Transcript: FarCon Rome: State of the Union keynote

all right thanks everybody thanks for being here

now before we start i think thank you limon for organizing this it takes an incredible amount of hard work to like put something like this together so very much appreciate the work you and team have done feels very much like farcon one in many ways when that happened in boston it was a similar small group of people and it's exciting to be at the beginning of something new and i'm glad all of you flew out here or drove or took a train or whatever to be in this room so thanks to all of you as well appreciate you coming in now this is a small room so at any point if you think hey don't not interested in this presentation just raise your hand and be like you know stop what you're doing let's just do questions and that's fine and we'll just move to questions and cut out the rest of the slides so just like raise your hand and and kinda speak up so with that let's get started is this working okay so we'll start with looking back at last year so last year is in the last twelve months and from there we'll keep going into like what the next year might bring so last year over the last twelve months we have had mini apps developer and creator rewards and the forecaster wallet inside the client and all of them have had a variety of different experiments that have worked to some extent some more than others and now we are at this point where we are like hey what are we doing next like why are we here and why why bother right like why come to this and put put this event on and then talk about all of this so i wanna start with that which is like why farcaster and our goal on our team is to think about why are we putting all this energy in and the top two things we consider for this broadcaster community is that we think it's a special place we think the community is optimistic the network is supportive and these kind of places are hard to find and when you do find them and the community is organically building up to this it's actually quite productive and useful to try and carry it forward in a way where you know hopefully this product and this protocol will live on long after we are gone our team is gone and maybe you all are gone too and that's the whole goal of a protocol and that's why we consider it like our job is somewhat as shepherds to like come and move this on to the next juncture where we can kind of bring this community together and move it on to the next stage and that's one of the reasons why we're here and why we think barcaster is worth putting time and effort into and now as we think about that i think it's important to start with what our goals are like why are we doing this and what do we want this to look like i'll go over these goals once now and then if you have questions we can also go over them at the end so the first goal here is to give individuals control over their data this is marked green because this is already doable today like everybody can go and get copy of your data you can build stuff on top of that data that's how companies like started twitter doesn't have for example a different third party provider that provides its developer apis even the ability to fork and to like create another copy of this data exists because the data is open so anybody in this room can go and do what you want with your data and that's available on the protocol now the next goal we have is free speech for every individual and what this means is if you are on this protocol you have the ability to go and say what you want this today is marked yellow because there's more work to do here and i'll get back to this slide in a second but we think this is an important goal for the protocol where if the protocol is going to live long after we are gone then we need to go fulfill this goal and the third goal here is to grow forecaster as a result of doing this and the reason this is marked red is because real impact only comes with real scale like if you had bitcoin today and only like 200 people used it it wouldn't really have decentralized much it would exist in theory and it would work as a proof of concept but the world of finance would still be centralized and that's where forecaster is today where it works or might work in theory but unless we achieve real scale there's no real impact so as we look at the first two goals we have to add this third goal in where if we want to have real impact in the world we need to achieve real scale and these three goals are kind of what drive us to like what are we gonna do to go and achieve success on these so going back to yforecaster and looking at the goals we just looked at our goal here is to improve on the status quo today and when you think about the status quo today that's like the competing products in the social network space in the consumer space overall and we think about improving on the status quo on three axes one on the philosophical level which is permissionless network just spoke to that a little second ago we'll come back to that the second the status quo today is like skeuomorphic to like tech two decades ago it's obviously changed a lot since then so how do we make consumer social more ai native and

we'll get into that so we'll start with the permissionless network spoke to these two things here the next tactical thing on the permissionless network side for us to do is to add new validator operators and we're going to be adding two new validator operators over the next couple of months if you read the latest dev call update or were part of the dev call then you might have heard us talk about it the focus on adding two new validator operators will be in new geographical areas so not only can we add more validator decentralization but also regulation decentralization and after adding those two validator operators our validator pool will be down to 60% today it sits around roughly 80% once we are below 50% we are not the majority validator runners anymore and our goal is to keep driving this number down so once we are post 50% you can go and run a validator but just asking the other validator operators to add you in that's the first thing here the second thing we are looking at is how to improve the status quo on ai native social what does this mean we'll dive into it by the way if you have questions at any point just raise your hand i'm happy to address them as we go along

