BaseCamp 2025 — Building Agents People Can’t Stop Talking To

Awesome. So, is everyone familiar with agents? Have you guys used Rocky?

Yeah. Okay. Matteo over here built Rocky. Round of applause. Give it up for him.

In one week also.

Cool. So, this is our agenda. We're going to go through why agents, why messaging, and then we're going to cover some product principles. And then we have, I think, five demos today from a bunch of builders in XMTP. And then, yeah, we'll talk a bit about how you can build with us and how you can get started.

So Shane's gonna kick us off. Why agents?

Great. Thanks, Claudia. How we doing? Doing good. All right.

I was actually saying to someone earlier, I usually hate event bots. I've been in messaging for way too long. Um, and I built a lot of bots on top of Messenger and WhatsApp back in the day. My first company in '09 was acquired by Blackberry. I worked on BBM.

RIP BBM.

It was a good product.

I loved BBM. Uh, it's a good reminder of how wrong you can be about needing a keyboard. Um, but I think seeing what you did with the advention was fantastic and it splits you out in different groups and the the mini apps and the agents working together and it just feels like a whole new surface area. So for messaging I mean it's 50% of the usage of the internet is messaging apps and the thing that I think we need to think about is what's different here and what's different here is there's a native financial rail that's global and money plus chat to me is the the largest surface area to build things that we've seen in the last 10 years. And it's not even just about the agent.

It's actually about many apps too and how it all works together. The agent is just the interface that is a better thing for you to use in groups, which is what we do all day. And I think when I think about it, the biggest opportunity is that these messaging apps have been so closed for 15 years. And so you can't build on top of iMessage. You can't build on top of WhatsApp.

You can maybe build a little app that you can use inside of it. There's no access to money. So you have global messaging. you don't have money. And these closed platforms have been killing innovation for over a decade.

And to me, the opportunity here to really go build on open platforms that combines the use case we use every day that everyone in the world understands with a thing that can take action and from a word, from an emoji, from a reply, it can place a bet, it can send money to friends, it can do a prediction. And you know, I was using Banker the other day and there's 25 different people from seven different countries and I just said, "Hey, Banker, send everyone 10 bucks." And in the the experience, it's kind of simple. underneath the hood. I'm like sending seven different countries USDC in their local currency under the hood in one click to anywhere in the world instantly for free is one of those moments when you see it pop up on the screen and it literally went and built a smart contract that loads up money in every single person's wallet.

That's the foundation to build [ __ ] that we've never even thought possible. And when I see things like that, WhatsApp can't do that. Vinmo can't. You can't even send a Vimmo to someone in Canada, let alone seven different countries. And I think that foundation of secure chat, money, and agents.

And actually, we announced yesterday, if you go to xMTP.org/build and chat, agents now are connected to many apps. And so, you can make every mini app talk really easily with the new SDK. It's not about one or the other. It's how it all comes together to use in a group.

And I'll end on I think the most important thing about building agents when they can transact is it's not about agents that say more. The worst thing that happened with the agent world is they're so verbose and they never shut up and they just keep talking and they're so long- winded. But actually the magic, the design of things that can transact is the speed you go from a prompt, a reply, an emoji or a message to the magic moment and everything else is hidden from the user. And you know, Banker is a good example. Brackchie has it.

I mean, I went in Brackie and I, you know, you literally, I think, placed a bet on an F1 race. In the middle of the race, the response back was that the F1 series of the positions would change because it started raining. So, your prediction probably with the tires he's using is off. I have no [ __ ] clue about F1. I don't even understand why.

I have no idea. And I was like, I didn't even I was like, I just bet on Versappen cuz I heard he's cool. And I replied back and said, yes. and it literally went changed the whole thing, sold my position, put it in the other person that was the highest prediction market to and like I don't I was just like blown away and then he responded back with okay and put another hundred on this and I responded with the handshake emoji and the emoji places the bet and when you see moments like that I just want to encourage everyone to don't think it's like don't think about it as oh it's just agents versus mini it's this experience where you spend every day with your What's the fastest way you can go from prompt to magic? That's the game.

