO of 0:00:10.480 --> 0:00:13.960 snap known, of course, for Snapchat, the messaging app, famous 0:00:13.960 --> 0:00:17.520 for its disappearing photos and videos. Snapchat burst onto the 0:00:17.560 --> 0:00:21.119 scene in twenty eleven, a heady time for social media startups. 0:00:21.400 --> 0:00:24.360 Facebook had just bought Instagram, Twitter was on the rise, 0:00:24.640 --> 0:00:28.400 and Spiegel had started Snapchat with two college friends at Stanford. 0:00:28.840 --> 0:00:30.160 What was it that made you say I need to 0:00:30.200 --> 0:00:31.840 drop out? I needed to do this. 0:00:32.320 --> 0:00:35.320 We're waiting for our venture investor's light speed to fund 0:00:35.760 --> 0:00:38.120 and I was refreshing the Wells Fargo app. In the 0:00:38.159 --> 0:00:39.919 back of the class saw the four and eighty five 0:00:39.960 --> 0:00:42.680 thousand dollars. Sit walked up to my teacher Craig and 0:00:42.760 --> 0:00:44.360 just said, you know what, I'm so sorry, I'm gonna 0:00:44.360 --> 0:00:45.240 have to drop this class. 0:00:46.080 --> 0:00:49.200 Within a couple of years, Snapchat was rising in popularity, 0:00:49.440 --> 0:00:52.720 stirring up its fair share of controversy and making Spiegel 0:00:52.840 --> 0:00:56.280 the world's youngest billionaire by age twenty five. What was 0:00:56.320 --> 0:00:57.680 her first I made it purchase? 0:00:57.840 --> 0:01:00.680 My first purchase was a pre owned for I went 0:01:00.720 --> 0:01:02.520 to the Ferrari dealership and they must have looked at 0:01:02.520 --> 0:01:03.440 me like, what are you doing. 0:01:03.960 --> 0:01:06.320 Spiegel has also grown up a bit, but is still 0:01:06.440 --> 0:01:09.479 very much in founder mode. Now in his thirties, he's 0:01:09.520 --> 0:01:12.600 running a multi billion dollar company that's locked in competition 0:01:12.760 --> 0:01:17.000 with both trillion dollar rivals and smaller startups. To stay 0:01:17.040 --> 0:01:19.440 in the game, he's betting big on smart glasses and 0:01:19.480 --> 0:01:23.000 the future of augmented reality, even as tech's biggest players 0:01:23.000 --> 0:01:26.800 crowd the field like Meta, Google and Apple. In a 0:01:26.800 --> 0:01:29.880 recent letter to shareholders, Spiegel wrote that Snap is at 0:01:29.880 --> 0:01:33.119 a crucible moment, faced with many obstacles that will define 0:01:33.160 --> 0:01:35.920 its future. We sat down to talk about what it's 0:01:35.959 --> 0:01:38.520 like to fight in a land of giants, how Snap 0:01:38.600 --> 0:01:42.040 navigates the dark side of social media, TikTok and China, 0:01:42.400 --> 0:01:44.640 and why he thinks the screens of today won't look 0:01:44.680 --> 0:01:50.160 anything like the screens of tomorrow. Okay, so it's easy 0:01:50.160 --> 0:01:52.400 to forget because it's been so long now, but Snap 0:01:52.440 --> 0:01:55.120 popularized so many things that are just like a fact 0:01:55.640 --> 0:02:00.000 of social media today. Disappearing messages, stories, filters, and lenses, 0:02:00.120 --> 0:02:04.560 is even vertical video location tags. What was it about you? 0:02:04.680 --> 0:02:07.120 And Snap's culture that enabled that. 0:02:08.000 --> 0:02:10.000 I think one of the things that was really unique 0:02:10.040 --> 0:02:12.280 is that we were mobile first, so we were one 0:02:12.280 --> 0:02:15.680 of the first services to start building mobile only. A 0:02:15.680 --> 0:02:18.320 lot of our competitors actually had started on the desktop 0:02:18.320 --> 0:02:21.040 and were migrating to mobile, but we had grown up 0:02:21.040 --> 0:02:24.560 with smartphones, and everything we were building was mobile first. 0:02:24.639 --> 0:02:26.840 And many of the things that we have built have 0:02:26.960 --> 0:02:31.480 really just been about making computing or making technology more human. 