two things here one when we think about ai native consumer social we think about what is software as a social media format so we have looked at media like multimedia images and video as social media format but what does software look at as social media format the second thing we look at is what ai can do for better human connection not just ai being ai which is what things like multiple are if you're familiar with them but how can we improve human connection with agents in ai so we'll start with software as a social media format just to talk about this for a second what this means is today when you think about software it's not really meant for social media timeline so if you go look and look at what companies like lovable replit bolt claude everybody's making it's for building software that is longer heavier in some ways it's sort of like the netflix or youtube video of software so if you were looking at consumer cameras that came out like two decades ago before consumer cameras you would have to put a lot of effort into making media that's where software is or at least was so far where you'd have to put a lot of effort into making software and then when consumer cameras came about the first thing that came about was youtube where you could take your long form videos and you could go put them online and it was a video storage platform once we had consumer cameras on smartphones that's when people started making bite sized snackable consumer centric media that's where we had instagram snapchat tiktok software today hasn't had that metamorphosis we have had the consumer grade software tool with ai but software today is still built in long form video format and it's not meant for social media timeline and so how do we think about software for social media timeline and what does it mean to have snackable ephemeral bite sized software we started with mini apps two mini apps here dota az flynn really good work fotocaster which is more recent i by the way just picked two there's a lot more mini apps here so if your mini app is not on the screen don't get offended

and mini apps today are still somewhere in between like the youtube and snapchat story so if you're going and looking at software centric media formats it still takes a little bit of time to go and build like deep experiences that people wanna come back in and that's that's good like we want the full spectrum of experiences but that does mean that we have a fair bit of the spectrum to go fill it where you can come in and both create and consume software quicker our next iteration was with snaps so this is sayangel's snap on snap magoce this wouldn't exist without ai driven software development which was snackable software where you come in on your timeline you can tap a few buttons and you can play a game and this is not meant to be a company this is not meant to be venture funded this is just like snackable bite sized come and play forget move on do the next thing and takes anywhere between a minute to couple minutes to make this one i saw yesterday i think somebody made it during the snapathon thank you and again if your snap is not on here sorry i just had to pick one from the timeline this by the way i think is a great example of what social media timeline software looks like it's timely it's meta driven so we have the us iran war going on the strait of hormuz it's not going to be relevant ten years from now because people won't really remember what happened here and it will live on the timeline and move on just like a tweet or a cast does and hopefully like in the next few months won't even take the amount of time it took nikolai today and that's another example of where software can be built for social media timelines and one of the things that we are exploring so this is an example this by the way should be out to everybody now or rolling out to everybody now live where in your cast composer you can go and build these today

thank you the thing to note here is what you find today is the worst it'll ever be like it can't be worse than this and so we are only gonna get better from here and today it's possible to make these in less than a minute like if you have a simple thing just go and then you can make it in less than a minute that's how much time it takes to like shoot a snapchat story and that's what we are looking at when we think about software on a social media timeline yes

no just update your app

there's a snap icon at the bottom yeah yeah

yeah cool any other questions

okay there will there will be time for questions in the end so i'm just gonna keep keep moving on for now the second thing for ai native social is how do we think about ai making human connection better so for any of you who have been familiar with things like multbook that's where agents talk to agents the best agents there are humans masquerading as agents agents today are not really entertaining unless you put a lot of time into them which we know because grain has put like a year's worth of time into making neymar agent entertaining and there's a long way to go still so if you just introduce agents on a network it doesn't really add that much what we really need to do with agents we think is make human to human connection better and that's something we are looking to explore as part of our journey into making ai native consumer social so this is a simple screen where agents kind of become part of your app experience from onboarding i don't know if the text is readable but essentially what's happening on the screen is as you onboard onto the app you tell the agent what you're interested in and the agent makes it a goal for it to ensure that you become integrated with the right parts of the community so here the user just tells the agent hey i am a designer i'm gonna follow the design channel i'm in new york and the agent's like okay i'm gonna try and get you integrated into the new york design community that's it and you start and i'm gonna run through a few examples where we'll see how this happens so the first one super simple they'll get slightly more complicated along the spectrum super simple designer in new york you have the forecaster app and you might have casted something about design like five weeks ago six weeks ago somebody else comes in and asks a question somewhere else on the protocol you don't know them you haven't followed them but they are a designer and the agent figures out that hey you actually talked about this thing five weeks ago do you wanna go respond to this person and tell them what you said because they are real and you're real and you guys should talk and

