48 Hours Inside the Startup Building the World’s First On-Device AGI

Steve 0:00 Most people's experience with AI is still chatbots. The future is that you just talk to your device and it does things for you.

Div Garg 0:05 Imagine something like Jarvis from Iron Man. Wherever I am, I have an AI that I own. It's my personal AI. That's possible right now with, like, things we are building.

Speaker 0:13 I want a Chipotle steak, brown rice, guac, sour cream, and some cheese.

Speaker 0:20 Food is here.

Div Garg 0:22 We can have an army of agents that can basically go and do everything for you. The whole humanity gets equalized on the same basis, and we want to make that happen. You will have a whole suite of new devices, like smart wearables.

Speaker 0:32 This is gonna go mega viral.

Speaker 0:33 Yeah.

Speaker 0:34 We have a line of, like, pendants that you speak into.

Speaker 0:37 Brew me a coffee, please.

Div Garg 0:39 We also partner with, like, uh, chip companies like Qualcomm.

Speaker 0:41 How do you make that demo into a real business?

Steve 0:45 Product. Make the product good. Make the agent good. Everything follows from there.

Div Garg 0:49 And that's what, like, people are coming to us for because we are the only ones who are actually really solving that challenge.

Speaker 0:53 This is, like, a 9-9-6 type of company. I end up doing, like, 9:30 to 9:30 pretty much every day.

Div Garg 0:58 If you love what you're doing, it doesn't matter. I dropped out of my Stanford PhD and, uh, turned down a billion-dollar offer from OpenAI to build AGI Inc. We built our own on-device models that can go and take actions on your phone. So you can control your whole phone and use any app for you. I think this is the most impactful thing we can do for the whole humanity right now.

Speaker 1:26 Signing my contract. Yeah. Uh, field AGI.

Speaker 1:31 Uh, it's my day one at AGI. Uh, my name is Sharik. Today is going to be my onboarding, and I'm super excited. Before this, uh, I was working a lot with, uh, voice AI and voice LLMs. Our interns don't pick up coffee.

Speaker 1:46 [laughs]

Speaker 1:47 They build agents that make us coffee. Open coffee link. And click Brew under Coffee.

Speaker 2:02 Is he successfully onboarded? [laughs]

Speaker 2:05 He passed my test.

Speaker 2:06 [laughs]

Speaker 2:06 Next is Namani Div. That's a whole different test.

Speaker 2:08 [laughs]

Speaker 2:10 So I, I talked to Div at the AGI Inc. uh, party, the... or the co-working session that they hosted, uh, a couple of days back. I kind of asked him about other competition that exists for AGI. I asked all three of them the exact same question, and I love their answer to it. They don't consider them to be a competition in the sense that they are doing something to kind of push this field ahead as well, and the more competition there is, the better you're able to build. And I feel like that's a very wholesome thing to kind of have as in, in your founders. Thinking about as a whole kind of local and voice, uh, machinery is moving rather than, you know, how their product itself is

Speaker 2:51 just moving.

Div Garg 2:51 I'm Div Garg, the founder and CEO of AGI Inc. We are building the world's, uh, truly personal AI that will be running on your phones and any other devices, including smart glasses and smart watches. We built our own on-device models that can go and take actions on your phone, so it can control your whole phone and use any app for you. Right now, we are actually number one in the industry on, uh, benchmarks like AndroidWorld.

Speaker 3:13 Welcome to the first and introductory lecture for CS25.

Speaker 3:18 I heard that you make the first transformer-

Div Garg 3:20 Yeah

Speaker 3:21 ... class in Stanford, right?

Div Garg 3:22 Yeah. I was at Cornell for my undergrad.

Speaker 3:24 Mm-hmm.

Div Garg 3:24 At Cornell, I wa- I was doing a double major. This was in physics and CS. I was just very good at physics. I actually did some International Physics Olympiad, uh, where I won, like, uh, gold medals in India. So, like, physics has always been easy for me. And then I joined Stanford for my, uh, graduate, uh, studies. This was actually when I was in my, in my first year at Stanford in the graduate program. I kind of noticed, like, okay, like, Stanford didn't have an actual course on transformers. It seems like the whole-- everyone in the industry cares about transformers because that's a big thing, and it's kind of weird that, like, that Stanford does not even have a course on it. And so it was me and along, along with, like, two other friends. We says... I had this idea, like, what if we just go and create this course? Uh, so we actually wrote, like, a three-page proposal that we want to create this course. And it's, it- and it's not normal for s- Stanford students to create courses. And so we actually wrote a three-page proposal. We sent it to the Stanford dean. And we were like, "Yeah, we, we're very confident, like, this will be, like, uh,