0:02:31.639 --> 0:02:33.880 I mean that's sort of been the key insight time 0:02:33.919 --> 0:02:35.919 and again, even things as simple as back in the 0:02:36.000 --> 0:02:37.680 day when you would open your camera you had to 0:02:37.760 --> 0:02:41.320 choose between photo or videos. Really annoying to set this toggle, 0:02:41.360 --> 0:02:43.639 and we realized, like, it's just a moment. Either it's 0:02:43.720 --> 0:02:46.240 a single moment in time, or it's a longer moment, 0:02:46.280 --> 0:02:48.160 which is a video, And so we invented this button 0:02:48.200 --> 0:02:49.880 where you would tap to take a photo or press 0:02:49.880 --> 0:02:52.160 and hold for video. And it was just all about 0:02:52.160 --> 0:02:54.880 making things much more intuitive for people to use. The 0:02:54.919 --> 0:02:57.360 same of course for stories. For vertical video was so 0:02:57.720 --> 0:02:59.639 obvious to us. I mean, all the video at the 0:02:59.639 --> 0:03:03.080 time cut horizontally. All the social feeds would have these tiny, 0:03:03.120 --> 0:03:06.440 little like postage sam sized videos horizontally on your phone, 0:03:06.560 --> 0:03:08.720 and of course the vision pitched at the time was like, oh, 0:03:08.720 --> 0:03:11.239 you would just turn your phone sideways and watch video 0:03:11.600 --> 0:03:13.480 that way, But it was very uncomfortable to hold your 0:03:13.520 --> 0:03:16.000 phone like that and not very practical, and so we 0:03:16.040 --> 0:03:17.960 just realized people were going to want to create and 0:03:18.080 --> 0:03:21.560 watch video vertically the same way that they held their phones. Now, 0:03:21.600 --> 0:03:24.000 the challenging part about that was because everyone was so 0:03:24.120 --> 0:03:26.440 used to horizontal video, especially our advertisers. 0:03:26.720 --> 0:03:27.440 They were like, this. 0:03:27.480 --> 0:03:30.120 Is totally crazy, We're gonna have to shoot totally new content. 0:03:30.360 --> 0:03:31.840 But at some point we ran a study that the 0:03:31.880 --> 0:03:34.640 completion rates were nine times higher when it's full screen 0:03:34.680 --> 0:03:37.760 vertical video compared to these little horizontal videos, and over 0:03:37.800 --> 0:03:39.680 time folks realized it just made more sense. 0:03:39.760 --> 0:03:43.440 So it was all about making things more human, more intuitive. 0:03:43.080 --> 0:03:46.360 And easier to use, designed around the way that people live, 0:03:46.440 --> 0:03:48.480 instead of forcing people to change the way they live 0:03:48.560 --> 0:03:49.680 to use technology. 0:03:49.880 --> 0:03:52.200 I actually remember the first time I saw Snapchat in 0:03:52.240 --> 0:03:54.360 the wild. We were at an ice cream parlor and 0:03:54.440 --> 0:03:56.200 there was this crowd of kids in front of us, 0:03:56.240 --> 0:03:57.880 and they all had their phones up and I was like, oh, 0:03:57.880 --> 0:04:02.240 they're taking pictures off the ice cream. Husband whispered, there's snapchatting, 0:04:02.400 --> 0:04:05.280 and I was like, oh, this is becoming a thing. 0:04:05.400 --> 0:04:07.760 Like it was really a moment, was there a sort 0:04:07.800 --> 0:04:10.280 of an inflection point where you felt like things really 0:04:10.280 --> 0:04:10.760 took off. 0:04:11.800 --> 0:04:14.200 One of the things that's very special about snapchat is 0:04:14.200 --> 0:04:18.359 it actually changed people's relationship with photography. So I'm dating myself, 0:04:18.400 --> 0:04:20.440 but back in the day, people used to take photos 0:04:20.440 --> 0:04:23.480 to save a moment, and even further back, people would 0:04:23.520 --> 0:04:26.320 like scrapbook, so it's really about recording a moment or 0:04:26.360 --> 0:04:30.000 recording something special. Snapchat changed people's relationship with photos and 0:04:30.040 --> 0:04:32.839 made them about communicating. And it turns out people communicate 0:04:32.960 --> 0:04:35.240 way more than they save things. And so there was 0:04:35.279 --> 0:04:37.680 a moment in time when we realized that the number 0:04:37.720 --> 0:04:41.040 of photos people were taking and sending on Snapchat eclipsed 0:04:41.080 --> 0:04:44.280 the total number of normal photos people were taking out 0:04:44.279 --> 0:04:45.840 in the world with their smartphones. And I think that 0:04:45.920 --> 0:04:47.800 was really the moment that we realized there had been 0:04:47.839 --> 0:04:50.640 this huge shift that like the majority use case or 0:04:50.680 --> 0:04:55.120 the primary use case for cameras was communicating, not saving things, 0:04:55.320 --> 0:04:57.479 and that I think was a pretty meaningful shift. 