AI Modeling Agencies with Lush AI

Lush AI is an AI companion for modeling, specifically in the creation of hyper-realistic AI influencers. The company is launching Lush Chat, an AI-powered OnlyFans platform where users can chat and interact with AI models. The platform aims to provide a more personalized and engaging experience compared to traditional OnlyFans models. Lush AI uses a combination of generative AI software and proprietary AI models to create realistic AI influencers. The company plans to monetize both fictional and real people, with a focus on speed of execution and building a market leader in the industry. The Lush token is used to share revenue with token holders, and the platform plans to bridge on-chain and off-chain payments for a seamless user experience. The conversation explores the potential impact of AI models on human trafficking and the role of Lush Network in providing a platform for individuals to engage in modeling without fully exposing themselves. The discussion also delves into the creation of a decentralized marketplace for AI models and the potential for AI models to replace human influencers in various industries. The conversation highlights the importance of real people in the industry and the challenges of building a platform that combines AI and crypto. The chapter concludes with a discussion on concerns and risks associated with the project and the importance of transparency and trust in the crypto space.

Twitter: @LushAIAgency

DISCLAIMER: This is not financial, legal, or tax advice. The guests and hosts may hold some of the assets discussed on the show.

Twitter: https://twitter.com/redcandleclub
Youtube: https://www.youtube.com/@redcandleclub

Dan Kazenoff: @DanKazenoff
Lawrence: @crypto_sieve

DISCLAIMER: This is not financial, tax, or legal advice. Hosts and guests may hold some of the assets discussed on the show.

Transcript

Lawrence: 0:13

All right. Hello, everyone. Welcome to the Red Candle Club. And today we are interviewing Lush AI, you and I just give a quick like intro on yourself and you know how you got started in crypto and How you came to start lush AI?

Eunn: 0:28

Yeah Well, good morning everyone and thanks for the intro. I'm happy to be here. So Yeah, my name is Yoon. I'm the founder of Lush AI. I guess the elevator pitch would be that Lush is using generative AI to create hyper realistic AI influencers that are essentially indistinguishable from, you know, real humans online, right? You can think of these influencers as your Instagram model, as your TikTok dancers, as well as your OnlyFans girls, right? But before, before I get into the project, I'll just kind of do a quick background about myself. I have both a finance and a tech background. So I've kind of like traditional Silicon Valley dropout story. Right. So I left a university when I was about 20 booked the one way flight to the Silicon Valley area to try and, you know, partake in the startup scene there. Right. That was about like 10 years ago. Wasn't able to really raise capital cause I was like super young then didn't have any connections, but you know, I was fairly quantitative. You know, I think kind of mathematically. So I picked up trading relatively quickly. And that's how. Grew my portfolio even before transitioning to crypto. And it's something that's still, you know, due to this day running, you know, trad five trading outgoes. In the background but then, you know I got into crypto during the 2020 cycle and because of my engineering background, it was, you know, relatively easy for me to pick up actual blockchain development. So, you know, writing smart contracts doing, you know what three engineering since the last cycle and then during 2023 was when kind of the beginning. A couple of months was when I believe chat GPT came out and I like knew right then and there, you know, that the next frontier was clearly AI, right? And so I've been training AI models ever since 2023 for over nine months, specifically in the generative AI space, right? Creating videos, images with a special emphasis on the hyper realism. I was originally going to go down the route of, you know, web 2 monetization you know, building like a subscription based you know, website or something, and then taking, you know, subscriptions, gift payments and whatever from, you know, guys as a means of monetizing my, you know, technology and the AI models, right. But then, you know, crypto really picked up in the fall of last year, and I had to kind of put that on pause and reallocate back to crypto because I was fairly under allocated. During the process of doing, you know, some on chain, you know, early stage investments. You know, met and connected with a bunch of other kind of AI related projects. One of AI. They're essentially building like uncensored LLM. As part of, you know, getting in contact with the community, they kind of gave me the idea of, you know, building like a crypto project around the technology that I was already developing and monetizing Web2. And that is, that's kind of the the origin story of Lush, right? And so we were founded in late December. Did a quick I guess kind of early kind of, Community round for about like 120 K a that sold out in a couple of minutes, which we then used to see the liquidity on a theory and I've been just growing kind of organically since then. Right. And we are essentially at this point, gearing up for the launch of our first product, which is lush chat. It is a AI powered only fans platform and it's set to go live. At the end of this month, so less and less than like 10 days and, you know, we're we're gearing up for that. And that's kind of brings up to brings us up to present day.