It's not speak a lot. It's not a thing that can chat a lot. That's bots. Let's not build bots anymore. That was that was old old can't really understand.

Can't keep it. You know, this is about simplicity and doing things behind the scenes that makes a group experience with people magic. If you're with a group of 10 people and you're having dinner, you should be able to explain a use case, everyone joins and have them all feel the experience and the magic in under a minute. And I think even under 10 seconds, that needs to be that quick. That needs to be the goal because when I w I'm in Nashville and all my friends are sports betterers.

I don't really bet on sports. And they all do this. One guy has a DraftKings account on a VPN because it's illegal to do it. That guy's the money banker. He has a notes app.

He then has a group chat because he doesn't want it in his iMessage chat. And then he's talking to all his friends. They're paying in on Cash App, Vinmo, Cash, people that ran out of money because they lost other bets or IUS. And there's just like a whole list of things happening. And then they're talking all day.

And then he goes back, he places a bet. They're all going to Google trying to find the secret tips, trying to look at what's the right answer. Coming back, hey, oh [ __ ] I want to place another bet. Okay, put it in the notes app. Vinmo me more money.

Go over here. Do this. Oh, the VPN's down. [ __ ] I gotta put my VPN in Missouri now. Great.

Okay. And like it's insane. That's literally what happens. And all of a sudden, I have an experience where I'm talking about it. Someone says something about it.

I can do research about it. I can ask more questions about it. I can learn from someone else in the group doing something. I can observe others using the experience. There could be a mini app that pops up to see all of my past bets and whatever.

And the fact that I can then say, you know what, Alcarz, US Open, $10,000. And then he says, "Actually, yeah, let's do that US Open, too." And then 50 replies come in in a group. I've never seen friction reduced that much. And for me, the last thing I'll say is when I see friction reduced that much that can do something that unbelievable.

And you can kill eight apps with one, that's where the magic lives. And it's really hard to do. Building agents is not easy. It's the most messaging and agents are the single most um I would say underestimated surface areas in technology. Everyone's like I'll build a messaging app.

Like cool. And then you start and it's really really hard. Claudia knows. And then you go to agents and you're like I just I just do a little agent. But the agent needs to do the work for the user.

Don't make the user do the work. Right. It What's this thing called? How to make agents that you can't stop talking to. You know how to make them do that is shut the [ __ ] up.

The agent should stop talking. Go do the magic and come back and make it happen.

And the speed from prompt to magic moment is the game, not speak more. And I think that actually is one of those design philosophies that's really hard until you really get into building agents for a long time. manage the complexity and coordinate all the systems and bring eight apps together and you have a way to look at a prediction market and go look information on Google and talk to Chad GPT and talk to a payments rail and see what the liquidity is and go across Koshi and predictions and draft kings and poly market and then bring back and say this is the thing you should do that's hard but that's where the magic lives and that's where the money's at and so I think this surface area is about many apps it's about agents and it's about group chats and it's about a global financial rail where anybody in the world can be in the same chat and it doesn't matter where you live the money works and Telegram can't do that iMessage can't do that can't do that because they don't have the native financial rail and they have financial laws in every country that they can't turn it on and there's a window here where they're all going to do it they're going to turn on wallets for everybody in 5 years they tried to do NFTTS with Instagram but there's a 5-year window of so much regulation that is the next messengers will get built and the next social experiences will built and the agents are the new app and chat will be the interface and all of you can sit at dinner with your friends and family who don't understand what the [ __ ] you do for a living and they'll get it. Yeah. If there if there's any other apps that launch uh all of your agents and many apps work in those new apps.

So anytime a new application launches on XMTP it's fully portable. So more distribution more apps. There's another major tier one app launching really soon. Um, another app that is someone here launching pretty soon. Um, so I think you can look out for that.

The cooler thing is that it follows you around. So the Brackie experience is a great one. My bets that I'm talking to Brackie and I'm in a friend a chat with eight friends. I'm then going to a new group group chat with six friends, but my bets with Brackchie come with me. And so then I'm talking to those friends.