Div Garg 4:10 amazing." I just decided, okay, like, this is such a cool concept. I should just reach out to the best person I, I can invite. At that time, that was Geoffrey Hinton, who, who was, like, the godfather of AI and has won, won a Turing a- a- award. So I just went to his website, and i- his e- email is public, actually. So, so you can actually... Like, anyone can email him. So I just emailed him, like, "I'm starting this, like, very cool course series at Stanford, uh, with the first of its kind, and do you wanna come and give a talk?" And he actually replied in the ne- within 10 minutes saying, like, he's down. [laughs] And that's probably the best person you can invite in AI. So we got, like, pretty much every famous person in AI to come and give a c- lecture at some point. Andrej Karpathy, some of the co-founders of Anthropic, some of the research leads from OpenAI. So we were able to get, like, some of the best of the best people to come to Stanford and, like, ta- talk to students.

Div Garg 4:57 And I dropped out of my Stanford PhD and, uh, turned down a billion-dollar offer from OpenAI to build AGI Inc.

Speaker 5:04 Why was this problem worth betting everything on?

Div Garg 5:08 I think this is the most impactful thing we can do for the whole humanity right now. How you spend your day and how you interact with technology, I think that's what matters most. One way I like to des- describe this is, like, again, like, uh, going back to Jarvis from Iron Man analogy. Like, imagine, like, you had this, you know, like, a personal AI butler, uh, that just knows everything about you, and you can ask, um, it to, like, do things for you. The difference between you and a billionaire is probably the billionaire has a army of people who can go and, like, do things for him. And that allows the person to be, like, more... him or her. That allows the person to be more productive. But now if we can have an army of agents that can basically go and do everything for you, like, that gaps get removed. So you can imagine, like, the whole humanity gets equalized on the same basis, and we want to make that happen.

Speaker 5:49 How do you describe Steve?

Div Garg 5:51 I think his priorities, right, so he's very, like, organized on, like, what matters, especially as the company's growing. What are the things we need to do? He's also, like, from Stanford and, like, where he was doing some research

Steve 6:01 I'm Steve, co-founder at AGI.

Speaker 6:03 Why is this technology so hard to build?

Steve 6:07 I think initially I was more focused on action. If you ask it to do this, then get the agent to do that. We've done that, and now more of the rate limiters are memory and context, and having the understanding of you, even if you give it something that's kind of ambiguous, to still understand and be able to do what you meant, to recognize your un- underlying intention, and then ultimately to, to proactively know that you would want that and then to do it for you.

Speaker 6:38 Then what do you need to get there?

Steve 6:40 There's a relationship between trust and capability. If you trust the AI more, if you trust the tool more, then you'll give it more context, more understanding about you. You'll allow it to have more understanding about you, and if it has that, then it can be more capable and ultimately proactive. It goes into trust. Ultimately, we're working towards running it all on device, and that means that the data never leaves the device. There's no cloud to trust. That's the ultimate goal.

Div Garg 7:09 So this new orchestrator that we just launched today, uh, that we're calling Eve, uh, that seems really powerful. I think that's really powerful. Can you just make this PC loader get it fixed for all 40s of ASAP?

Speaker 7:17 I'm actually working on, like, another piece here that's, like, a very high priority. A lot of our clients actually depend on it. Like, if we don't get in, then we'll lose this million client.

Naman 7:25 I'm Naman. I'm one of the co-founders at AGI Inc. I'm leading most of the engineering and research. There is my brother, so I've known him since I was born.

Speaker 7:34 For non-technical people-

Naman 7:35 Yeah

Speaker 7:36 ... could you explain how AGI technology work?

Naman 7:39 Our stack is, like, divided into, like, three places. So you have your screen, and then you have, like, multiple apps. When a user sends a task, we get a task, and we also get a screenshot. Our agent is able to take a look at this screenshot and then understand within the screenshot what action needs to be performed, and it does it in a loop. So we have a action model that looks at the screen and predicts the next action. That action is divided into click, type, scroll, and this action gets transferred back to the screen to perform task, and then this keeps running in the loop until we have reached our end state. This is where, like, our harness is, and we have our inferencing engine. So our stack is divided into application layer, the agent harness, and then the

Naman 8:26 inference.

