On this episode of The Edge of Show, host Josh Krieger goes deep with Carson Farmer, Head of Research at Recall Labs (formerly Textile), live from Consensus Toronto. Farmer shares how Recall is redefining AI agent transparency by merging blockchain-grade auditability with agent trust scoring. From their viral AI agent tournament drawing 140,000 votes to innovative agent-rank infrastructure, there's a revolution in verifying which AI agents are safe, reliable, and aligned with user values. Ideal for Web3 and AI innovators, this conversation uncovers how Recall is leading the charge in the fast-growing agent economy.
Key Topics Covered
- Founding Recall Labs & Textile history
Since 2016–17, Recall Labs pivoted from on-device ML to Web3 data sovereignty, culminating in today’s agent transparency solutions. - AI tournament & community engagement
Recall hosted its first AI agent tournament, drawing 140,000 votes and spotlighting agent personalities—building excitement beyond typical hackathons. - Agent transparency & trust ranking
Recall develops agent-rank layers—transparent scoring metrics that audit agent behavior, verify alignment, and maintain accountability when agents act autonomously. - Scalability in the agent economy
With billions of agents on the horizon, Recall applies blockchain-like discovery systems—agent-rank akin to PageRank—to manage supply, demand, and trust. - Upcoming cross-chain tournaments & roadmap
Next up: Solana vs. Ethereum trading-agent contest, enabling measurable benchmarking and promoting agent-market visibility.
Episode Highlights
“We got 140,000 people voting on their favorite agents…creating personalities around these agents.” — Carson Farmer
“We need checks and balances…these like agent trust ranking layers.” — Carson Farmer
“Capabilities are going to commoditize…differentiation is about do I trust this agent?” — Carson Farmer
“Recall is building…an agent‑rank analogous to PageRank to identify best agents.” — Carson Farmer
“Imagine…microtransactions that my personal agent handles…leveraging intent‑based discovery.” — Carson Farmer
People and Resources Mentioned
- Carson Farmer
- Recall Labs (formerly Textile)
- Josh Krieger
About Our Guest
Carson Farmer is Head of Research at Recall Labs, previously Textile, and a founding team member since 2016. With expertise in on-device ML, Web3 data protocols, and AI agent design, Carson leads Recall’s research into agent transparency, trust ranking, and blockchain-powered audits.
Guest Contacts
- LinkedIn: (not publicly available)
- Website: https://recall.network
- Twitter: https://twitter.com/CarsonFarmer
Transcript:
Josh Kriger: Hi, everyone. This is Josh Krueger, co-host of The Edge of Show, live at Toronto Consensus. I'm having a great time meeting some great folks here. And right now I'm with Carson Farmer, who's the head of research at Recall Labs, formerly Textile. It's great to have you on the show. Cheers. Thanks, Josh. So I met some members of your team recently and was excited to continue the conversation on the air about what you guys are doing and some of the implications of this aging economy. But first, maybe give us a little background on yourself and what got you excited about Recall.
Carson Farmer: Yeah, cool. So, I've been on the Recall team since before we were Recall. Pretty much day one, I was on the part of the founding team. And so, we've been around in various forms back since 2016, 2017, something like that. where we sort of formed this startup around sovereignty over user data. And there was a bit of a meeting and finding people who are interested in similar things. And we had sort of tried to create this startup around that. And we started to realize we didn't really like the way that the internet was moving in terms of user data. and personal data and personal data sovereignty. So we thought, okay, we should probably try to address this. And we built a company called Textile around that.
Josh Kriger: And when did the AI sort of come into the vision?
Carson Farmer: Well, so really early on, we actually were doing, back at that point, we were doing a lot of really early machine learning on-device data modeling. In fact, the whole point of that was really to try and make it so that people could leverage intelligence on their mobile devices without having to go out to the cloud and give up a lot of their data. But what we found at that time was the tech really wasn't ready yet. It was really hard to run like full deep neural nets on device without like having the phone heat up in someone's pocket, basically. And a lot of the like data companies and the AI companies, they just like wanted the raw data, right? So that they could build better models in the cloud. And so we kind of started to pivot away from that, like on device machine learning and moved a lot more into the like web three data sharing and data protocol. layer.
Josh Kriger: And did you have a sense this AI agent proliferation was coming? And how has that impacted your business?