0:04:57.600 --> 0:05:01.760 Lucky for your server bills, they didn't get saved correct. 0:05:01.640 --> 0:05:04.640 Very unlucky for our server bills, but we were so fortunate. 0:05:04.680 --> 0:05:06.920 We were one of the first companies to build cloud 0:05:07.000 --> 0:05:09.720 first as well, so we were mobile first and cloud first, 0:05:09.839 --> 0:05:11.880 and that's one of the things that allowed us to 0:05:11.960 --> 0:05:14.640 really innovate and spend our time and energy building new 0:05:14.640 --> 0:05:17.240 products instead of racking servers. Most of the companies that 0:05:17.279 --> 0:05:20.040 came before us, a huge scaling problem for them was 0:05:20.120 --> 0:05:24.520 literally server infrastructure, physical infrastructure, and Snapchat built on the cloud. 0:05:24.560 --> 0:05:27.200 So while the bills were aggressive, the scaling was a 0:05:27.240 --> 0:05:27.760 lot easier. 0:05:27.880 --> 0:05:30.760 Elon Musk once said, starting a company is like staring 0:05:30.800 --> 0:05:32.479 into the abyss and eating glass. 0:05:33.480 --> 0:05:35.159 He's got away with word he does. 0:05:35.640 --> 0:05:38.239 If you could go back and talk to college Evan, 0:05:38.400 --> 0:05:39.360 what would you tell him? 0:05:39.800 --> 0:05:41.800 I would tell him, don't worry, It's all going to 0:05:41.839 --> 0:05:43.480 work out in the end. I think it can be 0:05:43.560 --> 0:05:46.560 really challenging starting a business. You know, we had no 0:05:46.640 --> 0:05:49.960 idea what we were doing. We had absolutely no money. 0:05:50.120 --> 0:05:52.359 I had ten thousand dollars my grandfather had left me. 0:05:52.520 --> 0:05:55.080 We maxed out our credit card bills, and we were very, 0:05:55.160 --> 0:05:58.120 very fortunate to get investment from light Speed in those 0:05:58.200 --> 0:05:59.760 early days, kind of just as we were about to 0:05:59.839 --> 0:06:02.120 hit the wall on our own ability of fun the thing, 0:06:02.120 --> 0:06:03.520 and my dad was like, there's no way I'm paying 0:06:03.560 --> 0:06:05.839 for people to send disappearing photos like this makes no sense. 0:06:05.880 --> 0:06:07.919 So we've just been so fortunate to be able to 0:06:07.920 --> 0:06:10.440 build a great team and step by step just grow 0:06:10.480 --> 0:06:11.520 by serving our community. 0:06:11.640 --> 0:06:14.760 My understanding is that you're religious about running a design 0:06:14.960 --> 0:06:18.960 first company. Is that true and how does that work? 0:06:19.120 --> 0:06:21.200 I wouldn't say design first because a lot of the 0:06:21.200 --> 0:06:23.880 way that we built our company is actually modeled on 0:06:23.960 --> 0:06:28.320 my relationship with Bobby. He is engineering first, but he 0:06:28.360 --> 0:06:30.760 has a real appreciation for design, and I would say 0:06:30.800 --> 0:06:34.240 I'm designed first, but I understand appreciate engineering. And so 0:06:34.400 --> 0:06:37.200 we've been able to build this dialogue over time, and 0:06:37.320 --> 0:06:39.560 especially in the early days where the way that we 0:06:39.600 --> 0:06:42.600 would invent things would take a lot of iteration, both 0:06:42.600 --> 0:06:45.480 on the engineering side and the design side. So, for example, 0:06:45.680 --> 0:06:47.400 one of the first pieces of feedback we got was 0:06:47.440 --> 0:06:49.760 sampshat doesn't work. You can always just take a screenshot, 0:06:49.920 --> 0:06:51.880 and so we did a lot of work to invent 