Speaker 8:26 Where is the bottleneck? What is the hardest part?

Naman 8:29 Getting the agent to be fully on device. This part is fully on the cloud side right now, and we are looking into making this running entirely on the device itself. So all the inference, everything happens within the device itself, and it's challenging because as soon as you reduce the model size, the capabilities also decrease. And there it's our, like, research challenge where we are looking into how do we ensure that model is still, like, as accurate as, like, running on, like, a bigger machines, and also it's fast enough. And this is what truly unlocks privacy and security for the users.

Speaker 9:05 What if OpenAI or Anthropic someday launches their new feature-

Naman 9:11 Right

Speaker 9:11 ... they can control their phone-

Naman 9:12 Yep

Speaker 9:13 ... with their voice.

Naman 9:14 Yep.

Speaker 9:14 What happened to AGI Inc., and how is it different from that?

Div Garg 9:17 So a lot of stuff we are doing is, like, actually will be running on the device itself, and it's also very different from the business models frontier labs have. A big part of our vision is, like, you're not renting a model running on someone else's cloud. So it's, like, something that will be, like, running, running fully on your device. So, like, you own the model, you own the data, and this is something that frontier labs don't, don't want. So for example, like Anthropic is not, like, ship your model that will run on the, on the device, right? Because how will they

Div Garg 9:39 make money in that scenario?

Speaker 9:40 Our team uses Wispr Flow Notetaker when we work across time zones. Right now, two of our designers are making odds merch between Seoul and San Francisco. Their time barely overlaps, so they get one call, and nothing can get lost. That's where Notetaker comes in. Before it starts, Notetaker already knows who's joining from the calendar invite and their own terms from the dictionary. No bot joins on Google Meet or any other platform. When the call starts, it just runs quietly in the background.

Speaker 10:05 Hi, Jiyeon. How are you? Hi, Soyoung. I'm good. So here's the cat watermark sample for our new series, Odds.

Speaker 10:13 When the meeting is over, it double-checks its own work. The summary's done with real names, not speaker one and two. The next morning in Seoul, a teammate who wasn't on the call asks Claude what was decided and picks it up from there. And the merch made it all the way to the AGI, Inc. team to San Francisco. Wispr Flow Notetaker is free to use, and you can get one month of Wispr Pro with unlimited meetings with the link in the description.

Speaker 10:38 [upbeat music]

Speaker 10:42 My name is Will, and I covers the scope of growth as well as the wearables and, uh, hardware departments.

Speaker 10:47 Can you share what can we do with AGI?

Speaker 10:50 Absolutely. So everyone's seen the traditional demos of ordering Ubers and, uh, responding to emails, but there's a lot of other things you can do. We spend a lot of focus on the consumer side as well as the accessibility side. Imagine if, like, the, the grandmother could just say, "Hey, can you help me order, um, DoorDash for some fresh sundubu?" Or, "Hey, I don't know how to use my phone. Can you pull up a FaceTime and call my

Speaker 11:13 grandson?"

Speaker 11:14 Then can it, like, order lunch for us right now?

Speaker 11:16 Absolutely. I'll show you how it works. Right here you see a bunch of presets from all the different channels.

Speaker 11:21 Mm-hmm.

Speaker 11:22 But we'll just do a custom prompt today. So, uh, I guess there's, like, four of us. We'll do, like, a custom, like, Chipotle set. I'll let you pick, like, the, the ingredients.

Speaker 11:31 Hey, um, I want a Chipotle, uh, steak, brown rice, lettuce, corn, guac, salsa, sour cream, and some cheese.

Speaker 11:42 And, uh, order it on DoorDash. So it just pulled up the app hands-free. As you can see, it's hitting each individual, uh... So right now it's clicking the steak. Got the brown rice, black beans. So it just placed the order. It's about to place order, and then right here it says a prompt, "Would you want to place this order for $91?" We always have it as, like, a guardrail. Nice. So you guys can see the order is placed, and, uh-

Naman 12:23 Task complete

Speaker 12:24 ... task complete.

Naman 12:25 [laughs]

Speaker 12:25 That's dope. Hell yeah. [upbeat music]

Naman 12:33 How do you make that demo into a real business? Yeah. You need to find, uh, customers, and you want to have people, like, users who love your product. Uh, okay, like, it's like, okay, like, you can build this new cool technology, but, like, why would someone buy it? Why would someone want to use this? And what value can you create? And so we've been spending a lot of time, like, understanding, like, the friction pain points people might have in their daily lives.