so real real human to human connections being made second one you've all seen the follow lists where you join and you see a bunch of people you can follow it's not really anything special about varcaster every social media app does this and some people use that list some people don't results vary we think it's actually more useful to tell you who to go follow on the network depending on what you're doing and what they're doing so again on the forecaster app you've been doing what you're doing and at some point your agent comes back and says hey you know there's not much happening on your feed right now but two people keep showing up for me based on what people you know are talking about what you are talking about and they seem kinda close to what you've been doing on this network and what you've said what your interests are they are real people you should go talk to them super simple can be done today but not really talked about again improving human to human connection third one now we are getting slightly more complicated which is you like this event by the way this was inspired the next two by what people are doing with farcon and forecaster fridays which is irl hangouts and we think there's gonna be a lot more demand for these kind of irl hangouts as things become more digital native first so a lot happening in the digital space and people also want physical space so this example here the agent says hey you only use your follow feed you don't use the home feed so you're going to miss some algocasts which is okay but i spotted that somebody's doing dinner in new york involving others in the design channel do you want to go attend that dinner just wanted to make sure you don't miss it and this is a very specific thing without having to go through the rest of algo feed and just focused on hey do you want to go and attend this event i know there's a few spots left i know the organizer is still looking to add people and given the place they're having the dinner at i know it'll cost you about $25 now if this is up your alley i'll go tell them that you can attend the other side of this is you can actually host meaning if you're a designer in new york and you've been trying to do something and maybe you've said in passing like you know i'd love to host something like this at some point the biggest barrier to hosting is the work needed which limon can tell us about like insane amounts of work and you wanna go and host a gathering with like a few people and here inside the app it can tell you hey based on network activity i think these people will attend you should reach out to them first so that you have your core group host it here it's under $30 worth of budget you can have a dinner once you have your core group of four to five go and spread the word around and maybe other people will join just let me know and i'll make the reservation and for you it's a one tap like okay let's go and try to do this thing irl and now we have gone all the way from like joining the network knowing who to follow knowing who to respond to to maybe attending an event to all the way to maybe hosting an event now we'll see how far on the spectrum

human connection by embedding agents into the network and into the application from the very get go and that was ai for better human

yeah great question i think all experiments should in most cases start with mini apps like we might start this experiment with a mini app because building into the client is much more expensive than building into a mini app and so to go and test out our use case we might build this into a mini app see how it does and if it does well we'll turn our own mini app into the client and i like we are just getting started with what agents can do and these are ideas we thought of in like chatting in our team for like a day and so as agents like you know keep doing better and more people think about it i think the number of ideas today is like infinite and it might not be exactly infinite but it's definitely more than four so there's a lot lot of stuff that can be done here and i think anybody starting with the mini app today like maybe an agent plus a mini app today should probably be in good shape

cool the last thing on how to improve status quo is internet native finance and maybe this room is more familiar with internet native finance crypto digital rails payments but i'll quickly talk about what we think about when we look at internet native finance so the first thing we think about is the simplest payments like if you are on a global distributed network you should be able to purchase and sell to anybody anywhere in the world on day one and this is obviously already possible today but we think it's an important part of building a social network today like financial rails are gonna be increasingly important both for humans and agents and today the rails exist they can be improved they can be made better for agents and that's our first thing for internet native consumer finance the second is access to capital meaning if you're now we are in the like all the people in this room are probably in the 1% of power users who are not only users of the network but maybe builders on the network so when you look at a larger network it's like one nine ninety distribution where 90% of people are lurkers 9% of people create good content and then one percent of people are actually building something that makes them money on the network and this room is probably more on the 1% side and so if you're on the 1% side not only do you need the network to function well where your users are but you also need access to capital in some way that allows you to go build what you want and access to capital looks interesting in crypto where today there's a few different ways to do clanker today allows launching tokens that gives you access to capital based on trading revenue we think there's more to do there than just the meme coin launchpad today especially with regulation in the us market that's about to pass in the future there are different ways of introducing different commodity or utility tokens that allow you to do more than just a meme coin this is a longer road we think it'll take like anywhere between eighteen to twenty four months where the regulation passes and you can do interesting things but it is something that our team is focused on the exact shape will depend on what kind of laws pass but there might be ways to give revenue back to your token holders there might be ways to pass yield on if you're a larger team you can actually launch a token that is useful to your business and to your token holders and those are things we'll think about as laws come to pass