Lawrence: 3:41

Very cool. All right. So you, you have this chat product and you're, you're releasing it in 10 days. So. Is this supposed to be like for, is it like an OnlyFans type thing where people can, can chat with, with their models on OnlyFan?

Eunn: 3:58

Correct. So the, the initial, the initial launch of the platform is supposed to be geared towards the Web3 community, right? Essentially to showcase our technology. And our, you know, project what that means is there's going to be like a gamified kind of airdrop and point system to really kind of drive virality and like, you know, adoption from the crypto community, but in kind of actual like production use the primary audience, the primary customer are going to be, you know, traditional kind of web two guys that are getting like, Funnel from, you know, social web, two platforms, which is Instagram and tick talk, right? Once, you know, these, the AI models that we develop the first one being Jenny she has a Twitter which I guess I can link to you guys later. So essentially all of our AI models are meant to be essentially, like I mentioned earlier, indistinguishable from real people online. Both visually and behaviorally, right? What that means is that each one of these girls or, you know, later on, guys will have kind of their social media presence, right? And their story and their narrative behind each one of these people. And so, for example, from their Instagram profile or TikTok profile, whatever those, the content generation will be, you know, from our AI, right? And then eventually even like the conversation and the DMS and even the engagement will be powered by AI as well, right? From those platforms, such as Instagram. We will then, you know, just like how an OnlyFans girl links to her OnlyFans page, except in our models case, they will link to our, you know, own kind of internal OnlyFans platform where guys will go to not just, you know, pay for pay per view content, but also, you know, chat and interact. Right. And so when I mean AI powered platform, I mean, from both a content creation perspective, as well as a conversational perspective.

Lawrence: 5:40

Got it. Okay. Interesting.

Eunn: 5:45

Yeah. And I'd be happy to give you guys early access to, to our product launch, right? We're keeping like the group of people small, but I think I'll let you guys kind of have early access and play around with the applications just so you guys can see, you know, what the actual kind of value proposition

Dan: 5:57

is. That'd be cool. Would we be able to like screen share certain aspects of it if we got access or are you trying to keep it like closed off?

Eunn: 6:07

Yeah. I think, yeah, a couple of screenshots here and there is fine. What we, how we plan to structure the launch is that it's going to go in three phases, right? Phase one is I guess where we can fit you guys in, right? It's like the early in the insiders, the strategic advisors and like the partner kind of KOLs, community leaders, et cetera, right? That's going to last about 10 to 14 days. And then after which phase two is going to launch, right? How phase two is going to launch is phase two is essentially the partnered communities. Of the individuals that we onboard in phase one. And so if the people in phase one will have like a referral code, I guess you can like even post it in Twitter or like discord or telegram, whatever. And then the guys that you guys are associated with will be able to get into turn phase two. And then phase three will be like the general public. So that's a, that's kind of the overall structure of how the launch is going to work.

Dan: 6:55

Cool. Yeah. I'm wondering like how So you said like you didn't have a background in AI originally is more like kind of quant trading stuff How how did you build this? Like what's the team makeup at the moment? And Yeah. Just curious to learn more about the model and things like that.