Speaker 12:54 So at that time, we still had a lot of ideas. We're, uh, trying different ideas. One time even we went to the, uh, Golden Gate Park, and there was this, like, music festival going on. We went in front the, the, uh, festival. I was giving away the, like, Powerade and stuff to kinda get engaged with the potential customer and then see if they would get any feedback. And then we had, uh, um, the meeting already set up, so it's not something that I can take months. So I have to really understand what I can deliver out of the million expe- uh, iteration. Finally, the... I was able to invent the One Day Works.

Naman 13:31 Mm.

Speaker 13:31 I was literally crunching, like, two weeks to literally sleeping in the office, like, three, four hours to really, uh, get things done. Miraculously, the... we... I was able to invent the, the first time ever, uh, the Android agent that no one had been ever built.

Naman 13:48 You guys are, uh, partnered with big companies like Qualcomm- Yeah, yeah ... and other big companies. So how did you get those companies?

Steve 13:56 Like, it felt, it felt very pure in a way. It was like, it was tech demos. They'd say, "Wow, I've never seen, I've never seen a phone do that before." And, uh, that's our phone. That's our chip that's running that. Cool. Uh, what more can we do? I think it's that we saw that there was a need for this by these big companies and no one was filling it, and we could. I think it's like pick your battles. If you can automate spreadsheets, that's a great business. You will do well if you can fully automate everything people do in spreadsheets, but there's 10 other companies going after that. Large proven market. There's this more unusual path where we can build it for edge devices, and there were definitely companies that wanted it, especially the ones that make the devices. No one was really going after that. I think we're still the best. Companies we work with still say we're the best in, in that category.

Naman 14:49 I would say there's a big difference between building a, call it like a demo prototype versus a production level technology. Because, like, it's possible, like, the demo only works 10% of the time. It's fine because, like, you're, like, testing out what's possible. But then if you're, like, shipping something to a billion plus users at some point, then you just need to, like, rigorously test it. So a good analogy is, like, self-driving cars. Like, if you think of, like, Waymo and Tesla, they have... They actually started, like, working on self-driving cars, like, uh, more than a decade ago, and it took them, like, 10 years to go from, like, something that was a concept to an actual thing that can drive on the roads. Because there's so many, like, edge cases in the real world where, like, things can fail, like, so, like, something might go wrong.

Speaker 15:24 I think with the complex ones it's, like, 20, 20-ish percent.

Speaker 15:28 In the kitchen I saw you were explaining and showing something to-

Speaker 15:32 Yeah

Speaker 15:32 ... founders. Could you briefly explain what was it?

Speaker 15:35 Oh, that was our baselines of our agents. So basically we were trying to categorize how well we perform across different apps. Some apps it was good, some apps it's not the best [laughs] because, like, users are very unique. And what I mean by that [laughs] is, like, some people ask some really interesting things to do with an AI agent.

Speaker 15:58 For example?

Speaker 15:59 One person asked to, like, file their taxes. Like, so, you know, [laughs] uh, can't do that.

Speaker 16:03 By their phone?

Speaker 16:04 Yeah. So, uh, that's what I mean. There's some, [laughs] like, pretty interesting use

Speaker 16:10 cases.

Naman 16:11 A single agent can be, like, perfect in a demo, but as soon as you take it to the real world, like, as soon as you take a agent from a sandbox environment where everything is fully perfect to real world where there's so much chaos. And for us, like, that's where, like, the real challenge and real value add shows is can the agent actually do... work really well on production? If the agent is, let's say, 96% accurate, on a single turn that is fine, but let's say you're running across 10 or 20 different tasks, those errors actually compound. So if a agent was 90% accurate, after 20 tasks, it actually decreases around 40% accuracy, and that's what we are, like, trying to solve, and that's what, like, people are coming to us for because we are the only ones who are actually really solving that challenge.

Naman 16:54 So when memory can decide-

Speaker 16:55 Yeah, because I, I, yeah, I feel like that would give us more context. Okay, so, uh, I'm, like, allowed to kind of do a whole kind of redesign of what is being stored in the support system?