as part of that one other thing to note like these are things we are thinking about as we go on but simpler things we'll continue to introduce as we go simpler things like many people asked for hey can we have a way to connect our forecaster wallet to other web apps like opensea or uniswap so this is live now if you're on the latest version of the app you can go connect your forecaster wallet to any link that supports a web wallet connection and use your forecaster wallet to go and interact on chain it doesn't have to be just mini apps or payment sends anymore the second thing we are gonna do here is we are gonna introduce a second wallet inside your wallet and that'll be a private wallet which is not tied to your social profile and it also won't connect to many apps by default so it'll be more secure and that'll allow you to go then transact on chain in ways that you don't always want to be public about and both of those will be available shortly

cool so going back to where we started we are gonna innovate along three axes on the philosophical side thinking about a permissionless less network on the social side thinking about an ai native social network and on the financial side introducing better internet native financial tools and we are just getting started so we have new formats which are rolling out now and new ways to create them we'll have more useful agents and we'll have monetizable products built on top and i think that's it so questions

yeah so i think from a architecture standpoint of the protocol itself forecaster is slightly different to either blue sky or mastodon so mastodon is more like decentralized discord servers where the discord servers are centralized but the network of servers is decentralized so if you're if you're part of one server and the server host takes down the server you still lose everything that was part of that server forecaster is global meaning the in network or protocol state is stored on every node so there is no risk of one node going down and your data going down with it blue sky is a little bit more like email where you have to host your own email server or be part of somebody else's email server so if you're using gmail you're part of google's email server or you can host one on your own computer so most bluesky users host their bluesky data on bluesky's email server there's nothing wrong with that approach it's just different forecaster again is global so any node has all the network data you're not on anybody's server your seed phrase is yours and it's self custody your accounts are on chain and that allows you a slightly different shape of decentralization than the other two allow i think we'll see what exactly works in the market and there are slightly different approaches to product building like blue sky obviously took a more political leaning approach to go get users which worked out to a reasonable extent and then we we are taking a slightly different approach to that it remains to be seen what actually works but we are pretty excited about actually building things that are useful for other people instead of looking at twitter cloning it and making it a different political spectrum

yeah mostly a lot of work stability and resiliency of the network today so if you look at the forecaster updates channel we recently posted all the work we did on snapchained to make it more reliable and snapchained today is relatively brittle when you think about how different validators operate with with each other so even though we only have five or six validators right now we have to every time something moves something simple we have to do a lot of work to ensure snapchain reliability and we are actually improving that reliability like right now like we did something over the last two weeks and two weeks of like continuous work to like get it to the next level where we can now add two more operators snapchain block times are not conducive to geographically distributed operators right now where if you are in a very faraway zones you'll actually be handicapped by how quickly the blocks go and and manan should be speaking about this better i'm kind of speaking on in terms of what i know but we have to think about how block times can work better before we can geographically distribute and that's the next work we have taken on right now and we'll just keep moving along the spectrum where we can be at a point where anybody anywhere can run a node and we expect to get there but if we just did that today it would just crash like the whole network would crash we'd have to go figure that out as part of figuring that out we'd be back to where we are now just after a lot of other hassle and so we think it's more prudent to go and fix this stuff and add as we go and then at some point we'll be in full days full decentralization mode

by doing all the things that we talked about here which is focusing on the philosophy to start permissionless network focusing on the social aspects which is ai native consumer social and the financial aspects which is ai native consumer finance i did have a slide here because i knew somebody would ask a question like this which is the way things grow is you have to build something good and you have to build something useful and they both have to be true at the same time and the overlap between them is very small especially for consumer products the bar is very high and the intersection of those two circles is increasingly narrow and if you just build something good like we could build a good protocol that works in theory but it's not useful doesn't matter and if you build something useful but it's actually not good to use also doesn't matter so we have to do both and it's a very hard problem and so the high level answer to your question as to how to grow the network

over there thanks farish like thanks for inspiring and give us some kind of things to look look forward so i i wanna double down on this kind of comment that that you made so i think like one aspect is kind of making it decentralized but today like startups and businesses are shut down not because of government but because they are ran out of money and in order to stay alive you need to be good and useful and make money and i kind of simplified either you create something that is so useful for me to pay for or you're helping me make money as a builder so i would happily pay for that so and in the state today there are great people in the network but how can i either build a business or or do something that i'm willing to pay to to to to use to get to this point that we all kind of build something that grows and serves us all