0:06:51.920 --> 0:06:54.159 a new form of screenshot detection, which at the time 0:06:54.240 --> 0:06:56.560 wasn't just an API on iPhone. We had to figure 0:06:56.560 --> 0:06:58.880 out how to determine someone had taken a screenshot and 0:06:58.880 --> 0:07:01.120 then send that note if it came back, and to 0:07:01.160 --> 0:07:03.400 design that both in terms of the user interaction, which 0:07:03.400 --> 0:07:05.800 required people to press and hold on the image while 0:07:05.800 --> 0:07:08.000 they were viewing it, but then also in terms of 0:07:08.000 --> 0:07:09.720 the engineering and what the client had to do to 0:07:09.760 --> 0:07:12.240 detect that a screenshot had occurred was something that would 0:07:12.320 --> 0:07:16.120 only happen with this dialogue and really tight integration between 0:07:16.160 --> 0:07:19.560 design and engineering. And so we've really found that innovation 0:07:19.760 --> 0:07:22.320 lives at kind of the intersection of those disciplines, maybe 0:07:22.360 --> 0:07:23.880 a little bit of art as well, and that's how 0:07:23.920 --> 0:07:25.480 we tried to model the company. So I don't think 0:07:25.480 --> 0:07:27.720 it's design first, it's not engineering first. It's really about 0:07:27.720 --> 0:07:31.760 that relationship between design and engineering that allows us to innovate. 0:07:31.840 --> 0:07:34.840 So is that more important than product or growth? And 0:07:35.240 --> 0:07:37.800 how does that compare then to TikTok or Meta. 0:07:38.480 --> 0:07:40.120 I wouldn't say it one is more important than the other. 0:07:40.160 --> 0:07:43.239 It's really a team based approach. But our general view 0:07:43.520 --> 0:07:46.800 is that growth follows great product and great innovation of 0:07:46.880 --> 0:07:49.560 service that people really love to use. And that's really 0:07:49.600 --> 0:07:51.920 how we focused on building Snapchat, especially because it was 0:07:51.960 --> 0:07:54.640 such a new behavior. If you didn't have a friend 0:07:54.680 --> 0:07:57.920 teach you how to use it, it was really hard to understand. 0:07:58.000 --> 0:07:59.960 And so just by building a product that people love 0:08:00.120 --> 0:08:02.120 having them teach their friends how to use it, Snapchat 0:08:02.200 --> 0:08:03.400 is not a lot of fun if you're just using 0:08:03.440 --> 0:08:06.320 it by yourself. So that is what really drove our 0:08:06.320 --> 0:08:07.160 growth over time. 0:08:07.280 --> 0:08:10.880 How is AI changing snap and what is design in 0:08:10.920 --> 0:08:11.760 an AI world? 0:08:12.640 --> 0:08:15.200 Well, I think AI is an incredible creative tool. That's 0:08:15.200 --> 0:08:18.360 certainly how we're using it as we approach the product, 0:08:18.680 --> 0:08:22.720 and especially because Snapchat is a visual first service, we 0:08:22.800 --> 0:08:25.560 open into the camera. A lot of our AI investments 0:08:25.600 --> 0:08:28.440 on the product experience are focused on image on video, 0:08:28.680 --> 0:08:32.040 on three D asset generation and really enabling people to 0:08:32.200 --> 0:08:35.199 express themselves in new ways. So I think that's really 0:08:35.200 --> 0:08:37.280 been our focus on the product side. Of course, AI 0:08:37.840 --> 0:08:41.280 informs every product surface that we have, whether it's the 0:08:41.360 --> 0:08:45.200 content recommendations you see or the lens recommendations. Of course, 0:08:45.240 --> 0:08:47.320 everything we do to make sure that you're finding friends 0:08:47.320 --> 0:08:49.640 that you know in the real world, all of our 0:08:49.679 --> 0:08:53.440 trust and safety infrastructure increasingly more and more customer service. 0:08:53.600 --> 0:08:55.560 The only way we can serve a community of nearly 0:08:55.600 --> 0:08:58.160 a billion people really well is with the support of 0:08:58.240 --> 0:09:01.080 all of the AI infrastructure we've built over the years. 