Naman 17:03 Yeah. That is done. So maybe create a plan-

Speaker 17:05 Okay

Naman 17:06 ... and if you want to experiment it, and then n- show me. Okay, these are the improvements. This is how this could just be better.

Speaker 17:11 Yeah. I, I'm, I'm working on making a plan. I'm not editing anything in the code right now.

Naman 17:15 Yeah.

Speaker 17:15 So I'll show it to you by the, by like, in, like, two, three hours, and then-

Naman 17:19 Okay

Speaker 17:19 ... uh, I can make it run.

Naman 17:20 I could also invite you to the both Cloudflare and Warsaw so that way you have access to, once you make changes-

Speaker 17:27 Okay

Naman 17:27 ... you can actually spin up the environment to test it on your support.

Speaker 17:30 Cool. Yeah. Perfect. Okay. Yeah.

Speaker 17:34 What did you ask to him?

Speaker 17:35 Uh, so I was asking him about the data that is being stored on the database. It's not seen very quickly at companies to be working on production grade code at the very start, at the very first day.

Speaker 17:48 When you hire a new employee, how do you decide what are they gonna be in charge of?

Naman 17:53 So it just depends on the role, like, what's the role and, like, in what capacity are they joining us. So typically what we do is, like, we actually do, like, trials And then do a full-time hire. So in this case, he's on a first week trial. So we give him, like, a, "Here's the feature implementation that we want, and here are the outcomes." We are seeing, like, how good are is the person in terms of, like, taking this, like, a bit more ambiguous request in and actually coming up with, like, a really good solution. So what we're trying to do is also see how good are they at, like, coming up with product UX. And we wanna do, like, m- more, like, brainstorming and iteration cycles. So here's a feature you can think of, or this is what I see is, like, really useful for the user, and then we wanna iterate and, like, keep refining it. We see this as, like, a constant iteration. It's you just keep refining, optimizing it until it keeps attracting the users and make them actually truly sticky users. And for full-time employees, we actually like them to be, like, independent thinkers. Typically, I wanna basically be the person who's validating the idea is correct or just, like, be the final layer of judging. [upbeat music]

Steve 19:07 Hi.

Speaker 19:07 Hi. Did you finish your meeting?

Steve 19:10 Yes. Yeah.

Speaker 19:10 Do you guys usually have meeting at this time?

Steve 19:13 Yes. So that one started at 10:00, and that's with the remote team in India, so it's morning for them. But it's good 'cause we have space. We have space in the evening. No one else will bother us. We can just focus on product.

Div Garg 19:25 Yeah, I should just make sure I show this name so everyone knows what I'm gonna show.

Steve 19:28 That's a good, good-

Div Garg 19:29 Yeah

Steve 19:29 ... idea.

Div Garg 19:30 Uh, we can choose the IS option for the real stuff.

Speaker 19:33 AGI Inc. seems like there are many people who are night owl. Where, where does this night owl-

Naman 19:38 Yeah

Speaker 19:38 ... culture come?

Naman 19:39 I sleep pretty late, like, sometimes, like, 2:00 AM or later. I just find it easy to focus at night because you don't have any distractions. No one is pinging you or sending an email. There's no distraction. There's nobody pinging you. It's just you and, like, the things that you are trying to solve, and that's why we see the, our culture where, like, people like to stay longer because that's where they see they wanna keep contributing. They wanna get it.

Speaker 20:01 My personal question is, is that, like, really sustainable, working 24/7?

Naman 20:07 I think, like, if you love what you're doing, it doesn't matter. It's like if you really are looking into building something that you truly see as a really value add, people would wanna use it. It's like a meaningful sacrifice that you are going to do. I would say we try to be very autonomous. There's a certain independence on things you're working on, as long as you're, like, performing at, at a certain bar. So you, you are, like, in charge of your decisions, how you make them happen. People really like it because, like, they can, like, un- unleash all their creativity and their talent. And I think that's very fulfilling, um, to, like, see, like, something you've created by yourself is now getting used by millions of people over time, and I think that's, like, one of the most fulfilling things that you can do in your life.

Speaker 20:45 [upbeat music] Happy birthday to you.

Speaker 20:54 [clapping] Woo!

Speaker 20:56 In-person events are expensive. You have to pay for food, staff, and everything, so why does AGI keep hosting events?