yeah the network or forecaster generally the product the network the app any layer you talk about the best it can do is provide the right tools and it can give you the components you need to go make money but the truth is nobody can make money for you like that's just not a thing in this world you have to make your own money and the tools we can provide so go all the way from like simple monetary tools like launch a token to in the future launch better tokens to trade to payments all the way to grants so we just set up clanker ecosystem fund where we'll actually be giving out financial grants to people building on the protocol and the network and the grants will sustain early teams but if your team size is too much the grant might not be enough that's a different problem you have to solve in a different way but access to capital is a full spectrum of options and today where the product is we can fulfill a certain part of that spectrum and in the grand scheme of making money or making revenue everybody has to make their own money so we can give you the access to the best tooling and we very much want feedback on how we can improve that tooling so that your life is easier and then every founder or builder for themselves has to solve distribution and revenue and we'll work with you but that's kind of what the overall package looks like

useful the people who use it if there's enough demand it's very easy to tell when there's good demand you look at a chart and instead of going like this it goes like that and we have public charts today they are on dune you can go and look at broadcaster protocol activity and you know sometimes there are spikes because there's internet volatility people will come in and try to farm something etcetera etcetera but if you try to look at what long term growth looks like for any company or product that works it's the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the charts

one's discovery

great discovery is largely broken into two components proactive and reactive which is what you were talking about like hey how do i find something that is already happening and i can just see it and then how do i ask for something that i'm thinking about that i can then go find so proactive discovery is what is solved with feeds generally and we have made a bunch of improvements to our feed at this time where today if you're on the home feed instead of the following feed you can actually go and find casts that are relevant for you recommended by our system and it is based on personalization on what you have done or said on the network and again this is the worst relevance is gonna be it'll only get better from here the second thing we have done is we have also added different ways for you to see content that you might have seen but has changed so if somebody's having a good conversation the conversation comes back on the timeline because the substance of the conversation has changed we have also on the other project discovery side if you're looking for like you mentioned nfts different kinds of hackathons we have a different feed on the app page that's just for projects i don't think many people use that feed today because looking just for projects is not actually that interesting but if you're specifically looking for projects there is a place to do that and finally on reactive you can ask the the nanr agent today to go and find stuff it won't find things all the time in which case if it fails on something let us know but it does find stuff fairly reliably and so once we build the network agent into the app you can instead of always asking on the timeline you'll be able to ask within the app you can do this today with the nanor mini app it's one added step of taking the mini app goes back to my experiment with mini apps first point but i think we are at a point with the neymar mini app where we can slowly start building that back into the full application and you can go and ask the application for things you want to

i think the success here looks not like twitter which is where we are starting to think about what different native software formats look like and what interaction models that are different from just scrolling a feed look like exactly what that shapes up to be we'll have to find based on experiments and what people do and they tell us and they give us feedback and we look at the data it is you're right a twitter clone today and it's obviously not been successful at moving people off of twitter in a forecaster and so we don't think that's what success looks like in the future

interesting our models are based more on what we expect the use case to be versus what the ui looks like because predicting ui isn't actually super useful all the time predicting use cases does actually get you to the ui based on whether your use case experiment works and so our use cases that we imagine is people creating new kinds of software first content and consuming new kinds of software first content and like i said the snaps are the first experiment for that it might not be exactly what it shapes up to be and it might look completely different in a year from now but that's the use case same for like ai native like how to help human to human connections we think that's the use case where we wanna help humans connect better with each other what the exact ui looks like i think we'd have to find out like the default ui today would be like the chat style interface and every ai product is in that interface whether that is the interface that sticks five years from now i think it's very hard to predict

how we solve to incentivize artists who are not looking to be commercial

maybe a question back for you is what are the blockers experimenting

yeah i mean the closest we have to that is the clancor ecosystem fund and our company is not positioned to give out capital like a venture capital firm like we are not a venture capital studio and we are also not a i don't know large endowment and so the only capital we can give out is that comes in through our revenues that we can pass on which is what we are planning to do with clinker any capital outside of that is the same as the rest of the world you have to go figure it out and we had to go figure ours out and that's true for everybody else so i don't think it's actually honest to go and make promises about how we are gonna get you capital outside of the things we already have and that's just the reality of where we are