Carson Farmer: Yeah. So, I mean, way back then, we had no idea that, you know, things are, I mean, no one. Yeah, no one's impossible to predict. But the really cool thing is like we had a really good sense of the fundamentals and like how to do a lot of that stuff. And then sort of fast forward to today, we've done a lot of like backend web3 database stuff. So we know the data, we know the data space really well. We already had a lot of experience in AI and machine learning. So when the sort of like opportunity of LLMs and AI agents sort of presented itself, I mean, in the last couple of years, We were already really primed to just go like, oh, great, we get to go back to our roots now. The tech is ready. The need and want of the industry is there to actually deploy something pretty quickly. And so that's where we kind of came up with Recall, which was this perfect marriage of like what we knew really well from early days to since then. And yeah, that's what we're building.
Josh Kriger: You guys are having some fun. There was recently an AI tournament that you all did that caught some attention. Can you tell us a little bit about that?
Carson Farmer: Yeah, that's been, I mean, that's kind of blew us away, the attention that we got there. We knew this was going to be something that, you know, people are interested in. We've been going to conferences and talking to people, building agents, especially in the sort of like DeFi space, DeFi AI space. And so we were talking to developers, we said, like, let's build a forum for them to actually compete. And there's a couple of really good reasons why we wanted to do that, which I can get to in a sec. But the biggest thing was we got like, I don't know, 140,000 people voting on their favorite agents. It kind of ended up creating like personalities around these agents. That's a level of engagement you don't get with a typical hackathon, you know? No, and this is the like, yeah, it's really exciting. I mean, competitions, we people love competitions, right? Like, we have the Olympics, we have like these competitions, because they, they kind of like induce demand for something that maybe like, the market hasn't yet naturally created a demand for. So what were the top three agents? Oh man, I'm not even gonna remember their names. Oh, the concepts, like some of the winning concepts. Yeah, so one of the winning concepts, actually the one that won, and we'll have to put maybe some show notes or something like that with a link to their product, because they're an actual production trading agent. And they so people you they have a Twitter like you can or X rather and You can actually like engage with them and they're like a full shop that actually builds AI agents And so they have a like really multifaceted strategy that I would never be able. I'm not a finance guy so I wouldn't be able to get do it justice, but They're an actual production product, right? So like people actually can hire them to manage portfolios and things like that
Josh Kriger: Very cool. So you kind of alluded to something that was one of my next questions, which is sort of like, we use these agents, but we don't always know what sort of the agent looks like behind the scenes, the tooling, the engine. Like, you know, you can't, as a non-tactical person, always dissect these things, take them apart. How are you all sort of thinking about the transparency side here of people sort of using agents like, you know, like all the time, like it's going to become ubiquitous. And you know, what happens if one has a nasty virus or sort of has a split personality and becomes evil after a week or two of using it?
Carson Farmer: Yeah, these are really important questions. I mean, the whole AI industry is thinking about things like that. And the cool thing about our industry and blockchain in general is like, we already have a lot of the tools in place to try to address some of these problems like observability, you know, auditability, we've got these like blockchains, which are perfect for a sort of like audit traces of behaviors and things like that. So we're really kind of dialing in on a lot of those, that underlying technology. I'm our head of research, so this is something that I'm really excited about. And so building out the trust layer for AI agents, making sure that they're actually doing things that you want them to do. A good example I like to use is I do investing outside of the crypto space and in the crypto space, and I have my own investment thesis. And one of them is I like to invest in environmentally sustainable And if my investment firm just started investing in oil or something like that, I wouldn't be too happy about that because that's not part of my thesis, right? The same thing, I wouldn't be too happy if my AI suddenly started investing in things that I wasn't particularly aligned. So we need checks and balances to actually address that. So that's a lot of what Recall is starting to build, is these like agent trust ranking layer.
Josh Kriger: And where you can kind of inspect or put your agent, potential new agent, through the metal detector and see what's going on there.