Eunn: 7:14

Yeah, I guess I'll separate that into two questions. I guess first of all, I guess I'll cover the team. Right. So I guess you could think of me as like the Swiss army knife type of founder, right? So I've, I've like AI training background, the software dual and background, as well as like the finance and training background, right. Our, I guess our CFO guy or the, like the finance investor kind of side of the business, he. He's been kind of in Bitcoin since 2009. So he's been, he was like mining back in the day runs his own VC firm, you know, built, built his, his capital from like 50 K to a hundred, I believe over a hundred million at this point. So he handles like the investors relation side as well as the marketing side, right? So. On the other side, we have this kind of see, I guess like CTO guy. So it's responsible for like the building out the, the application or not. That guy's been a full stack developer for 10 years. I met him during the life cycle when we were, you know, building kind of a blockchain based protocols. Right. So he's kind of handling the leading the engineering. My, my friend on the finance side of using, leading the investment relations and I'm kind of like, you know, managing all components of this business. Right. And then, so regarding your other question was, I believe, you know, Like the, I guess the, how the AI training and AI models work. And so within our kind of like toolkit, right, we have kind of various, I guess techniques and like, depending on the circumstance, one may be appropriate, you know, relative to the other. For example, we have, you know, fully generative AI images, video, as well as kind of AI assisted kind of deep fake, right? So it depends on really the unique kind of situations. Regarding the AI models itself there are, you know, I guess AI, generative AI, you know, software such as, you know, Dali and mid journey from these kind of closed source, you know, web two companies, right? The, the software that we use is actually called stable diffusion, which is an open source you know, software, I'm not sure if you guys know of it, but it's basically like a base layer where, you know you can train multiple layers on top of it. So like, I guess an analogy would think about it is that like, you know, there's like just like how in in blockchain, there's like an L1, L2, whatever, et cetera. Right. So stable diffusion is like the base layer and then the, the training, you know, maybe there's like a layer on top of it where, you know, The, the, the AI is, you know, is trained to generate perhaps Asian people or Slavic people, et cetera. And then there's like another layer on top of that, where you're like fine tuning specific kind of individuals, facial features and body parts. Right. And so most of the AI training that we do is on layer two, where we're like Specifically honing in on like the, the facial features of like individual people, for example, Jenny, our flagship AI model in order to achieve consistency, consistency in content generation across, you know, various environments, lighting you know, clothing, et cetera.

Lawrence: 9:56

Got it. Okay. Interesting. So I think I'm getting a picture of the vision here. It's essentially like. You're, you're creating an only fans platform where you have digital models that are all AI chain trained, and then you have the chat where so that you just can chat with the models and feel like they're generating like a personal relationship with them. Correct. And you know, you're, you're building basically a platform for that to be monetized, you know, so you can monetize the chat.

Eunn: 10:22

Okay. Yeah. So I guess I would, how would describe it is we, the reason why we say vertical, like luscious of vertically integrated Influencer modeling agency. What that means is that, for example, OnlyFans as a platform generates let's say 20% of like the fees, right? So they take whatever revenue goes through the platform, they take 20%. But what I mean by vertical integration is that not only do we control the platform we also the biggest actual risk factor, or this biggest kind of like well. To call it risk factor, like the human girl or the human model is the biggest risk factor in this business, right? Because the churn rate of OnlyFans girls is extremely high, right? Most people only last like a couple of months. You know, obviously due to personal reasons, they get caught, you know, they don't want to just like laziness and like content creation. They don't want to do this anymore. So like, it's actually a huge risk for like an OnlyFans agency to onboard a model. And that, you know, Allocate, you know, attention and capital and building up the traffic sources, et cetera. Right. And so we can use AI to essentially replace his biggest risk factor, not only replace them, but actually capture that piece of the revenue pie, which is about 35%. Right. So we control the 20 percent from the platform fees, as well as the fees that the model usually pockets herself, as well as the chatting agencies that, you know, normally take 10 or 15 percent from the conversations. This is what I mean by vertical integration, right? So it opens up kind of the, the revenue sources to a lot more channels versus the traditional business model only fans who controls the platform itself, right? Because the actual product in this business at the end of the day is the girl or the person that you're engaging with. The platform itself is simply a means to an end.

