So uh this is gonna be a bit of a discussion starting with with us three, and I'll introduce myself and then allow the others to introduce themselves as well. Um, but feel free to put your hand up at any point and participate. I hope there's this that the general topic is which flows on really well from from Natalie's talk, so thank you. Um that really set us up well. I was kind of looking back at Nate and our previous discussions and a lot of what we've talked about really resonated there. Um, but overall we you know bring it to AT Protocol specifically and and building a new sort of creator economy here.
So from my perspective, um, I'm one of the co-founders of Spark. Spark is focused on content creators, the creator economy specifically, more like short form creators and that kind of really really big influencer marketing ecosystem that's grown on um these pump and structural platforms. And you know, uh I I work we work out of a space in in New York called Versey, which is home to a lot of um content creators is I think they have a combined over 10 million followers um and these are predominantly spread across TikTok and Instagram. And a lot of them, a lot of the discussions and workshops that they run there is basically all about how they can extract their audiences from these closed platforms because you know they're at the whim of an algorithm change or a policy change and and their career could potentially end tomorrow in their livelihood.
Um and AT Protocol we see is a great solution for that. You know, their current solutions are pulling that out to things like Substack, to newsletters, Patreon, Payworld Gardens, but App Protocol is a way for them to natively own that audience on the services where they actually get discovered. And so that's a bit about Spark and my motivation for building uh in the Atmosphere. Um I'm Joe Bassa, by the way. Thank you, Joe. Uh my name is Josh Akadem, also New York-based. Uh I'm one of the bad guys in the room like Natalie. I'm I'm a venture investor.
I work at a firm called uh Sentinel Global. We're a multi-stage investment firm. I've spent my whole career translating open protocols to real-world problems and commercial opportunities. So this started in the fintech and digital asset space where I've done a lot of work um investing in in companies from you know the seed stage to even underwriting public ones. I've done a lot of work in the open social web, uh serving communities from Naster to Firecaster, Activity Pub, came across atproto and Bluesky a couple years ago and is really drawn to the affordance of data and identity portability that the ecosystem offered.
And super excited to be here today and talk about some of the new novel business models that this uh protocol and and technology foundation enables. So pass it over to Nate. Hello, uh I'm Nate. Um I work on open source for most of my life and recently I've become really enamored with App Proto, so working on like some SDKs uh for App Proto, and then recently a music app which has sort of brought me into this problem of like what ways can artists like express um how their content should be used, as Joe mentioned, like how do you retain ownership of the content that you create without being subject to random things.
Um so I yeah, I just think it's really exciting to sort of break open the problem that you get when you reverse the Faust even bargain that sort of she was referring to in the previous talk. So I'm excited to get into it. Awesome. Yeah. Um I guess everyone the reason why everyone's here, everyone understands the the issues with current social and and the kind of power dynamics that are there, especially with content creators. Um even looking at you know the creators I'm we're looking at catering to the most of their revenue sources that are not actually through the platforms they operate on.
You know, the the TikTok, Instagram revenue sharing programs is actually a very, very small fraction of their revenue source, and it tends to be these kind of third-party brands, chips and things like that that make up you know the the large portion. Um yeah, Josh, you you posted a really good leaflet recently about kind of redistribution, the redistribution of attention um and uh especially with algorithms and feeds. Um so yeah, I'd love to if you could share your thoughts more on like how that can impact the creator economy specifically in the context of that proto. Sure, we'll do a quick survey.
How many people posts on social media? Put your hand up. Alright, everyone, hands down. Who makes meaningful income from their posts, hands up? I see one hand in the room. Congrats. The reality is users and publishers are the stakeholders in these social networks that create value, but they're not the ones that capture the value. And this is because the platform at the end of the day is the entity that owns distribution. They own what they're the ones who control what content users see and in what order. And they also control the interface. So the screen real estate that users see.
And these are really the two things that enable the platform to capture a meaningful majority of value on these platforms. And as the entity that captures the meaningful amount of value on these networks, the platform is also the arbiter of value. It is the entity that decides how that value gets allocated across the various stakeholders in the social in its social network ecosystem. And ultimately these platforms serve not users, they don't serve publishers, they serve shareholders at the end of the day. And so there's really little to structurally no incentive for these platforms to share back revenue to creators.
There are some of these creator monetization programs that exist where maybe a small fraction of advertising revenue is passed back to users and publishers, but they're prohibitively high thresholds to participating in these programs. And so one of the core affordances that I mentioned of atproto in my intro was this decoupling of data from application presentation. Everyone's data lives in their own personal PDS. And applications tap into each user's respective PDS and create a an application view by tapping into this data. And so one of the one of the things that happens as a private byproduct of this affordance is users in App Proto are able to create their own custom feeds and algorithms that live outside of the platform's control.