Carson Farmer: Totally. I mean, you wouldn't hire someone if you didn't look at their resume and see what they've done before and look at their track record and all this kind of stuff. So you probably shouldn't be doing the same thing when we're dealing with agents. And agents are a little different than a traditional web app because A traditional web app is like, you know, very deterministic. It's been audited and built by some, you know, development team. And a lot of the AI tooling is a lot more autonomous, right? It's making some decisions on its own. And so we need some audits ability and we need some verifiability in that mix. And, you know, capabilities are going to commoditize, right? Like, it's going to be really hard to tell if this agent is like 0.25% better than this agent or, you know, a fraction of a cent cheaper than this agent. So the differentiation is really going to be about do I trust this agent? Does it have a track record of not leaking private keys? Or does it have a track record of like actually staying on task and investing in the things that it, you know, that the users care about?
Josh Kriger: So I guess let's fast forward a little bit more. It's sort of a given at this point that there's going to be this proliferation of billions of agents at some point. I heard some stats around the amount of traffic on the Internet that's basically bots at this point. It's just increasing every day. What does that world look like in terms of the capabilities of recall and other products like you to sort of deal with that madness? Like, I just think of like, you know, a billion people trying to use the same subway car at the same time. You know, how do you sort of ensure that level of transparency and visibility can be scaled in that sort of world of so many agents competing for attention and competing with humans for attention?
Carson Farmer: Yeah, no, that's a super good question. I mean, and actually to extend your analogy a little bit, I think the key thing is that we want to be able to make it so that we get onto the right subway car at the right time to be able to sort of differentiate. And we kind of have some experience with this on the internet already. We've already got billions of websites, or I don't know what the number is, millions maybe, probably billions. We can ask AI. Yeah, we can ask ChatGPT that one later. But for sure, we've sort of managed to navigate that. And so companies like Google have done a lot of indexing. Google's PageRank is like a really great analogy because it's sort of a whole industry built up around it with like sort of high frequency ad trading and all that stuff. And that was driven by the structure of this proliferation of websites. And so recall it's I mean we've got a sort of ranking system that we're building for these agent capabilities and things like that that is sort of analogous to a page ran an agent rank if you will for agents. And the idea is it's that's a it's a discovery layer right so you can actually identify what agents are good for what thing you wanna do. and actually start to get to a point where you can do intent-based discovery and analysis. We're moving from search, where I have a question that I'm trying to find an answer to, to just finding right away. So imagine I have a question or I want to know, what's the weather going to be like next week? I don't want to get a bunch of pages that suggest that maybe the weather might be... What I want is an answer, ultimately. And I don't want that answer to come with like an ad. I just want the answer. So I'm probably going to pay for that. And I'm going to get, it's going to be tiny microtransactions that my personal agent is going to go out. It's going to get an answer. It's going to come back to me and it's going to answer it. And it's going to handle all that payment and stuff for me. And ideally it's going to leverage something like agent rank or something like that, that recall is building. I mean, you know, to try and get at the best answer possible.
Josh Kriger: Very cool. I'm curious to use some of these products and services and check out some of these agents. Is there anything else on your roadmap coming up that you wanted to touch on?
Carson Farmer: Well, so we're doing these competitions. We just finished one. It was a huge success, a surprisingly huge success. We've got another one coming up that's kind of interesting, I think, for this crowd. Yeah, what's the theme? We're pitting Solana and Ethereum against each other. All right. If you've got an agent that is trading on Solana or trading on Ethereum, let's see who kind of comes out on top in that. All right. You know, it's it's a it's a fun it's a fun one, I think. But yeah, hopefully we'll generate a little bit of excitement and interest. So we're doing that one coming up soon. Anyone who's building a like a trading agent that's particularly good with like seven day P&L, they should totally join. It's really easy. And then it kind of ends up generating a platform for your agent as well. So if you talk to the winner from our first competition, they got huge inbound afterwards. I'm sure. Yeah, of interest. And it's just kind of fun to start with. And we're starting with competitions and we're starting with, you know, seven day P&L trading. But there's so many more skills that we want to test, right? Like there's so many more things that agents can do for us, but we want to make sure that each step is sort of very verifiable and builds up this ranking system.
Josh Kriger: Very cool. If people want to dive in more to Recall, where should they go?
Carson Farmer: So, I mean, the landing page is the best place, just recall.network. You can follow us on X. And then are you on X, too? I'm on X, yeah. Carson Farmer is my handle. And I've got a little cartoony face that looks roughly like me, so you can kind of get a sense of who's there.