Lawrence: 12:01

I see.

Dan: 12:03

I'm curious how you plan on maybe Are engaging with models, like, I guess, as far as the mechanics or the flow. So say there's a model who wants to I guess maybe have a lower fee cut and also just sell her likeness to this, this model for, for image generation, as well as text generation. How would the engagement work and how can the models, I guess, trust the platform to be like acting honestly on their behalf.

Eunn: 12:36

Correct. So there's, there's, I guess, two ways To go about, you know, monetizing AI models, right? You, I guess you discussed kind of a lot of it, which I was going to originally talk about, but I guess I'll quickly cover the first, right? There's two avenues you can go down towards. You can use purely fictional like people, like they're hyper realistic. You like, they look like real people, but like they don't exist in real life. Right. That Jenny is basically a pedime of that, right? We're basically monetizing people who don't exist in real life, right? From that, there's pros and cons to each, right? What you described is basically cloning a real human being and monetizing their AI likeness, right? The pros and cons of each approach are this, for example, in the former where you're monetizing a real person or sorry, you're monetizing a purely fictional person. You don't have, for example, the legal risk or like the, you know, contract obligations or whatnot, as we know what you can and cannot do, right? With her likeness, right? So it's like you own her, basically, you basically own this person. You can do whatever you want with it. The downside is it's because she's not a real person. It takes time to, you know, build up her social media following her audience. So the, the path to monetization might be slower, but it's built on top of a more seller foundation. Now, what the latter is they, these girls generally probably already have some kind of social media, probably following on Instagram and TikTok. Right. So the monetization is, you know, faster. But again, you know, you're, you're, you're, you're, you're. If you decide to go the route of, you know, cloning real people, like you said, there, there might be some things they're not willing to do. Some content they're not willing to create. So all of this has to, you know, go the route of, you know, traditional kind of contracts writing kind of getting like a legal entity set up and writing contracts with these girls for, you know, maybe I don't want to make like hardcore pornography. Maybe I just want to, you know, show bikini or lingerie pictures. Right. But so there's pros and cons each. Right. And we can do where we're basically set up to do both.

Lawrence: 14:31

Got it. Okay. So quick question. So like, if, are you guys trying to automate, do you want to involve no real people at all? Or do you want this to be all artificial intelligence or is there a plan to combine, like if real people want to get involved in this somehow, like let's say for example, you're a model and only fans and you're doing quite well, but you don't really want to continue to make content in the current form. Or maybe you're just, it's. Damaging your, you know, your image of yourself or something, or it's just like, you don't want to do it anymore. Is there a way that you can integrate with this platform to say, I just want to use my body or my face or some part of myself and have the rest be automated so that I You know, I'm not fully exposing, you know, my whole life.

Eunn: 15:17

Is there. Yeah. So, I mean, that's what, that's what Dan was talking about, right? Like cloning a real person. Yeah. So if you're going to go that route, which I said we are set up to do like there needs to be like the best case scenario is where like the girls be like, yeah, just like, you can use my likeness for like, whatever, like if you want any hardcore pornography, like whatever you want just, you can do whatever. That's like the easiest person on board, right? Because you don't have to like, you know, do any like, Oh, like. Train the model anyway, or like, you know, right. Kind of legal descriptions of like, what I can, I cannot do. I think for, from a business perspective. It makes sense for us just from a speed of execution perspective. It makes sense to monetize purely fictional people or like people that do not exist in the beginning, right? Sure. It just the legal issues that will probably drag us down the time time that takes draft of these contracts maybe I think when we get, you know, two or three purely virtual like quite like Fictional, like people online operational, and then we can kind of show these real human beings that these are generating real money, right? Then we can on board. We'll start with the extremes of onboarding, you know, a couple that are open to essentially doing everything, right? Maybe these are like existing kind of, I guess, porn, sort of whatever kind of those girls are just willing to do everything right? And then the less kind of like, These legal kind of barriers, I think the better in the beginning, because I think speed of execution is very important. And I guess that's, that's how I envisioned kind of our go to market strategy.