your thoughts please the answer to this is actually very simple which is the protocol sensors nothing today and the client sensors things that shouldn't be on the client so people here don't know but there's a lot of bad material on the protocol like we had to censor csam we had to censor malicious mini apps malicious like accounts that try to like scam people out of their seed phrases and if you go to some of these malicious accounts you'll actually just see on the client that says this account is nerfed on the client but the protocol still has the data and it has everything that they have shared and you could theoretically have a different client that surfaces these kinds of information but obviously the us laws don't allow that and we don't agree with that and based on our philosophy internally to the client and based on the laws of the country we live in or the app is operated in the client will have its own rules but the protocol is open and you can go and find stuff on the protocol so when i say free speech i mean at the protocol level the client level does have its own like app store rules us laws our own team philosophies and all of those levels are applied

thank you question do you have internal goals for the daily active users on the platform for like this year or maybe like two years three years i don't know something like that

yeah i think we don't set like the way to think about goals is if you're a large company you go and set a goal and now your goal for the quarter is to go hit that goal and that would be slightly disservice or a slight disservice to where we are both as a team and a protocol and a product today like we can't set quarterly goals that's just too long a time frame we have to be able to innovate daily and kind of move forward the ball every day and things might change very quickly so we don't usually set quarterly or yearly goals but from a larger high level standpoint yes i think the protocol and the client needs to go 100x and that's what we are focused on when we think about if you go back to the goal slide the third bullet grow fore caster as we do the first two that means 100 x ing from where we are today and whether that happens in like a month i mean i don't think it's going to happen in a month but whether that happens in a month or a year we don't really set goals based on that because we want us to be able to like change on a daily basis depending on how things look but on a high level yeah i think anything less than we are at what like couple thousand maybe on an app level like $2,030,000 dau 20,000 dau the protocol is lower because some people come in and just lurk and so you don't see that on the dune dashboard but if you can't go from 20,000 dau to 2,000,000 dau there's not really much happening and so that has to be at least the first outcome when we think about growth

do you know do you know from this 20 k how many are like builders versus people coming to use farm and rewards rewards like the characteristics of the users

i think the one nine ninety applies where 1% is power users 9% or extreme power users let's say nine percent is somewhat power users and the 90% is a mix of lurkers and maybe some spam accounts so so on and so forth

yeah great question i think if we are still here ten years from now the demographics will have expanded from where we are today just by definition of growth but talking about just like where we are right now you're right that a lot of social media skews younger first but it's not true for all kinds of social interactions so like reddit and twitter are good counterexamples where twitter didn't actually start young reddit didn't start young either and today now we have teens on both but it's taken a long time to become popular among the younger people who started media first on instagram facebook snapchat tiktok tok and so i think there's enough variation on the spectrum where you can build products for the right demographics depending on what use case you're looking to solve the use cases we demonstrated today are unlikely to be teenager use cases they're more likely to be young professional use cases which is largely the room we have here so i think that still lines up with where we are going and then if we are here ten years from now we will have expanded the use cases and the opportunities where we might be in a broader demographic

one more so you you were talking about how farcaster is more about bringing people into like real life back into real life organizing like in real life meetings all the like mainstream platforms they're all about opposite sleep less stay online don't go anywhere scroll click advertising so and unfortunately these things works very well to enslave the people's mind so how you wanna fight that thank you

so

there's so i'll say two things one the product isn't focused on just doing irl hangouts those are two examples i had the product is focused on increasing depth of human to human connection now that could be digital or irl we think there are some good irl use cases and to that point of why we think this is a or at least an interesting approach to take is exactly because what you're saying i think there's a ceiling of demand for how much time a ceiling of demand

and we hypothesize that there will be a surge of demand for things that people can do with other people instead of just with their computers or phones and that demand should

those of two things to happen in parallel and because all the other networks are focused on optimizing digital's time we are focused on optimizing or not just digital time but like time with agents let's say like half the tiktok videos today by the way are agent generated they're not real people and that's the kind of thinking where we are coming in from a slightly different direction and say hey instead of you looking at just agent generated content can we connect you with real people and then we'll see how that works

yeah that's right and i think some of the work like even the snapcomposer work you look at that's free tokens you can go build software without paying for anything and those kind of things are not usually looked at thank you emma as things we are doing to like bring costs down but that is a free agent and you can go do things with it and if the tools we can keep increasing to gm forecaster's point earlier so if you want better tools or different kinds of tools that we can come do and that's kind of the space we're in