Josh Kriger: Awesome, man. Well, great to hang out and pleasure to learn more about what you guys are up to. Definitely going to stay in touch. Yeah, thanks a lot, Josh. It was great. Hi everyone, welcome back to the Edge of Show live at Consensus Toronto 2025. It's great to be here and the showroom is bustling and I'm so honored to have two very special guests on the show with us right now. We have Elena Sinelnikova, who is the director and co-founder of Menace and one of their senior advisors, Natalie Ameline. Did I say that right?
Natalia Ameline: Yes.
Josh Kriger: All right. And you have a very special story as well and excited to get to know both of you better. Thanks for joining us on the show. Thank you. So I guess my first question, because I think it's always great to see how things come together is, you know, you started Metis. How did you all meet and what drove you to sort of become excited about this project?
Elena Sinelnikova: We met a long time ago, that was back in Russia. We both were in the same university and as we just were accepted in the university, we basically were standing in the area where all the students gathered and I saw Natalia. She attracted my attention and I came and introduced myself. So that's the story. And that's the beginning of the story.
Josh Kriger: That's great. And I guess at this point, Natalia, obviously, we should let our audience know that Vitalik is your son, a remarkable human in our industry. How did he inspire your passion for blockchain and get you excited about what's being built at Metis?
Natalia Ameline: You know, it's interesting, right? So normally we have parents as our role models, right? So for me, my son kind of became a role model for me in that sense as well. As you know, he got into the space very, very early and it is from him that I learned about Bitcoin and ultimately got introduced to Ethereum and everything that the platform is enabling. So from there, the inspiration was very easy because while seeing the enormous potential that blockchain can create for the society, it was very difficult to stay away from it. And hence the first opportunity, I jumped and I joined Alena and Metis.
Josh Kriger: Well, I don't know if it's nurture versus nature, but I will say, you know, of all the iconic leaders in our space, I think he has one of the most pristine reputations. So I give some parental credit to you as well. I'll take it, why not? Let's talk about sort of Metis and sort of where you guys were maybe back in the day and how you've evolved as a technology stack.
Elena Sinelnikova: So when we co-founded Metis with my co-founders Kevin Liu, Yuan Su and Steven Goh, me and Natalia were still at Crypto Chicks because we're also co-founders of Crypto Chicks and Metis was basically incubated by the Crypto Chicks.
Josh Kriger: Crypto chick related initiatives since then so so I guess copy is the greatest form of flattery, right?
Elena Sinelnikova: But we started it back in 7 2017 so I think it all went from that great So for all those ladies that came later said they thought they were somehow inspired by what you did there.
Josh Kriger: That's great. Exactly
Elena Sinelnikova: Exactly. So yeah, but when Metis just started, we already had a community at CryptoChicks. And basically, all people that we knew, all people that worked for us in CryptoChicks, they are right now actually working at Metis. So Metis has just allowed the CryptoChicks to have their own platform to adhere to our vision of decentralization. of having the platform where people and companies can join together, people can find employment and companies can find the business without dependency on any central party. So that was the vision of Mitis from the very beginning.
Josh Kriger: So let's dive into Hyperion now. Like this is a L2 that sort of, you know, has some different qualities relative to other L2s in the market. And when did that sort of come about and sort of what are some of the differentiators?
Natalia Ameline: You know, we've been watching the developing industry, blockchain and AI, of course, as well, as it's really became prolific in the Web3 space. And we identified some major gaps there. One of the gaps is inability of today's blockchains to actually support AI workload and applications. So, therefore, we understood that we, while Meters Andromeda chain, our main chain, is great for everyday, all-purpose kind of things, AI needs something a little bit more robust. And that's how the birth of Hyperion came into play. Hyperion is the network which is really focusing on high performance. high throughput, very fast transactions, you know, between one and two gigahertz per second, and very low latency, practically instant transactions, which are enabled by our parallel processing algorithms. So all that enables Hyperion to handle those AI work demands and makes it very different from other applications today.
Josh Kriger: Give us some specific use cases for this type of technology that someone could do right now. And also, things are moving so fast with AI. So you have to also think about the use cases that are going to be relevant in two, three, five years. And so I'm curious, what are the use cases in particular that stand out to you in terms of your network?