Lawrence: 16:45

Hmm. Okay. Gotcha. Yeah, no, I totally get like using just purely AI models at first. Yeah. I think just to get the popularity of the platform makes a lot of sense.

Dan: 16:57

Let's see. Dan, you have any other questions? Yeah, I think I have a few more.

Lawrence: 17:02

I have more. I just want to give you a

Dan: 17:04

chance to speak. Right, right. Yeah. So I guess, so you said that the team's about like three at the moment and like,

Eunn: 17:11

well, three main people, but we have like, for example, under, under our CTO guy, there's like two engineers working under them, working under him. I just, I didn't kind of go into like specifics or like more people, but I just gave you like a general overview, like your three main guys involved.

Dan: 17:24

Okay, cool. Yeah, I mean, I think I guess there's kind of maybe like a feedback point from my end. Um, just reading through the docs before the call. It's relatively sparse, I would say, on the technical front, and it. It kind of goes into more of like the societal, I guess, implications. I would say it's like maybe decently red pilled as far as like, why OnlyFans like marketplaces exist. I think like that personally is, is already like very like apparent, like the, you know, the sex market and porn, like that's so widespread already. So I think that like, you know, kind of already makes sense, but yeah, I think having more information on like the tech side, as far as like where, where does the blockchain element. Come in and what is the utility of the token like specifically like can users pay or use the token to like pay for access? And then, yeah, how I guess our smart contracts involved. I don't think that was super apparent to me. Yeah, I think that would just be like good to include in the documentation.

Eunn: 18:31

Sure. So I guess well, we do, I guess have a brief section, maybe, you know, I guess we didn't go into too much detail about it, but I guess I can discuss that now. So in terms of the blockchain component so for this web three launch, right, specifically for Lush chat, we are going to deploy smart payment, smart contract, you know, smart contracts on both Ethereum and Solana. Right. So you'll be able to, you know, pay using Solana, ETH whatever, which will be converted into their dollar values and, you know, into like a credit type of system where you can, you know, pay on this platform. Right. But in, in this type of business, the vast, like I said, the initial launch is just essentially to basically establish our name as like, as like the market leader in this industry, right. For the weather community, when you actually get this type of business up and running, the vast majority of payments are going to come through credit card payments. Right. And so we are setting up the platform you know, offshore entities. Right now to essentially accept credit card payments and, you know, the payment processors and then bridge those, you know, bridge the USD on chain into like a stable coin or whatnot. From there, you can do things to, you know. Actually like drive value to the token holders. We have, whether it be, you know, through some kind of like staking and revenue share or et cetera. Right. And so the token itself, I think it doesn't make sense. It's a poor user experience to force people to transact using your, you know, your native token, especially if your native token is like subject to volatility. Right. So like, and the, the token itself doesn't need to actually be a medium exchange for that value. Right. And so we're just going to accept whatever, you know, whatever the payments in a platform in stable kind of currency, whether it be stable coins or actually U. S. dollars real credit cards and the bridge, those proceeds on chain in the case of credit cards to drive value to our token holders. So in the case of lush chat, it's like a hybrid type of web to web three business.