we actually

we we do need more dogs we actually have done something for the dogs which is if you're interested in dogs cats or pets and you have been liking or following or recasting those casts your personalization is better and you'll see more dogs cats or pets

yep

there you go there you

great question i'm also gonna sit oh sit sit sit sit sit sit so business model wise you're right that most social media companies today focus on a screen time business model we are actually fortunate in the sense that we don't have to focus on just a screen time business model like clanker's business model has nothing to do with screen time and it makes a bunch of revenue simply because people use the product that was built and same on the forecaster side when we think about increasing depth of connection thank you

thanks lamont because we are focused on the depth of connection there's actually different ways you can make money as a product so for the irl events if our agent for example is helping you organize the event and making transactions there's ways you can charge as part of the transaction there's ways you can charge as part of the concierge that is the agent itself there's just like as new product categories open up different kinds of business model and revenue lines open up and we think both internet native finance and consumer like ai native consumer has just different kinds of business models like ai today just has a different business model it's not focused on screen time it's focused on subscription revenue and you could argue that hey the way you use ai is on screens but you could also use ai through voice and it's the usage of the tokens that generates the revenue and so as social products move into these new kinds of tooling we'll just have new kinds of business models emerge out

like the difference between how he and i cast

oh i see

i i mean to answer that well i would have to know the deepest philosophies that they had as a team and i can only go off of information that everybody in this room has and on surface level there's actually reasonable similarities in the sense that we want to build a product that

i think we might be thinking about it differently in terms of i think the way dan and team approached it one they were pre ai so they just didn't have a lot of the ideas that we have now simply by being in weird to say a different time but that was just a couple years ago

and i think there was maybe a difference in philosophy in terms of they attracted a very builder centric network and did a lot of hard work for that like dan hand onboarded a lot of early people including me on one on one one calls

yeah it does but then the last turn of products they built were maybe for a different audience than the builder centric network they onboarded and we are unlikely to do that we are likely to more build for people that we are like manan and i like to build products for ourselves and we generally have this philosophy at our company that if we can use it we can ship it to others and test we are unlikely to build something that we can't use now i say this is a difference but i think this is also a similarity because they were trying to ship something that they couldn't use and then they didn't right and that's why the protocol changed hands and so i think it's actually very hard to ship something you don't use and it's possible that the kind of products we use never achieve the kind of scale that is needed but we are unlikely to ship things we don't use so we'll just find out whether we can achieve that scale or not

right that's right we did hire an in tori

that dan did not hire

yeah i think i mean for building a good frame took like a week and building a good snap hopefully takes minutes so but again this is just like different times

yeah

yeah i agree that's good feedback i think we can change the terminology somewhat everything on the protocol is public what we mean by control your data is that we don't single handedly control your data you do and it's public and you can go share it or by by phone yeah yeah that's right i think it's reasonable or at least somewhat reasonable to assume that the protocol will expand to account for private data so for example dcs are not on the protocol today and the previous team never got around to putting them on the protocol we might end up doing that i don't know like we don't have timelines we haven't talked about it but it seems reasonable to do we might have other kinds of private data in terms of private channels or private lists etcetera that could end up living on the protocol and then the term control would actually encompass controlling your data the way you were talking about but you're right that that's not the case today

so in my talking to a lot of people here one thing that i hear often is that people really love channels what's gonna happen with channels

thanks grin for people who don't know by the way grin started the original far con and is a reason why we are here today so

channels so we just haven't spent enough time thinking about channels is where we are so we haven't thought about you know a negative take or maybe negative depending on who you are but some would say hey let's just you know not deal with them let's let's remove them they're centralized anyway let's just deal with what's on the protocol and some might say hey you know we love channels to grint's point let's do more with them and we just haven't had the time to go and think about either of those approaches and so they are as they were we will hopefully get some time in the near future to like think about them and work through them the one change we did make was it required 2,500 warps or something to go make a channel it's usdc now so you don't need warps but outside of that we have largely left them alone tbd exactly when we can go revisit them we have a long list of things we would like to go and change and fix channels being one of them forecaster pro being another of them it's just a limitation of time and priorities but yeah at some point and you should tell us what to do and we can add it to your linear board