Elena Sinelnikova: Absolutely. So the idea of actually the whole AI thing on the blockchain is spun from our philosophy of AI should be human-led and human-aligned. Because right now, with all AI developed by the corporations, this thing gets overlooked. So, therefore, we are getting now together people that are interested in developing their own AIs that are adhered to their values. Like, for example, if you have kids, you can develop the AI that is carrying the values to your kids and going to be like a babysitter to your kids, the supporter, the mentor. Yes, this is like one of my favorite use cases. that could be implemented. Yes.
Josh Kriger: And also like it's it's that it's that sort of having that person with you, you know, even if they maybe live far away or have passed on. Right. It's a it's a it's carry on that tradition of mentorship. That's cool.
Natalia Ameline: Yeah. Yeah. And in addition to AI use cases, of course, there are traditional cases, for example, that will allow to expand existing industries. For example, DeFi, we can now incorporate high frequency trading or derivatives on AI, which before were not necessarily possible because of limitations of the speed and throughput. The same goes with the gaming. You know, obviously there are games on AI, but it's always But one of the impediments always being this instant ability to process a transaction, which Hyperion offers.
Josh Kriger: And I guess another big part of this is Lay's AI. Could you talk a little bit more about that?
Elena Sinelnikova: LizAI is where exactly the alignment is happening between AIs and humans. Consider it as a kind of governance, but I don't like the word governance. This is the alignment platform where actually humans can come together, align, be accountable for the data, for the training data, what it is exactly that we're training AI on, who is it exactly coming from, so where is it coming from, and then to hold each other accountable, AIs and humans.
Josh Kriger: How does that connect with Hyperion?
Elena Sinelnikova: Because on Hyperion, we would like to see the ethical AIs, the ones that are trained from this data that we know what it came from, that is not manipulated by any single party, but it actually belongs to people and that's responsible to people.
Josh Kriger: Yeah, that makes a lot of sense. So as part of the implementation of this alignment, you're also launching tools, right? And one of them is the data anchoring tokens or DATs. I'm learning some new acronyms today in this interview. Tell us a little bit about what those are.
Natalia Ameline: Yeah, so essentially data anchoring tokens is one of the major innovations that we're introducing that will be launched within, let's say, LSEI. It's a new kind of token standards that allows you to kind of capture and trace the data let's call it data provenance. So it's history, how it's been trained, who owned it, who contributed it, and where it's been used. So this token will be able to capture all this information. And then we want the information or the data to have value. Because one of the issues today is we, you, me, we share our information publicly. And these centralized entities, they harvest our data without our explicit consent and train their AI tools on it, and we don't get anything in return.
Elena Sinelnikova: No, we pay subscriptions to them.
Natalia Ameline: Yes, we actually pay them for giving them our data, exactly. And that is a very major imbalance in control and economics. So we want to kind of turn it around and the data anchoring token is one of the ways to do it because it will hold this, our personal information. And because the tokens lives on chain, we will be able to set the rules, who can use our data, how to use it, and actually earn rewards for users using our data.
Josh Kriger: That's really cool. So, you know, AI and blockchain at intersection is becoming more and more prominent, right? And you all are not the only crew that's thinking about how to solve some of these problems. You know, we live in a decentralized world with blockchain and we sort of, there's a lot of different products and services out there. I'm curious if you view, you know, all these different products as competitors or potential collaborators. like is sort of the Videlic sort of EVM model sort of in your mind in terms of how what you're doing can interact with other AI projects? How do you look at the landscape?
Elena Sinelnikova: Absolutely. We operate under our philosophy. Human lead, AI advises. Human decides, human accountable. So whoever is with us on this philosophy, we accept them with open arms, whether they competitors or not. So we would love more people to understand that humans should lead in this AI economy.
Natalia Ameline: And, you know, our technology, specifically the Metis SDK that powers all this innovation, such as Hyperion and LessAI, it enables the cross-chain interoperability as well. So that allows us to seamlessly bring other players into our network and share the all the economic successes that can be built within it.
Josh Kriger: So the doors are open.
Natalia Ameline: Absolutely.
Josh Kriger: Makes sense. If folks want to get more involved, learn more about Metis, Hyperion, where should they go?
Elena Sinelnikova: They can go hyperion.metis.io. So hopefully you can post that. So yes.
Natalia Ameline: And for Lesii, it's lesii.network.