I'm like, "Hey, look what I'm betting on over here. All my stuff's there." And the fact that it like comes with you across surface areas because there's the same kind of you know uh agent financial rail is is a really cool powerful experience. And so I'll just encourage all of us to think smaller to think bigger here. Build agents that actually do less talking to build better experiences and make agents and mini apps work together seamlessly.

And our goal with the new SDK is you can make any mini app talk. And so with now you can go anyone who's built a mini app click deploys an agent and now you have distribution and group chat for that mini app and now that mini app has retention. People can invoke that mini app. People can use replies and emojis and all these other things to make the the digital experience great. And I think this combination of agents and mini apps and secure groups and money is just a playground to not be underestimated because every single person in the world gets it.

And if I'm a developer and I'm building an app, I want to be able to sit with my friends and show them the experience and we all use it and they like it. Uh, and that's the that's how that's how [ __ ] really spreads. I think the idea that it all spreads in some magic internet way, the biggest consumer companies actually spread word of mouth with friends. Uh, and that's the opportunity of groups.

Awesome. Thanks, Shane. So, now we're going to go through product principles for building agents in the base app. We already have 10 amazing agents featured in the base app. a lot of them are here and yeah just going to walk through like if you're interested in building like what you should keep in mind and like what we're seeing drive success for agents.

So yeah four principles here the first one is like just making it really simple like it should do one thing really well and be easy to understand. So like brackchie sports betting makes total sense. Squable, play with friends, earn money makes total sense. Bite, buy food on chain, use USDC, it's fe no fees. Um, so yeah, just having like one clear value prop and really think to yourself like how would I explain this like to my mom or to my friends not in crypto.

Like would they use it? Would they understand it? And that's like a good guiding principle. Um, the second thing is social. So like what we're seeing is that agents that work in group chats tend to perform the best.

So, we just had a fun time in our group chat and like really designing your agent to be like easy to talk to, funny, things like that really go a long way. Uh, the third thing is rewarding. So, something we think about on the base app team a lot is reducing time to earn. So, if your agent can help people earn money, like you're so many miles ahead and that's like a really good way to like build your agent for success. Um there's also nonmonetary rewards like just like social connection like fun like so think uh broadly with that but like in general like monetary rewards do go a long way.

And finally the best agents are sticky. They're not designed for one-time use. They really like build a habit and like get people coming back day after day. So yeah going a bit deeper into this um the first thing first principle is just to make it frictionless. So the onboarding message is like the front door of your agent.

So really recommend investing the most time here. I'm working with 20 agent teams right now. And like the most common feedback I give is to make the onboarding message more clear and more actionable. So Momo did a really excellent job here with these quick select buttons where it's like, "Hi, I'm Momo. How can I help you?"

You click I want to get started. You learn a bit more. You get the context and then I can just say I want to deposit USDC and then you just say how much. So I barely had to type. It was like so easy.

It really makes DeFi like way more um intuitive and less scary. So yeah, there definitely are some agents that you want to type more if it's like a social agent or like a chat GBT type of agent, but in general, if it's more um transaction based, I think making it um more easy to use goes a long way. And that usually is less typing. Second is um once you have that foundation done, really giving your agent personality is a good way to stand out. So your agent should have a presence.

It should like make group chats more alive. And even in DMs, if your agent's just delightful to use, if it's polished, if it's funny, if it's um kind to you, like people will want to use it more and talk to it more. So once you have the foundations down, like really recommend investing in like the tone and the polish and like working with a design a designer or a marketing team to really like figure out the brand of your agent. Having these Easter eggs of humor can make your agent go viral more easily. Cool.

And then on the rewarding front, so putting value front and center, whatever your agent is uniquely going to help them with should be super clear in the onboarding message. And minimizing time to earn. So the Flaunch team did a really excellent job here where they show you like when you're launching a token, it's like you're going to split the earning fees with everyone in the group chat and it's telling you how much you're going to earn. And then once you start earning, you actually find that out in the group chat as well. So I really like this like prominent focus on earning right in the chat.