0:09:01.800 --> 0:09:05.080 Twenty fourteen, you start dabbling in AR. Twenty fifteen, you 0:09:05.160 --> 0:09:08.160 built a research arm to explore AI in deep learning. 0:09:08.200 --> 0:09:12.800 Twenty sixteen you release spectacles. What gave you such conviction 0:09:13.080 --> 0:09:15.559 early on? How did you commit to those things so 0:09:15.640 --> 0:09:18.959 early on? Because there's always a choice, right or resource choice? 0:09:19.400 --> 0:09:21.960 Yeah, I think you know a lot of those ideas 0:09:22.000 --> 0:09:26.679 are the technological or engineering expression of making computing more human. 0:09:26.800 --> 0:09:28.840 So we just talked about a lot of the interesting 0:09:28.880 --> 0:09:32.480 product or user interface examples like stories or vertical video, 0:09:32.679 --> 0:09:34.600 or the ability to tap to take a photo and 0:09:34.640 --> 0:09:36.920 hold to record a video. Those are all elements of 0:09:36.920 --> 0:09:39.040 the user interface that we design. But at the same time, 0:09:39.360 --> 0:09:43.480 we've been making very meaningful investments in the underlying technology 0:09:43.520 --> 0:09:45.120 that we think helps make. 0:09:45.000 --> 0:09:46.040 Computing feel more human. 0:09:46.080 --> 0:09:48.800 I think AI one of the most powerful moments people 0:09:48.800 --> 0:09:51.600 had when they first started using CHATCHBT, for example, was 0:09:51.640 --> 0:09:54.520 feeling like they were sending messages back and forth with 0:09:54.559 --> 0:09:58.280 a person. That's I think epitomizes the notion that technology 0:09:58.360 --> 0:10:00.760 is going to learn how to interface with humans in 0:10:00.840 --> 0:10:03.560 a way that feels much more familiar, and that's really 0:10:03.600 --> 0:10:05.520 been the focus of a lot of these investments. 0:10:05.559 --> 0:10:08.000 What is the vision for spectacles? Do you want it 0:10:08.040 --> 0:10:09.000 to replace phones? 0:10:09.520 --> 0:10:09.880 Initially? 0:10:10.080 --> 0:10:12.360 I don't think it will replace phones, but I think 0:10:12.400 --> 0:10:14.800 over a very long period of time, it will replace 0:10:14.920 --> 0:10:17.959 most of the screens in your life, because if you imagine, 0:10:18.040 --> 0:10:20.960 glasses can provide a display area that's maybe the equivalent 0:10:20.960 --> 0:10:24.200 of four hundred times a smartphone, but at a fraction 0:10:24.559 --> 0:10:27.000 of the weight, and it's not something. 0:10:26.760 --> 0:10:28.079 You have to carry around. You don't have to use 0:10:28.080 --> 0:10:28.440 your hands. 0:10:28.440 --> 0:10:30.080 If you want a big screen on the wall, you 0:10:30.080 --> 0:10:32.280 can just hang it there in space through the glasses. 0:10:32.360 --> 0:10:35.000 So I do think over time people will wonder why 0:10:35.040 --> 0:10:37.440 did I, at one point in time have a tablet, 0:10:37.520 --> 0:10:40.560 a phone, a TV, I guess, watches or screens. 0:10:40.559 --> 0:10:41.880 Now refrigerators or screens. 0:10:41.920 --> 0:10:43.960 Everything's a screen, And I think at some point people 0:10:44.000 --> 0:10:46.520 will wonder, why did we ever do that? When glasses 0:10:46.559 --> 0:10:49.440 allow us to overlay that sort of technology and interface 0:10:49.679 --> 0:10:51.120 on anything in our world? 0:10:51.559 --> 0:10:54.880 Meta, Google, Apple, All of these companies have glasses or 0:10:55.000 --> 0:10:58.520 some form of wearables. Now, how do you compete? These 0:10:58.559 --> 0:10:59.800 are massive companies. 0:11:00.120 --> 0:11:03.720 I think focus is really important. In particular, we've. 0:11:03.600 --> 0:11:07.440 Been so focused on the glasses form factor since twenty fourteen, 0:11:07.840 --> 0:11:11.360 when a lot of people were investing in virtual reality. 