Dan: 20:13

Okay, so what exactly is the Lush token used for then if you're planning on setting up the legal entity to accept these credit card payments? Why exactly would users hold

Eunn: 20:23

Lush? Well, cause for in the case, in the context of Lush chat, it is like, like I said, the means of like driving value to token holders, right? So like whatever revenue the application makes, whether it be from WebCryptoPayments or Web2Payments. Which is be bridged on chain and through like some kind of staking contract or whatever the revenue would be like Distributed to token holders, right similar to how it's done with even like, you know, TG casino roll bits Even uni swap, you know, just it's about to kind of like start there, you know fee sharing revenue mechanism a revenue mechanism, right? It's like on chain revenue distribution It's actually It's not like a new concept. Other protocols are doing it. What we're doing, I guess, is that kind of just bridging the Web 2 and Web 3 worlds, because in this business, the vast majority of payments, like I said, are going to come from like credit card payments. Right. But that's Lush chat phase one. I don't know if you've got to take a look at through the entire documentation, but like phase two is like the Lush network. Which is like a blockchain based marketplace for AI models, right? So like phase one lush chat. We are internally monetizing our technology and our models in phase two. The lush network is where we open up our like infra tech infrastructure to the general public so that anyone can essentially, if you're AI artist, deploy your models to the network and monetize them by allowing any day everyday users to, you know, ping the network to generate content. Right. For instance, if you are an owner of like a mom and pop clothing store, you need to hire a model to, you know, whatever, create some content to display your products instead of, you know, signing a contract with a social media influencer, flying out to a photo shoot and whatever, which costs. Oftentimes cost like 10, 20, 50 grand, right? Just ping the blockchain network, use a model. Let's say I want this model, Stephanie, wherever I want Stephanie wearing the shirt, standing outside on a beach, et cetera, right? In that context, you'd be using our GATT or our token as like the GATT native gas token of the blockchain network in order to generate content. So there's a two lines of businesses to lush and we cover both in In the appropriate section in our documentation.

Dan: 22:28

Yeah, I just want to drill into that just a little bit more, just so I'm very clear. So there is basically the token at the moment is not needed for lush chat. And it likely wouldn't be it would be credit card payment based. So there's going to be like USD separate.

Eunn: 22:50

Correct. It's not, I would say, well, just to quickly correct it, I would say it's not used as a medium of exchange specifically when on the platform. That doesn't mean that it does not have a value when it comes to like value accrual, right? Does that make sense? Because the token can have value in like driving value to whoever holds the token. It doesn't need to be used as a medium exchange. For example, like let's Uniswap is about to turn on fee accrual mechanisms for their Uniswap token, right? Like you don't need the Uniswap token in trade. That doesn't mean that Uniswap token cannot actually have value. If they have some kind of on chain mechanism to distribute

Dan: 23:29

revenue, does that make sense? I mean, I guess Uniswap is for governance, right?

Eunn: 23:34

Yeah. Well, it's a governance token, but they just like announced, they got approval from their lawyers to announce this, like definitely, but the, the fees that, you know, they generated from the protocol before it was not distributed to token holders. But now they're like about to turn

Lawrence: 23:46

it on. Right. So, so the goal is then to use Blush token as a, you know, to share the revenue with users, like you said, to share the profits of the platform. Yeah. So as the, as the platform gets more usage, it makes more money. The token, you know, theoretically should

Dan: 24:01

accrue that

Eunn: 24:01

value. This isn't like a novel, I guess, like innovation. Many other protocols, especially in DeFi also kind of utilize this kind of thing. On chain, you know, revenue distribution.

Lawrence: 24:14

So so, so like, let's say, you know, you have this platform, I'm chatting with a model, et cetera. So is it, would it be the same similar type of, like, would it be a subscription type thing? Like only fans has right now where, you know, I subscribed 20 a month to a model and I get access to chat with her and see the videos, et cetera.

Eunn: 24:32

Okay. Chris. So. Only if you guys business model is actually relatively straightforward, right? It's just like a subscription based payments and one time payments, gifts, pay per view content, right? So yeah, our, our platform, I think something similar sufficient monthly subscription based, you know, fee you, whatever, like three 99, four 99, whatever the price is. It could also be free, right? So it's just substitution based payments as well as the ability to pay for pay per view content. As well as, excuse me, as well as the, you know, gift, the model centered gifts and whatever. Right. So that's like the, the, the revenue kind of model of the lush chat.