And this is also good because as people earn, they're going to keep using it. They're going to keep telling their friends and it's going to be a really good growth loop for your product. Finally, as I mentioned, you don't want people to use your agent once. You want them to be longtime users. So, really think about the endto-end journey.

Like, I'll use Momo as an example. Like, let's say I stake USDC on Momo. Um, how am I going to come back? Like, how am I going to like keep like hearing about my earnings and like coming back to the app? Um, so it's really good to think about like natural growth loops.

This could be like daily challenges. This could be uh a leaderboard and streaks to like get people like into the game and like addicted to the app and even like reminders and automations. Um, Banker added like a good automation where you can like set limit orders or set price alerts and that like notifies you every day. Um, but yeah, Squabble also did an excellent job here where they have a leaderboard. So like once you play the game, I used to lose like every day when I played this.

Um, I'm like so motivated. Actually, I'm third place. I'm not doing too bad. um I'm so motivated to come back and like climb the leaderboard. So, I think like having that history and that compet competitive aspect really goes a long way.

As you might have seen, agents are discoverable on the home feed and in the chat tab. We're also going to make it easy to add them to group chats directly, like coming next week, I'll say. And yeah, so it's a really good place to get distribution for your product in the base app. And what we're also seeing is that most agents also have a mini app. So, if you're building a mini app, highly recommend also building an agent.

It's pretty quick, pretty easy, and it's a low lift way to increase your distribution. And yeah, at Bass, like we love building in the open. So like once you have your agent working, like show it to your friends, add it to group chats that you're in. Like just get people talking about it and giving you feedback. Um post demos on the base app, post on X.

And we also have an agent developers group that I can add anyone who's interested in. And that's a good place to connect other devs and get feedback. So, our road map um coming up in the next few months, we're going to add markdown support for more UI customization like links, bolding text, things like that. We're going to add typing indicators so it feels a bit more like chatgbt or clawed. Yeah, it'll be great.

Um and finally, we're very excited about like spend permissions so you don't have to like approve every single transaction. The agent can um execute more autonomously. Cool. Here's a link to the documentation if you need it. Um I can also send this in the chat so no pressure to scan it.

Cool. Um, now we're gonna switch to demos. So, I think Squabble's gonna kick us off.

Hello, I'm Lemon, uh, base builder. Uh, mini app builder. I'm running builder's garden. We are mini app development studio, but we also build agents as you can see here. Squabble um was part of the last base batches messaging track and it's um, Scrabble in real time with real money bets.

And so the idea is that you can invite squabble in the group chat. Um we already passed that moment, but you invite the agent. You ask the agent to set up a new game. Uh you can set an optional buy in of I don't know $1, $2, whatever. People join the lobby.

Whenever everybody staked and and they can stake from uh any chain with any token using uh the game starts, so it's about you know words with friends, but you know the everybody is placing on the board letters at the same time. So this main goal of the game is around uh scoring more points of course. So placing more letters on the board and at the end the agent automatically distributes the uh pool uh amount to the winner. Um, of course, the agent is also keeping track of the state of the game, meaning that whenever somebody joins the lobby, whenever somebody is ready uh because they staked their buy in uh or maybe they left the game, uh the agent is always notifying the group chat about it as well as when it comes uh to an end. So, it's congratulating the winner.

It's uh showcasing the leaderboard as Claudia was was showing earlier. Um and yeah, I guess that's that's pretty much it. It's um very quick you know type of uh uh of game like two minutes round usually. So um happy to see you playing it out and if you have any feedback uh please come to me.

It's very addictive especially if you do like a high buy in then what's what's the highest buyin you've done?

Um I think like we've seen like um chain we did a high bahin uh at the beginning. I don't remember how much it was like 100 100 USDC probably if not 1,000 and then I I sent you them back because it was very high. Um

nice. Are there any features you're planning on adding to Squabler or like what do you think is next for Squabble? Uh I I think like the inline actions that you were also showcasing are a good way to uh you know um improve the agent experience cuz right now uh with Squabble you can only interact by typing which is fine but you know the buttons are a good way to help people you know creating the first game or setting things up.