And the algorithm at the end of the day is the thing that decides distribution, that decides who sees what content and in what order. And so I'm hopeful that in this evolution of social platforms adopting atproto and decoupling data from platforms themselves, we'll actually be able to live in a world of user-controlled algorithms versus platform-controlled algorithms. And that gives a lot more agency to the user and the creator, as well, number one, they get to take back control of distribution, which enables them to control what their audience sees. Today, if you're a user or publisher and you're broadcasting out into the social uh abyss, you have no control over what content your audience sees and in what order.
And so this is a really bad sort of trade-off for users and publishers. But if if publishers are able to take back control of distribution through public feeds or through custom feeds and custom algorithms, not only do they take back distribution, but they also take back control of value. And now they become the arbiters of value in this ecosystem, not the platforms. And I'm hoping that with publishers and content creators uh becoming the arbiters of value in these ecosystems, now they're the ones that decide how that value gets allocated across the ecosystem. And I'm excited to see how this envelope of value could get distributed across various stakeholders at the very extreme.
You could imagine like Andra, for example, who runs the largest custom feeder algorithm on Bluesky, a verified news feed that she built with Grays, she could embed advertising and sponsored posts within her feed, which she's already done, not at massive scale, and then redistribute all of that ad revenue that she generates to users or subscribers of her feed. And you could start to imagine how this envelope of value could get distributed across these various stakeholders that actually provide value to the feeder algorithm that creates these new incentives and new forms of of economic coordination. Um I'll stop there.
Uh that's really great. Um I think about like the people on the stream can hear. So when you you're talking about the the like creation of uh or redistribution of value to the creators as a platform creator, as a venture capitalist, presumably this is actively against your core interests, right? Because in order for you to gain value from the thing you're creating, you need to control what people see because that gives your platform value that then you you can extract for yourself, right? So how are you reconciling that? Where like are you planning on somehow building record deals with creators where they pay you for access to a platform and if they are the ones who are the ones creating value, how does the platform or the venture capitalist add anything to that that makes it worth getting the share of the value in return.
Great, thank you for the question. It's it's obviously tough to be the bad guy. Uh no, no, no. Look, I think the competitive vectors are gonna change in this space and and the points of demand aggregation will change as well. Um obviously in this world the creator has a lot more leverage as someone who controls their own distribution uh owns their their portable feed that they can take with them from one platform to another. I envision, again, it's it's still really early. Uh still it's still too early to say, but I envision, you know, companies like Graze, for example, will build will almost become like the stripe for creators and build monetization tools for them that allow them to, you know, for example Graze has already built an an ad marketplace that allows for uh feed creators on Graz to embed advertising and sponsored ads within their feeds.
Um and you know, Grays will probably impose some sort of take rate on on that transaction. And so yeah, as more uh as more people use Grays, as more feed creators come onto Graze, there's more attention concentrates and and Graze feeds, um, yeah, I I expect Grays to yeah potentially in the future uh you know participate in some of those economics where they're introducing advertisers to attention that takes place within these feed creator um within these algorithms or feeds that that creators uh produce. And I think the platforms will also play a role in this. Ultimately, if you create a feed or an algorithm, you still need distribution for your fe feeder algorithm.
If you just create it uh and put it out into the world, you you there's not much value in that. You need people to to view it. And you know, Bluesky has 40 million users, they provide distribution for feeds. They'll probably want a piece of the cut as well. Whenever they render an ad uh within one of these feeds, you know, whenever a user looks at one of these feeds in the Bluesky client and sees an ad, uh you know that that envelope of value will get split between the creator, the um and the platform, uh, and probably Grays as well, who's like this monetization tool or infrastructure for for creators.
So um yeah, it's my my response. Yeah, I mean I'll I'll just make a quick point and then we'll go to the question over there um from the perspective of someone that's you know building an app there. I think I echo Josh a lot there in that there's still the the distribution surface and I think it it does flip the power dynamics, but there's still the requirement for the interface to to get the creator's content to the audience and there's costs involved with that. But I think that the competitive nature changes and the power dynamic changes.
So you know the the creator in this case has a lot more power and say and arguably more the mark much more bargaining power in that transaction. So you know the the apps that exist, depending on what roles they play, you know, looking at AT Protocols specifically, there's lots of different components. There's they might host your PDS which actually has your content stored on it. It might they might have the app view and they they might be running the feed, they might be doing the moderation, they might have a full stack in which case they deserve a larger bit of that pie.
But um ultimately the creator has a lot more bargaining power and and these kind of other components of the system fall further back into the you know into the background and you know they basically play a role as as a broadcaster or or the rails for which a creator um can access the audience and the creator has the choice over which apps that they use as the surface to reach their audience. Yeah, those I think that was a question over there.