0:11:11.520 --> 0:11:13.440 So I think a lot of the investments people have 0:11:13.480 --> 0:11:15.360 made over the last ten or eleven years have been 0:11:15.440 --> 0:11:18.000 much more VR focused, and people felt like that was 0:11:18.040 --> 0:11:20.400 a stepping stone to glasses. But in our view that 0:11:20.480 --> 0:11:23.760 technology stacks are actually quite different, and the notion of 0:11:23.760 --> 0:11:25.920 putting a screen on your face to block out the 0:11:25.920 --> 0:11:28.240 world and enter sort of a different or an alternate 0:11:28.280 --> 0:11:32.400 reality is completely different than the vision for integrating computing 0:11:32.440 --> 0:11:34.600 into your life, into the real world with your friends 0:11:34.720 --> 0:11:37.000 and your family. So I think what we've always found 0:11:37.080 --> 0:11:40.040 is just a singular focus and execution against that over 0:11:40.040 --> 0:11:42.480 a very long period of time helps to compound a 0:11:42.600 --> 0:11:45.800 technological advantage that I think positions us well competitively. 0:11:45.880 --> 0:11:48.960 What does the market look like like? Will every company 0:11:48.960 --> 0:11:51.240 have one and we'll buy one of each or will 0:11:51.240 --> 0:11:52.840 one to two win. 0:11:53.520 --> 0:11:56.520 It's really hard to say now it's still very early 0:11:56.600 --> 0:11:59.960 on and people are playing in different parts of the stack, 0:12:00.120 --> 0:12:00.560 so to say. 0:12:00.840 --> 0:12:03.280 So, we have what's that in human language? 0:12:04.360 --> 0:12:07.040 We have a developer ecosystem, so hundreds of thousands of 0:12:07.040 --> 0:12:10.280 developers who have built millions of lenses on our developer tools. 0:12:10.640 --> 0:12:13.600 When they build those experiences, they run on our rendering engine, 0:12:13.679 --> 0:12:16.640 which is really deeply integrated into our operating system, and 0:12:16.679 --> 0:12:19.520 they see and experience those AR objects through our glasses 0:12:19.600 --> 0:12:22.160 using our own optical engine, meaning the piece of glass 0:12:22.200 --> 0:12:24.800 in the glasses and the little projector we design and 0:12:24.800 --> 0:12:27.280 build ourselves. And so when you look at that full 0:12:27.280 --> 0:12:31.040 stack integration and our ability to control every aspect of 0:12:31.040 --> 0:12:34.880 the experience, we can deliver a singular product experience compared 0:12:34.920 --> 0:12:37.600 to other companies that have various parts of the stack 0:12:37.640 --> 0:12:39.800 but are kind of cobbling together the rest of it. 0:12:39.800 --> 0:12:42.800 It's very hard to create an extraordinary product experience when 0:12:42.840 --> 0:12:44.360 you don't have that sort of vertical control. 0:12:44.679 --> 0:12:46.320 How does AI change AR? 0:12:47.160 --> 0:12:51.200 I think fundamentally, AR is the interface for AI. So 0:12:51.320 --> 0:12:53.520 a lot of people right now are spending time investing 0:12:53.559 --> 0:12:55.160 in building these really big models. 0:12:55.200 --> 0:12:56.680 There's a lot of competition around that. 0:12:56.840 --> 0:12:58.760 You can almost think about that as the back end 0:12:58.800 --> 0:13:01.920 of AI, the infrastruction ructure layer for AI. But where 0:13:01.960 --> 0:13:03.559 we're spending a lot of our time and where we're 0:13:03.640 --> 0:13:05.720 very interested is what is the front end of AI. 0:13:05.840 --> 0:13:08.040 How are people actually going to interact with and engage 0:13:08.080 --> 0:13:10.679 with AI and Certainly the chatbot is not the end 0:13:10.720 --> 0:13:12.280 of the road here for the way that people are 0:13:12.280 --> 0:13:15.120 going to want to engage with and use AI. And 0:13:15.160 --> 0:13:17.400 the reason why we think AR is such a powerful 0:13:17.520 --> 0:13:21.000 interface is because it brings AI into the world with you, 0:13:21.160 --> 0:13:23.800 so it can understand your environment, understand who you're with, 0:13:23.880 --> 0:13:27.400 obviously make the right recommendations in terms of your experience, 0:13:27.520 --> 0:13:30.600 but be very very helpful in the moment as well. 0:13:31.080 --> 0:13:36.000 How do the AI wars play out, what happens, who wins. 0:13:36.280 --> 0:13:38.520 I really wish I could predict the future in that regards. 