Yeah that could be good for the buyin. Awesome. Thanks so much Leone.

Thank you. [Applause]

Next we have Matteo from Rocky. Hey everyone, I'm Matteo. Um, and I want to introduce you to Rocky. I'm sure everyone's, you know, used it by now. It's been helping everyone with their scheduling, adding people to groups, uh, sending out reminders, broadcasts, and just helping everyone have a great better experience here at the event.

Um, the story started a couple weeks ago at the onchain summit hackathon uh, in San Francisco. So I built the first version of the event agent there, the summit concierge agent, connected with base devs, other base builders and after the event I started shaping the agent specifically for base camp and that's kind of how Rocky was born. Uh the idea behind Rocky is simple. Every event needs an agent. If we think of events as mass onboarding experiences for base and the base app, uh the best way to make you know then agents would be the best way to make that onboarding seamless.

Uh, Base Camp, Rocky's been supporting attendees with questions around the schedule, sending broadcast, and helping guests join groups. The agent works with both natural language requests and quick actions to help keep the flow smooth and keep people on track. um whether it's pulling up sessions, dropping people right into group chats where they can discover other agents like Flanche, Banker, Brackie, um it's a good way to connect and kind of give exposure to the other agents in the base ecosystem by creating those groups, adding people to the groups and it's just kind of working as that uh infrastructure or connector. Um and this, you know, this agent isn't just for Base Camp, you know, I can adapt it really easily. And here at base camp actually I've connected with dev connect doge day um base South Korea and all of these events should have an agent um you know on base.

So you know that's Rocky like a small hackathon project that turned into a um you know this. Yeah. Thank you.

Cool. Next up is Mo Luke.

Yeah let's just talk about it. Okay. So first of all my name is Luke. Many of you know me. I was the founder of Moonwell like four years ago and I'm also the founder of Momo like I don't know six months ago.

So just quick origin story like in DeFi summer 2020 when we were all locked down and I was working at Coinbase launching staking rewards we started dejanning on a bunch of food coins like you know yam and urine and like sushi swap and like it was a really magical experience but at the same time as everyone was locked down and like people were dying from COVID like we were making lots of money like farming airdrops and I learned a lot about traditional finance just in the crash course of DeFi summer. So my friends told me like, "Hey, you made like you turned your Bitcoin your 100K of Bitcoin into like I don't know seven figures really quickly. Like why don't you create a fund?" And I'm like, "I don't want to do that. I'm a builder."

So um I always wanted to create uh like a robo adviser like five years ago that could basically learn what I learned, which is the power of compound interest. And like you don't need to like dejen like the the bros at the sports betting thing. You can actually just like dollar cost average and do all those healthy habits that we all are told to do like flossing and brushing your teeth which is like put a $100 of every paycheck into like you know a USDC stable coin vault that's lending to Bitcoin borrowers on Coinbase and like you're earning two or three the rate of inflation and like just doing that and stacking that for $100 a week um or a month or whatever you can afford out of your paycheck you're going to ultimately do better than all those you know DGens that ed the sports betting or trading memes or like trading um call options on Robin Hood because those guys might make $10,000 one day, but then they lose $100,000 next month when the markets crash. So, I distilled my five years of yield farming expertise that helped me grow from like six figures to eight figures and then, you know, I don't know more now. And then when I was at Coinbase into Momo Agent and we launched the Coinbase AI hackathon, I think uh it was just an idea and I like told my team I was like, "Hey, the Coinbase AI hackathon is next week and we have like a team of, you know, now it's like 20 devs mostly in Brazil."

I was like, "Do you guys want to come participate with me?" And some of them were on summer holiday because, you know, in January it's like the summer holiday and if you have kids that's when they're out of school. But half of them joined me and they were just like, "Yeah, let's show up and just like be at Coinbase." And now we're a ventureback team and we raised $25 million four years ago when I left Coinbase. But we have resources you don't have.