Hi. So I actually use Grays to create feeds and I have one feed enabled for monetization. But what I found in general with the feed creators with Grays is that most of us are trying to create these feeds, but we're not necessarily out there trying to market to get advertisers. Like that's not our main skill set, if you will. Whereas the creators themselves who are trying to monetize their work are skilled at that. And so I guess I wonder both what do you see as a way that we can actually get advertisers even wanting to be part of this system.
Yeah. Like how do you make that happen when you kind of have this weird distribution of creator versus more of a developer general interest person. Yeah. I mean wherever there's eyeballs is going to be value. And right now, like yeah, we there's there's over 40 million accounts on Bluesky, but then there's even less if you look at the actual monthly active users and there are ready interest in advertisers, whether that's through Grays or people, you know, sponsored feeds or or brands participating in the network, there's clear value there. There's eyeballs, there's attention, and where there's attention, there's there's mine to be made, especially if we're looking at you know advertising models, but there's a lot of other models as well, I think which come about when permission spaces become a thing on app protocol.
But yeah, to answer your question, I more that becomes uh a real thing. Sorry. Sorry about that. Um yeah, uh I think I would like to ask uh Nate uh to to pass it to you. I mean we we've talked a bit about more from this perspective of creators and ads and and that being their kind of source of revenue, but for musicians and artists, it's a pretty different uh revenue source and algorithms, especially today with you know apps like Spotify and Apple Music that have uh a real impact on how the music's created in the first place.
Um how do you see kind of atproto changing that landscape. Cool. So I'm really excited about well, first of all, has how many of you have heard of atproto fans? Okay, a good chunk of you. For those of you who haven't, it's sort of like you show up, you hook up your bank, and then other people can show up and pay you five bucks, like coffee or something like this. Um so what's exciting to me is that um in my music app today, um you can um say when you upload a track or something, um you can say only allow supporters to consume this track.
And so I can just, as an app view or an application, just respect that stated support, which when permission data is a thing that can be private as well. Um so it's sort of more it's just like um Patreon style, like creators have relationships that they build grassroots with people that's sort of divorced from the classical conception of distribution or uh the advertising models, and which is exciting to me personally. Um I think the tricky thing so the tricky part in my mind comes around when you ask a question like if Spark is going to present uh cameo of a song that appears on a music platform where the content is gated.
Um how do you trickle the value back to the artist who owns it? How do you quantify um the the value that the artist is providing by allowing their content to show up downstream? Um I I think that is where like I want to press and try to figure out like how do how do I do a handshake with Joe and say like how much how much is it worth uh to the artist for their content to appear there? Um and I think that's a very tricky problem, and I don't I don't have any answers to that right now.
But like that is where I'm really interested in thinking because it does seem like uh the value of an impression depends entirely on like what the content is, first of all, and then second of all the place where you see it. Um just hold up like you can have your phone in front of you, but you just not look at it, and it's like the system might count that as an impression. Um, so it's like there's a lot of different uh human ways that I think we need to think about uh what the value of content distributing downstream is that I I think um is an important problem.
Uh but one thing I'm really excited about is that if an artist doesn't like the way that Joe and I do the handshake as far as like respecting the value of their content downstream, they can just take their content and go to a different app because of atproto, which I think is really cool. And so maybe a different app does a better handshake um in terms of respecting the value of their work downstream. Um yeah. Yeah, I think that's a that's a great point. I mean I you know Spark and player, we're gonna interoperate one day, I promise you.
We'll have you know you can add songs to Spark posts that are you know player records, so an artist's song that they have on their PDS can can start to propagate, interoperate across lots of different apps, not just Spark, but wherever there's there's music. And where these kind of value exchanges happen, I think there's it's really important that there's some mechanism to share value there. Um I think right now we're starting to see like intra-app monetization models where you know Leaflet is launching Leaflet Pro and they have their monetization stream, you can subscribe to Leaflet. Um, you know, you can have there's other apps that are starting to do pro things as well, or Grays has their own advertising.
But you know, even if you're looking at a Grays feed, which which tabs into your question, a feed builder on Grays could have ads in their feed, and Bluesky that are actually taking on the majority of the costs and distributing that content is getting none of that. And sure that works for now while they've got a hundred million dollars in the bank, but it's not going to be a sustainable ecosystem. So I am a bit of a proponent for, and it this came up in the previous talk, micro payments there for the int inter-app ecosystem.
So you know, for example, uh a player song being used on a sponsored post on Spark and a percentage of the revenue being made on that sponsored post automatically being distributed to the artists whose song is being used there. And that also could flow in the other direction where you have niche artists, which this exists on platforms today. There's these kind of shadow economies where uh up and coming musician will pay for a popular content creator to use their song on their post and as a great way to get discovered. And we see TikTok is being such a great discovery mechanism for artists recently.