0:13:38.520 --> 0:13:40.920 I'd probably make a lot of money and it'd be 0:13:40.920 --> 0:13:43.320 fascinating to see where this is all headed. I think 0:13:43.360 --> 0:13:45.520 one of the things that we have found that's happened 0:13:45.640 --> 0:13:48.679 very very rapidly is the commodification of this technology, and 0:13:48.720 --> 0:13:51.280 so I think there's been a huge proliferation of these 0:13:51.320 --> 0:13:54.320 foundational models, a lot of progress and open source, and 0:13:54.400 --> 0:13:56.719 so I think people are thinking a lot about how 0:13:56.720 --> 0:13:59.480 to differentiate on top of these models, and so I 0:13:59.480 --> 0:14:01.719 think a lot of the winners down the road will 0:14:01.760 --> 0:14:04.880 be people who figure out how to do personalization very well. 0:14:04.960 --> 0:14:07.720 So I feel like only one model really knows me 0:14:07.800 --> 0:14:10.800 better than another one. Or people who are really thinking 0:14:10.840 --> 0:14:13.120 about how to integrate these models with other tools so 0:14:13.160 --> 0:14:16.040 they become a lot more useful than just a chatbot. 0:14:16.280 --> 0:14:19.280 And certainly the business models are going to matter a lot, right, 0:14:19.280 --> 0:14:23.640 And what is the customer relationship and customer expectation around AI? 0:14:24.040 --> 0:14:26.720 Especially with a technology that's so personal? I mean people 0:14:26.760 --> 0:14:30.800 are having very very personal conversations with AI about therapy 0:14:31.040 --> 0:14:34.240 or about their health. These are really serious topics and 0:14:34.240 --> 0:14:35.840 I think that needs to be treated with just the 0:14:35.960 --> 0:14:39.320 utmost care. So the companies that do that well and 0:14:39.400 --> 0:14:41.960 who are able to build that trust with their community, 0:14:41.960 --> 0:14:43.560 I think that's going to be a huge predictor of 0:14:43.600 --> 0:14:44.480 success down the road. 0:14:44.760 --> 0:14:46.440 All of the top AI talent is going to all 0:14:46.480 --> 0:14:49.120 these other companies. Does that put you at a disadvantage? 0:14:49.240 --> 0:14:52.520 We actually do build our own models, especially in image 0:14:52.520 --> 0:14:55.480 and video and three D. We're an industry leader and 0:14:55.560 --> 0:14:58.280 one of the reasons why people really like coming to 0:14:58.320 --> 0:15:01.960 Snap is because rather just trying to build the biggest model, 0:15:02.040 --> 0:15:04.560 we try to build models that a huge number of 0:15:04.600 --> 0:15:07.400 people can use. So what does that mean for us? 0:15:07.720 --> 0:15:10.760 We take really advanced image in video models and we 0:15:10.800 --> 0:15:13.560 distill them to run on your device for free. 0:15:13.920 --> 0:15:15.520 And that allows hundreds. 0:15:15.200 --> 0:15:17.240 Of millions of people that use Snapchat to use these 0:15:17.280 --> 0:15:19.880 incredibly advanced AI tools that they would have to pay 0:15:20.200 --> 0:15:22.360 ten dollars a month, twenty dollars a month or more 0:15:22.400 --> 0:15:24.520 to use on other platforms. And so as we think 0:15:24.560 --> 0:15:27.400 about our responsibility to our community, you know, nearly a 0:15:27.440 --> 0:15:29.520 billion people, how can we bring these tools to them 0:15:29.800 --> 0:15:31.720 as quickly as possible in a way that they can 0:15:31.760 --> 0:15:34.520 really use and benefit from Because I think, you know, 0:15:34.560 --> 0:15:38.200 it's certainly one goal to build the perfect largest video model, 0:15:38.200 --> 0:15:41.120 but if it's super slow, if it's really expensive, if 0:15:41.160 --> 0:15:42.760 you can't run it on device, it's going to be 0:15:42.800 --> 0:15:44.720 hard for the broader world to benefit from that. 0:15:45.040 --> 0:15:47.080 And how do you get the talent right, because I 0:15:47.080 --> 0:15:49.320 mean it's like a war for talent for all of 0:15:49.320 --> 0:15:50.400 these top AI people. 