But the fun thing was like all of a sudden my entire team that had built Moonwell like over the last three years was like super engaged. We're building the future of AI and we were like and it resonated in one of the hackathons. So we're like we have to make this into product. And half my team was like, why are you building AI agents, Luke? Like you were building Moonwell for four years.

What the [ __ ] Sorry for the language, but um so we we kind of learned a lot just by kind of, you know, going deep on on what should anic experience look like. And the the short version was Moonwell was created to make uh lending simple because early DeFi apps were too hard to use. But as nice as Moonwell is, your mom probably couldn't use it without asking you like what wallet do you use? Do I use Coinbase wallet? Okay, how do I download it?

What's the seed phrase? like all these things that Coinbase is solving with TBA, they're still not solved problems. And so the idea is that MAMO is your personal finance agent, not your crypto agent, but finance meaning it should be very simple for someone that has never used a crypto app. You can just say go to mamo.bot or load m or talk to mamo in the base app and just say and ask it a question like, hey, where does the yield come from?

Well, there's Bitcoin borrowers on Coinbase that will pay twice the rate of inflation because they want to go buy a Lambo. And like your mom doesn't really need to know that because really it's just stable coins and it's the safest possible collateral because it's all regulated and it's approved and it's like now legal with the passage of the Genius Act. And so like you can get twice the rate of inflation by just lending it to some Bitcoin bro that wants to go buy a Lambo or a Rolex, right? But or more likely a elder millennial like me that just bought Bitcoin a few years ago and has been holding it this whole time and maybe wants to send their kids to college. And so that's where the yield comes from.

But then most people don't need to know that. It's just like, hey, I deposit dollars into the Sing. I can onramp up to $500 through my Visa debit card. So, I can go from Bank of America paying 0% to Momo with $500 in the base app in like a few minutes or less. And I and then the the Momo agent will just talk to me and say like, "Hey, here's how your money grew in the last week.

Way to go. Keep it up." And so the this gamifying kind of like making this this experience that people don't even really have to think about um whether it's crypto or not. It's just money because we've gone from like I don't know if you guys noticed but July 18th we went from being the crypto economy to just being the economy and everything is changing really rapidly. That's what Jesse was talking about.

So we just want to create make momo like an app for normies but also it's also an app for dgens like me who now I can just like basically the the key insight we had is just like I built staking at Coinbase is that you know staking needed delegation in the early days because you can't just delegate a hundred billion dollars to the Coinbase validator node and hope that it never loses your Salana or whatever but or your Cosmos or your Tezos or your Ethereum but agents need delegation and so we designed MAMO with a smart contract system just kind of like what you guys are building session keys where you know you give the agent limited rights so it can move you between vaults and markets to make sure you're always earning the best rate on your USDcoin or your Bitcoin but then you have control your funding wallet whether it's TBA or some other wallet like MetaMask or even Phantom or whatever is the only thing that can withdraw and so for most people that's going to be normies it's going to be a social login or it's going to be TBA and then for early crypto people like me it might actually be a ledger that's secured in a physical location across down and now I can actually put $2 million of my own stable coins or my own bitcoin in momo and can constantly compound and it will give me like nice words of encouragement like way to go keep it up you earned $50 last week compounding and by the way I made sure you were always earning the rest best rate so put it your portfolio on autopilot and then what are we building in the future that I'm really excited about because TBA is still early days you probably noticed that like we don't have the typing indicator like a telegram bot does and a lot of my team actually trade on telegram bots because like some of them are dens like I'm not I don't have time to I'm building stuff. But some of my team actually trade memes on Salana and they use um like Banana Gunbot or you know Bonkbot. If you guys have ever if you were early meme coiners on Salana, you probably use Bonkbot. And those things have been built for 5 years. So they have more sophistication, but like they have a typing indicator when you ask.