Speaker 21:03 Events are, like, big for us. We're building for on-device, so we need a lot of very smart researchers to join our team, and we just, like, love bringing them together and discuss the topic. Digital becomes discounted, uh, a lot. You know, it's like you don't know if it's someone's agent reaching out to you, and you don't know if the information is accurate. And, like, I noticed a l- like, a lot of times if people are recognizing AI slops, they just, like, stop reading. So meeting someone in person, in person is one of the best ways to actually build connection. And sometimes you don't know what's going to happen if you just have one discussion with a person, right? They can make an intro. They can introduce you to a client. They can refer someone to work for you. So overall, events, they just, like, compound a lot, and they, uh,

Speaker 21:46 drive a lot of serendipity.

Speaker 21:47 What's your favorite thing in this company?

Speaker 21:50 My favorite thing in this company is the, uh, the trust behind the creative freedom we're given. Early on, you identify, like, what you're good at, and then they just give you the freedom to really, uh, build out, like, the project. For example, we'll do a, a company trip, and we'll go release, like, 100 mini phones. Yeah, so, like, this is, like, an example of, like, a mini phone, right? A little bit chunky, a little bit like an iPod. Uh, the thought process behind this is like when you see something like this, it sparks a sense of curiosity. And you ask a question like, "Huh," like, "Are you texting on it?" Like, "How are you interfacing with this?" It's like a modern-day pager. You can, like, speak into it, and it makes all the actions for you. So this is just, like, one of the many different, uh, interesting experiments we're cooking in the company.

Speaker 22:30 So you love AGI.

Speaker 22:32 I love AGI. This is my actual first contribution to the company.

Speaker 22:37 Oh.

Speaker 22:38 I didn't know that.

Speaker 22:38 You made it.

Speaker 22:39 This is my first contribution. Uh, ever since this project, I've been, uh, I've been, uh, part of the family. The biggest part, like, I don't know. Look at this g- look at this man right here. CK, this is the big brother of the company. Without CK, we really wouldn't be anywhere.

Speaker 22:54 CK's a big part of team culture.

Naman 22:56 Culture is also very dynamic. I feel like a culture cannot be enforced. A culture is formed

Naman 23:05 by every single person that joins the team. Most companies have this tendency of hiring fast and firing fast. We take a different route. We hire really slow, but we fire very fast because it's difficult to find culture fit in a person. Someone can be a great performer. They can be the most skilled person in their industry or their job, but you don't really know the culture fit until they actually come and work with you. And so we're very cautious about the people that we bring into this environment because every last person that joins this company adds

Naman 23:37 on to the culture. In the future, we believe there could be, like, billions of personalized models, and, like, every person gets their own model. That model can be just, like, self-learning. It's learning things about you, your preferences. It, it's getting better and better. And you can fully trust it because you actually have the model. Like, it's like, like it's yours. It's like how you have a car, and you can modify your car and put things in and whatever. It's not like you have one generic car that serves every single person on Earth, right? And that's what's happening with models right now. So we want to have this v- way where you can buy the model, you can customize it, it learns things about you, and then over time, like, everyone gets their own very, very personalized model over time.

Steve 24:07 And so choosing that AI, which AI is doing that, what went into the design of the AI that does that for you, is very important. That might be a more influential relationship than any human relationship or a- at least as influential of a relationship as any human relationship.

Naman 24:24 And what we see is it's gonna keep changing. I really wanna be there in the world where, like, I actually contributed in making the world better. I wanna see how far can we take it, and, like, can I actually really contribute in the world, actually making everyone's lives better?

Steve 24:38 At the end of the day, you don't, you don't go to an everyday person and say, "Agent.

Steve 24:42 I made a great agent for you." One thing it can mean for people is that you look at your screen less, and you spend less time on the screen and more time in real life. You can just talk to your AI from earphones or your watch or whatever, and then it can be doing things for you in the background while your phone is in your pocket or at home. Now, the way that we think about what we're building is not just, oh, we're making this agent to control apps, but it is this whole system. It's this new operating system for devices that help you look at the screen less and be more in real life.

Speaker 25:15 Actually, we brought the small gift for you guys. These are the small gift-

Steve 25:20 Thank you

Speaker 25:20 ... for our guests. [laughs]

Speaker 25:22 Ooh.

Steve 25:22 Let's do it.

Speaker 25:23 Looks good.

Steve 25:24 Yeah.

Speaker 25:24 [laughs]

Steve 25:26 Cheese.

Speaker 25:27 Cheese. So thank you, guys.

Speaker 25:33 Thank you. [upbeat music]