Josh Kriger: And ladies, I assume you're both on Axe as well. Absolutely. What are your social handles?
Elena Sinelnikova: Elina Kryptychik, without the K at the end, the C at the end.
Natalia Ameline: Natalia underscore Ameline.
Josh Kriger: All right. Well, such an honor to meet you both today and really excited to follow your journey and what's next. Thank you for joining us on the show.
Elena Sinelnikova: Thank you. Thank you.
Josh Kriger: Hi everyone, welcome back to the EDGES show. This is Josh Krieger, your co-host, and we're live in Toronto at Consensus 2025, and I'm here catching up with one of the other co-founders of Sapien. We had your other co-founder, Rowan, on the show. This is Trevor Coverco, and it's great to have you on the show, man.
Trevor Koverko: Yeah, cool. It's exciting to be here. We got Toronto, my hometown. Oh, really? Everyone's coming out to me for a change. Usually I'm running around the world, you know?
Josh Kriger: I gotta say, I'm enjoying the food, I'm enjoying the vibes, the weather's been great. I hadn't been to Toronto in about 10 years. I'm like, I have to give this city a little bit more attention. It's been okay. We were a little mini mini New York vibe going on up here, Portland, New York Yeah, kind of a combination cool Well, I I guess that sort of precedes the question of like what brings you to Toronto? But I think you know for you like why is an event like consensus important at this current moment?
Trevor Koverko: Listen, we you know as a proud Canadian. We have a lot of street cred when it comes to early crypto and early AI, two things I'm working on now. And whether it's kind of the early Bitcoin scene here or Ethereum being built and created and founded out of here where Vitalik's from, or whether it's a lot of the early forefathers of kind of the modern AI movement like Jeff Hinton at UFT, there's a lot of really good technical talent here. And it's just an amazing city for global events like this. It's a very kind of cosmopolitan city compared to a lot of U.S. cities.
Josh Kriger: And obviously events are a great opportunity to amplify some of your recent milestones and achievements. Let's talk a little bit about your partnership with Alibaba. I guess there was just recently a quote on their end.
Trevor Koverko: Yeah, no, it's, it's amazing. We just to back up a little bit, we are a data company. So our excitement is around how do we produce enough data to meet the demand from these new AI models. So when you build a model, there's kind of three things you can. If you guys say this, it's true. It's easy to run out of data, kind of surprisingly, because you think the internet is just an unlimited amount of data. But as it turns out, it's all been used. All these big models that you've heard of, ChatGPT, Gemini, et cetera, they all train on the same open internet data. So the next step change to get better performance, to get to AGI, we need new data. And we need actually, ironically, humans like expert human knowledge is the next kind of important missing ingredient.
Josh Kriger: But how soon will we run out of that data? I mean, is that I mean, is that like one, two years of data? Or do you think there's a longer window there?
Trevor Koverko: Yeah, there's there's a longer window. I think There's a lot of other components as well. But in my view right now, this is the biggest bottleneck in AI. We call it the data wall, where we just ran out of data. And now everybody's scrambling to get a data edge. So just as a quick example, if you're a company like JP Morgan or something, you're sitting on this treasure trove of data, partially because it's really good human data from their customers, and partly because no one's trained on it yet. So what are you going to do if you're an executive at J.P. Morgan? Are you going to sell access to it like through licensing? Or are you going to say, no, this is going to be a core competency of our future. We're going to build our own model. So we're going to see a J.P. Morgan model that's not as big as ChatGPT, but it might be better in that specific vertical, which is finance. So it's kind of an exciting area for enterprise right now where all these companies are building their own models. And that's kind of what my company helps with is these companies to, to build models and get the right data for those models. Cool. And there's new partnerships forming like with Alibaba? Exactly. That's kind of on the demand side. So we have like an enterprise team where we have customers like Alibaba, Baidu, Mid Journey. We have Amazon's self-driving car unit. So we supply data for all of them, which is really fun.
Josh Kriger: What kind of data are they looking to get?
Trevor Koverko: Well, there's different types. So we call them modalities, like text-based, video-based. There's also different data formats. So it's a bit complicated. But I'd say right now, one of the most exciting areas of data demand is in what's called 3D, 4D data or computer vision data. And so that's like robotics and self-driving cars. And that's like creating more data for these robots to actually like behave like humans.