We don't have that. So we had to build the I emoji. I love the idea of using a handshake emoji to like confirm a trade for a trading bot. That's awesome because you can just do one tap instead of typing like yes confirmed or something. And so we're trying to think of ways to make it more friendly because frankly the Telegram bots a little bit ahead of us but we're rapidly catching up.

And then what we the idea we had that we're going to launch really soon is nice little P&L statements that people like to reshare those. So if you've ever traded like memes on Pump Fun or Met Meteor or any of the um like I don't know all the I don't really trade memes but I have friends that do. You get these nice P&L statements and people love re-sharing those and those create can make your apps go viral. So, if you're a trading app or you're anything related to money or even just like a cool cozy casual game and you want to share like your house or something because event Sorry, I get really excited. You got to wrap up.

But eventually, long term, we're going to have Momo share your P&L statements and then we're going to have a character studio. This is something coming later in the year. We want to make a cozy casual game kind of like Stardew Valley, but where you can have a Momo universe, but that'll come like next year because it'll take months for us to build and that's why we're hiring. So, if you want to work on like cozy casual games that help people learn good financial habits and help onboard the normies, like the millions of users, not just the few thousand DJ gens like me, come talk to me and I'd love to connect with you and figure out how our agents can work together and how we can build the best apps on TBA. Thanks.

Awesome. Thanks so much, Luke.

Okay, two more demos. Um, yeah, uh, Sam from Bite is going to come up next. Feel free to talk or

Yeah, sure.

Hey, everyone, I'm Sam Achilleshutddle.base base.eth on TBA and Farcaster. I'm co-founder of Bite AI. So, Bite, you chat with Bite and order pizza delivered right to your doorstep.

Pay with USDC right in the chat with a simple approved transaction. So, the idea is making it as simple as possible for users to navigate restaurants near them, place a food order, get it delivered to their doorstep. Uh right now we're partnered with Domino's across all you locations around the US and really looking forward to like more agentic experiences like Shane was talking about like we have the user's order history, we have their food preferences. So doing like automated reachout notifications to the user on the times of days and the days that they typically place orders and just coming with a prepackaged cart for them. So all they got to do is hit approve and get it delivered right to their door.

And then group chats as well. I know I mean Claudia and the team are extremely helpful. It's uh just a really great builder experience. Uh but we've talked about ways to you know enable by group chats where I can send pizza to my friend. I don't need to know their address right there or non we're non friends and it'll just go ahead and pay with my wallet but send the pizza to where my friend is uh right in the group chat and yeah so that's what we're working on with bite and then uh also you know looking to expand to more restaurants.

Uh, Gouff is a service that we're partnering with and looking to expand to them. And then also talking to Yum Brands to enable Bite to place orders at all 58,000 locations worldwide. So I think a big thing we've got is like uh there's a lot of people outside of the US that want to use Bite and so we're looking forward to, you know, enabling it across the world.

Awesome.

I I'm really excited about Bite too because I feel like soon you can just automate your orders entirely. Like if I want to order like a large cheese pizza every Friday, I can just do that and like it'll have like access to my account at some point where it's with spend permissions where it can just execute on its own.

So it's like very hands-off permissions. I know that other places.

Yeah. Very exciting. Thanks so much Sam. Um cool. Last but not least, uh Tim from Brackie.

Hey guys, Tim from uh Brackchie. I'm the uh token uh sports betting bro in the audience. Um, so I'm actually would would like to begin just sort of describing what Brackie does and like what problem uh he's solving generally and then to think about like how we think about um messaging like the messaging experience fitting into what Brackie does and then I'll just talk about like a couple things as we've developed of like kind of special challenges that we've encountered in messaging building our agent for messaging as opposed to like existing in the social feed. Um, so first, Brackchie just uh very simply it's he's like an old school bookie, you know. So before sports betting sites existed, you had to go to like Tony Sopranos henchmen and and you know place your football bet.

You had to find somebody to be the house and place the bet. And so Brackie is that in the sense that he's a personality that you can talk to in your natural language. But the very cool thing is that Brackchie is headless. So you can you can book those sports orders and sports bets and talk about sports with Brackchie anywhere. So we started in Farcaster and base and then went into XMTP messaging.