0:15:50.960 --> 0:15:53.520 I think we have a reputation in the industry for 0:15:53.680 --> 0:15:56.760 being an incredibly innovative company. We certainly have the resources 0:15:56.800 --> 0:15:59.520 to invest behind the talent that we have here and 0:15:59.560 --> 0:16:02.160 to really help them continue to innovate. And I think, 0:16:02.280 --> 0:16:06.000 especially as it pertains to images, videos, augmented reality, these 0:16:06.000 --> 0:16:08.920 are areas where people who are in the industry see 0:16:08.960 --> 0:16:11.520 how far ahead we are, see the boundaries that we're pushing. 0:16:11.560 --> 0:16:13.960 And so if your life goal is to contribute to 0:16:14.000 --> 0:16:16.880 this mission to make computing more human, to bring computing 0:16:16.920 --> 0:16:18.880 into the world so we can stop spending seven hours 0:16:18.880 --> 0:16:21.080 a day looking down at screens, I think Snap is 0:16:21.120 --> 0:16:23.240 the best place to do that, and it's certainly a 0:16:23.320 --> 0:16:26.480 differentiated place to do that compared to other companies that 0:16:26.520 --> 0:16:29.520 I think are pursuing a relatively similar strategy in terms 0:16:29.560 --> 0:16:30.440 of the models they're building. 0:16:30.480 --> 0:16:33.480 What do you think of Mark Zuckerberg's AI strategy in particular, 0:16:33.880 --> 0:16:36.200 It's very ambitious, but also there are a lot of skeptics. 0:16:36.880 --> 0:16:39.120 I think he's just illustrating how expensive it is to 0:16:39.120 --> 0:16:42.960 play catch up, especially with advanced technology. I think today, 0:16:43.240 --> 0:16:45.520 as we look at the areas where folks are trying 0:16:45.520 --> 0:16:48.640 to invest things like augmented reality, things like AI, these 0:16:48.800 --> 0:16:52.320 technical stacks are very very difficult to build. They require 0:16:52.400 --> 0:16:55.600 highly specialized talent, and as I mentioned, I think that 0:16:55.640 --> 0:16:58.960 the benefits compound over time, and so to be a 0:16:59.040 --> 0:17:02.240 late entrant into the these areas ends up being very expensive. 0:17:02.280 --> 0:17:04.160 And so we'll see how it plays out. 0:17:04.240 --> 0:17:05.000 Can he catch up. 0:17:05.040 --> 0:17:07.080 He's certainly done a hell of a job building his 0:17:07.160 --> 0:17:09.080 business and I'm on the edge of my seat. 0:17:10.480 --> 0:17:13.160 Well, on that note, your LinkedIn says you're the VP 0:17:13.280 --> 0:17:16.080 of product at Meta. How do you really feel? 0:17:16.680 --> 0:17:18.320 I wish that was my joke, because I get so 0:17:18.400 --> 0:17:21.360 much positive feedback on that. One one of the fun 0:17:21.400 --> 0:17:24.560 parts of being a designer is building stuff that lots 0:17:24.560 --> 0:17:27.159 of people use, and so it's very gratifying from the 0:17:27.240 --> 0:17:30.480 design perspective that so many of the things we've invented 0:17:30.520 --> 0:17:33.359 here at snap are so widely used across the industry. 0:17:33.359 --> 0:17:35.360 I mean, I think even LinkedIn has stories at this point. 0:17:35.480 --> 0:17:37.879 So I think it's been really fun to see the 0:17:37.920 --> 0:17:40.760 adoption of all this technology across the industry. But it's 0:17:40.800 --> 0:17:42.879 also meant that from the very early days, you know, 0:17:42.880 --> 0:17:45.600 as we saw how the competitive landscape would play out, 0:17:45.640 --> 0:17:49.240 that we invested in technology that was highly differentiated, that 0:17:49.400 --> 0:17:53.040 was very very difficult to copy, and that has platform characteristics. 0:17:53.080 --> 0:17:55.000 So I think you see that today with our augmented 0:17:55.040 --> 0:17:58.000 reality platform, that's the leading our platform in the world. 0:17:58.359 --> 0:18:01.240 We have people use lenses just inside Snapchat and the 0:18:01.240 --> 0:18:04.000 camera eight billion times a day. Google Search I think 0:18:04.000 --> 0:18:06.359