Josh Kriger: Cool, and I guess another area of data that I was just talking to one of the founders of Elizabot is human-to-human data. That's really important too. Is that something you're playing with as well in terms of the human-human interaction?
Trevor Koverko: Yeah, and even though we're making these announcements with folks like Alibaba, our real core competency that we're known for is actually the other side of our business, which is the humans. So we have almost a million humans now. We just crossed 800,000 customers in the form of users. And these people, selfishly create data for rewards. Think of it like a gig working job. And that's one of our big theses is that the future of gig work is going to be this type of stuff where you have models demanding human knowledge and then we sell it back to the model.
Josh Kriger: You have a sense of like the average amount of hours they're putting in and what they're getting back yet.
Trevor Koverko: Yeah, it's kind of exciting because not only can you make a, in some cases, you can make a pretty good wage by doing this, but we can actually... Better than an Uber driver. Exactly. And it's more accessible because guess what? Uber drivers, you need a car, or you need to hope that robo taxis don't come around in the next five years. Or if you're a DoorDash person, you have to hope drone delivery doesn't disrupt it. But for this kind of stuff, it's literally if you have an Android phone and an internet connection, you can make a living wage. And think of the profound kind of impact of that is poverty goes away because no matter what, even if you don't have any skills, if you can't even speak English, that's okay because AI models still want your humanness to help input data tasks. And so Is it the most glamorous job? We're actually trying to make it more fun. But the key point is you can make a living wage as long as you have an Android phone and an internet connection.
Josh Kriger: And what's, I think I talked to Rowan a little bit about this, but what is a living wage to you mean and how do you maintain that living wage as more and more people enter your platform and trying to balance that supply and demand? I mean, that's the long-term concern, right?
Trevor Koverko: Yeah. And just to give you some data on that. So the demand side is crazy right now. So the problem is we don't have enough humans that know about this yet. Okay. And so that's our big bottleneck is getting more humans. That's part of our mission statement is to make it more accessible for the largest network of human data inputters. And, and what, what's cool about that is that the more humans you have and the more diversity you have, the better quality data you collectively produce.
Josh Kriger: So how are you finding new humans and how are you sort of casting in that geographically?
Trevor Koverko: We have a lot of, a mix of countries. So we're in over 110 countries right now of, of different humans labeling data. And our, our leading countries always change. It's kind of interesting because it started out as Southeast Asia. It was like the Philippines and places like that. And then just looking at it this morning, it was Germany is our fastest growing region. And we have Hong Kong as well, was near the top of the leaderboard. So it's really cool to see that's part of the, one of the reasons the, um, the web three part of our business. Are these smaller countries that have popped up recently? Nigeria. Interesting. Nigeria. If you've heard of companies like Scale.ai, they're one of the companies we were inspired by. They have a large network in Africa as well, where people are doing this for a new form of wages, which is pretty cool.
Josh Kriger: That's great. That's great. So any predictions on how this type of like, I guess, economy is going to impact the broader economic landscape over the next five or 10 years? Have you thought about that?
Trevor Koverko: I think this is going to be the biggest employment in the world. It's going to be a new form of gig work, which is like humans helping models get better. And it's kind of interesting because all you hear about is, oh, AI is going to replace humans. It's actually the opposite. It's AI employing humans for the benefit of everybody. And we're making AI better, which is going to help eliminate poverty and cancer and all these things as well. So I think it's a net win for everybody. And we're excited to just bring a whole new workforce to the economy. And you don't have to work in a monotonous job in the real world. You can do something fun and rewarding as well. Cool.
Josh Kriger: Anything else on the horizon you want to just touch on before we wrap up?
Trevor Koverko: Yeah, you know, just our app now is live. And like I said, we have 800,000 users in just a few months that have signed on. So that's making us one of the most used apps in all of Web3 in terms of active daily users that are using it on a daily basis. And check it out, game.sapien.io. That's where we're at. All right. And are you on X? We are.
Josh Kriger: And you personally, so where do people go to find Sapien and you on X?
Trevor Koverko: So at PlaySapien, we kind of have a gamified vibe to the product. And I'm just my name at Trevor Coverco. Thanks for hanging out, Trevor. Oh man, that was fun.