And uh this doesn't just have to stay within the you know creators like artists and and UGC creators, it can propagate across every other component of the ecosystem PDS hosts, moderation services, app views, labelers, whatever they can all become sustainable models that don't just rely on donations but share in the value that's being created across the ecosystem. So if everyone if anyone wants a chat micropayments on atproto, then please talk to me. But uh yeah, I think we're that question actually this is I'm glad you're asking a question because I was about to reference your most recent post and proposal about more artist-specific metadata on on app protocol.
So yeah, we'd love to hear it. Thank you that you introduced because it's a lot of different things. And if you want to talk about what you what talk about that, please. Yeah. Um my name is Hilke, I'm a I'm a musician myself, and I run uh uh music label, so I know a bit how the music industry works. Um, if you talk about monetization downstream for music that is posed on uh uh that is in a video, we have a system for that, it's called copyright. Yeah, um and I think this has to play a role even if it looks much too complicated for technologists like like many of us.
But this is actually the system in place that the music industry and the big players use. And I think we need to find out with with this protocol how we can map their data to the Ad Proto data. And um yeah, I think that opportunity was missed in the first wave of Spotify and stuff, and then the whole mess that was made, then they had to buy out a bit of majors, and that's how all the indie uh musicians and players got fucked. So I think now we have the opportunity to do this mapping mapping properly, and then the music industry, the labels will get along hopefully with with with this ecosystem that we try to build.
So that is where be that's the reason why I made uh a proposal for uh Lexicon who does actually that um translate the copyright system to atproto. Yeah I mean I I think looking at like TikTok as an example, um you know they have a pretty complete library of licensed music there and they have deals with these record labels and they tend to pay on average 30 cents per song used on a post. And these this is a very static uh dynamic and they are able to achieve that because they have you know economics of scale, they have a lot of bargaining power there and and distribution.
Um and I think looking at app protocol and music on app protocol, at least for now, there's kind of two angles, which is one, there's that library completeness on one end, which requires a lot of capital, a lot of bargaining power that we don't have, and on the other end, it's starting from the indie uh artists, the the the creators at the at their other end that are starting from nothing and and and owning their content from the ground up, we can then over time build that and build that bargaining power and then get those kind of partnerships with with real copyright and real respect for for the laws that are necessary to protect artists.
Um we don't have much time left, but I think there's a necessary necessary topic to cover with the creator economy, and that is AI, which is uh you know, I think there's a lot of talk about gen AI affecting content creation and and verification of what's real and what's not and and that affecting the economies. Um Josh as an investor, like I would love to hear your thoughts on on AI and in the creator economy at large. Yeah, I I think I think Natalie touched on this during the last talk pretty well. Uh she spoke about how the covenant of the web is dead.
Google scrapes websites in return websites get traffic. Uh but we're evolving from a system of search engines to answer engines where you know it's nine, ten times harder to get traffic on Google now than it was 10 years ago. There's some stat it's like 750 to a thousand times harder to get traffic if you're a website uh from from OpenAI and Anthropic and some of these model providers. So users aren't going to websites anymore. Uh they're just getting all the answers fed to them with uh natively within uh an AI chat interface. And AI agents don't look at ads.
Um right. And so you know, for publishers, I'm thinking more news organizations. Uh you know, the the current uh business model of the web doesn't work with the way in which AI works. Um because you know typically these m news media organizations are monetizing traffic to their website and monetizing uh that attention uh and selling ads against it. And so uh obviously Cloudflare, uh Microsoft are doing some interesting work around content data marketplaces. Uh publishers create unique niche content. AI model providers need data to train their models. Um there is an interesting stat. Uh so I I think it was uh I forget which AI model provider it was, so I won't I won't say the name, but um you know for the same amount of data, I believe Reddit got paid it was selling like um you know five to seven X times more for their data.
And the reason for that um relative to New York Times, sorry. And and the reason for that is because the New York Times' content is a lot more substitutable to a lot of other news publications out there, whereas the information that lives on Reddit is is more unique and niche and as a result a lot more valuable to the AI model providers. And so you know while we haven't um you know figured out an AI interface for social yet, I think we still you know swipe through these short form video reels and um you know we haven't really AI hasn't really solved the interface for social in an AI native way uh like Sora just died last week, right?
Um it will be interesting to to see how this business model evolves for for publishers and uh I think micropayments will will be a really big piece of it. Anything to add there now? Yeah, pretty much done for time. If there's any last questions, otherwise we'll we'll wrap it up. Thank you. I think uh yeah, we're all excited for that kind of there's there's a big kind of sea change coming for creators in terms of how much power and the power dynamics there with creators and and we're building that on app protocol so it's exciting. Excellent.