yeah absolutely like one of the top things we have thought about if we are like when if when we go back to channels is ai moderation like one of the things that didn't work last time were all the moderation aspects and that required a bunch of experimentation and people kinda got confused about what to do and there's probably ways to just like automate it for most of it and humans can just come in for the last 5% to 10% we haven't gotten around to that yet but that is like if we were going to do something with channels that would be like top of our list i didn't actually voice this over but one of the screenshots did talk about introducing the person to a channel and the the channel there would be the design channel we think it's a little too early to do that before we do the moderation part because we'll start start introducing people to channels and the channels will be like full of spam but once we do the moderation part i think picking what channel to cast in for example that's a simple one introducing people to channels based on what you're interested in having you follow people who are active in those channels all of those things can be done again just have to get to it i didn't hear anything

that's a great idea i think we should think more about that for the safety side you're right that we have more guardrails today than on the onboarding like one of the things that i just talked about we'll introduce our private wallets where people on network won't actually know that there's a second wallet with that address that you can use more safely but yeah i think we could do more with that on onboarding and we can tell people the kind of things that they'll get introduced to and because this is a more intelligent system we don't actually have to do it at onboarding time necessarily because nobody wants to sit through a 20 step onboarding but we can do it at the right time so the first time you get a spam token we can tell you like hey just so you know this is what's happening you might see some of this happen here's how to report it and then start hiding it in the future the first time they go to the wallet tab or try to use the wallet we can tell them hey saw you went to the wallet tab here's how to use it here's what this means so yeah that's a great idea thank you

also limon feel free to kick me out whenever it's the right time to kick me out kick me out okay

i'm not sure i fully understood the question like what do you mean by 10 x increase in seat call to actions

yeah great question i'll tell you some other times the the same question has been asked the first time the home tv was introduced where people were like hey i don't have to go to the town hall to see movies anymore they're in my home and i can see anything anytime how do i control the next time or even before when the newspaper was introduced in the printing press and people were like hey now i can read anything i want anytime how do i control instead of going to like the village priest to hear the stories and this happens every single time like information can only go up and systems to control them just rise in parallel and as human beings we kind of have an infinite capacity to go and find the right thing and you just have to look back at history to see how we have done this over time and every single time people are worried about there's too much information and that's gonna kill us and it hasn't happened yet and so i kind of find it hard to believe that this time it's gonna do something where information is gonna be the problem i think people will create what they create just like you know before youtube you had to watch tv but now you have streamable video and people figured it out and i think the same thing will happen where people will have both creation systems and moderation slash personalization systems and both will come in parallel and your system today your information intake is based on youtube instagram twitter newspapers movies netflix and all of those happen at the same time you still have the same twenty four hours that like people five hundred years ago did and i think the same system applies five hundred years into the future

useful yeah everybody's agent will be able to figure out what they want

yeah so let me answer a higher level question which is how can we use instant global payments at both the client and protocol level and payments that can be used by both humans and agents so x four zero two is a specific protocol by base that can be used by agents but the larger level question that we can think about is how can we use agent driven payments and it can be x four zero two it can be tempo mpp it can be some other protocol that comes out in the future and the way to think about agent driven payments is agents have to want to do something on the network that they would like to pay for and at the client level that means maybe controlling your forecaster wallet where you set an allowance and say hey my agent you have the allowance to spend $20 a week or $20 a month or whatever your allowance is and you can make your independent decisions here's the kind of things i am interested in and now you can go decide and then you can tune your agent to see like did it make the right decisions decrease allowance increase personalization simple things like if you're an artist and you like specific kinds of art maybe it'll go do some open mitts for you or if you are interested in irl hangouts and there's something happening maybe it'll go purchase a ticket for you and these are at the client like wallet level you have an allowance set and your agent can go spend and that can happen through protocols like x404 too at the protocol layer agents will wanna spend things that the protocol requires them to spend on so today the protocol requires you to spend on creating a new account buying new storage and buying the farcaster pro subscription that's pretty much it and agents do that today by the way a a lot of spam comes because people pay on the protocol directly because paying on the client is too slow and we we we figure out that they're spam and we don't onboard them so they go and pay the smart contract and then they write to the protocol and that is an example of agents using x402 but just or not x402 but using payments but just not in the ways we'd want that are useful over time you can imagine your agent buying new storage when you need your agent buying new subscriptions when you need we'll probably introduce more paid systems on the protocol as the protocol expands and then there will be more ways to pay on the protocol directly and they kind of go hand in hand what lives lives on the client and the protocol your agent has a wallet it can go pay on either end yeah

have more questions come find me