We just recently launched Brackie on Twitter. And so in any of those spaces, you can say at brackie uh what what's the price for the Cowboys this weekend and you can figure out what the market is and book a bet directly from the feed. So we have privy smart wallets that we load with Brackchie token for people to use. It's like your free bet on DraftKings or whatever. So you can just say yes to an offer that Brackchie makes in the feed and Brackchie will send the will book the bet from your wallet with the free free Brackchie that you have.

So it's it's about reducing friction of participating in sports betting um as much as possible. Um, and and to the point like uh of whether it's good for someone's mom to do it. Um, you know, I I think that sports are a very social experience. So, the thing that I always think about is like when you're bored at a family party and you don't know what to do and you you have one relative who happened to bet on the game and they're screaming at the TV and then like everybody starts to kind of come around as like what's this person screaming at? like why do they care about this game so much?

And then everybody's just into it and seeing what's going to happen. And I think sports are actually super fun as a social experience that we can kind of be inside of together. Um, and so Brackie has allowed us to kind of recreate that fun experience of watching games with people and caring about the games with people inside of these different social feeds. Um, the thing that's exciting about messaging is that obviously like who you're watching the game with is what's making this really different from other sports betting experiences. And so, you know, uh, your mom's not likely to be on Twitter.

Um, but you can you can get your mom into a messaging app. And so, you know, bring bringing your mom into your group with some of your friends and teaching her through brackie about how football's working and teaching her how to make the bet while everybody else is doing it and kind of copying that behavior. Basically, messaging group chats can be social um while also being more intimate and you are obviously choosing who's part of that. So I think it's super exciting space where we expect that Brackie will be discovered by people in social feeds like Twitter. Like why is this guy saying brackie max bet on the Steelers?

Can I do that? You see other people doing it in social feeds but you end up using brackie. Um this is how brackchie is working in the social feed. Um, so yeah, the the um so the the problem being solved is like having fun with other people during sports events and reducing the friction of being able to make your first sports bet. Reduce it as much as possible.

Um, and let you do it with with other people socially. Um, and in messaging, I would say that the biggest challenge is like, okay, we made Brackie for the public social feeds and so he'll respond to all the messages. One of the biggest challenges in building an agent in a group chat is um how how often do you want the agent to be responding in the group chat, right? So imagine that you have a friend in your group chat and that every time someone said something, your friend always wrote a response. It's kind of annoying and it starts to feel like that person is dominating the conversation.

Um, so I think there's a really interesting design space of like I'm always thinking of agents actually um, uh, like my personal opinion is that uh, we don't need to make like a better version of the Delta chat bot, right? It's not needing to be like a better, more superpowered version of that. It needs to be like another friend or peer or even super peer like is in our chats with us. Right? And so they should know when to talk and when not to talk, right?

They should know what kind of what kind of thing would be a fun thing to contribute or what kind of thing would be an annoying thing to contribute. Um, and so I think that that's like a really interesting sort of design space for messaging in particular that's like super important uh to work on and and discover together.

Awesome. Thanks so much, Tim. Yeah, Brackchie will be live in um the base app later this week.

Yes. So try it out

and just so like if you want to play with brackie you can you can already start up a DM with Bchie mention Brackchie in any group chat and what I would recommend you do is that there's a pickle ball tournament today of there's eight teams left there's a Brackchie market for that everybody who mentions Brackie you'll immediately get free Brackie in your wallet that you can start spending instantly and so ask Brackchie to teach you about the teams teach you about who won yesterday, which teams are undervalued, and you can just have that conversation either in a group chat or in DM or in the feed. Get your free back brackie, make the bet, and then come watch the pickle ball, and you can sell if your team is losing and buy another team. So, it's actually a pretty pretty fun like endtoend experience.

Yeah, I like that action item. Let's do it. Cool. Well, that's it. Thanks so much everyone for coming.

Um Yeah.