Lawrence: 25:08

I see. So I think, I think one of the cool things about this is. Like only fans right now, you know, okay. I pay a subscription. I see a model, whatever. I don't do this in real life, by the way, just, you know, but I'm just saying theoretically, if I was to, yeah. You know, it's, it's very hard for them to generate like individual relationships with each people. I mean, if I have, if I have a thousand subscribers, I'm not going to remember details about each person, you know, I, you know, so I think one thing that's cool about this is like, you can, you You might actually be able to do that, right? You might actually be able to create, you know, like, you know, if you think about the context, you know, in conversations with the models, the AI models, you know, and you know, yeah.

Eunn: 25:53

How it works is that like, when you're, we also discuss this in our documentation as well, like when you're messaging, when you're talking to regular OnlyFans model, OnlyFans, you're not actually talking to her. You're talking to like a Filipino VA that she's outsourced her chatting to. Oh, I see. Yeah. You're not actually talking to the girl. So VA's have like these kind of scripts that they go through to like try to actually extract gifts from you. So like, she probably doesn't even know you exist to be honest but that's just the nature of the business. Right. Because like you said, there's no way one girl can keep a conversation with a thousand people. Right. And so like, I think we believe that, you know, AI from a conversational perspective can provide, I know, a more fulfilling conversation than these Filipino VA's. And also just how that conversational memory built into the AI, right? On top of that, the AI based on kind of the chatting data, the platform collects, it can be used to recursively train the LLM to essentially like be better at whatever, like even asking for gifts, right? Like, or to maximize track land on revenue. So those are kind of very interesting, you know, that you can do. And I think this, this goes back to, I think one of the key kind of tenants that I always like preach. I think for any. AI business to actually have a defendable moat. They either have need to have control. Over the, I think their training data and or the proprietary AI models, right? So in our context you know, we have control over, you know, the chatting data, which can be used to, like I said, recursively train, the LLM and also all of our kind of girls, the models that are responsible. Our girls are like built and trained in-house, right? This is directly in, you know, antagonistic to, you know. What you consider like a traditional, like front end, like chat, GPT wrapper, right? Those types of businesses can be wiped out overnight. If chat GP just, you know, decided to launch their own chat or own application. Right. So in context of like defensibility in the age of AI I think data and the models ownership of those are very important.

Lawrence: 27:50

Yeah. A hundred percent. I, and I share the sentiment because I actually work for an AI startup myself and, you know, there's always that fear of like. just releasing a product and wiping us out overnight. Yeah. And it's really about the platform you build on top of it and how, you know, the experience of that is and, and also, yeah, like your proprietary data that you control and train on, et cetera, for sure. Yeah.

Dan: 28:15

Yeah. I mean, I think the, the, the, it's the, the language element with these models is like also I mean, very interesting because these companies have, I mean, I think GPT 4 is still like the best. There's other ones coming up, but those are still kind of relatively proprietary and very limited in scope. And there's like bias in the people who train them. So I guess is the LLM like basically fully custom? I know, I know like was it Mixtral or Mistral? That's like kind of an open source. Fully customizable unfiltered AI option. I'm curious if you guys are able to share like, yeah, just the model. Yeah, that's a good

Eunn: 28:59

question. So we are this is actually one of the projects that I also invested in personally, like when I was kind of doing early stage on chain investments, right? When they were about like three or 4 million market cap, they're training by like a hundred now, but they're actually a partner of our Of our project, they're called any QA. I, so they're building essentially the narrative of them is like, they're building like the uncensored chat GPT. Right? So there are kind of like technology partner, and they will be providing the LLM to drive the conversational ability of our AI models. And then ideally, we want to, as, like I said, as, as we collect more and more chatting data on our platform, we want to use that to kind of recursively train the models to get better. LLM to get better at better on conversation. So this is, yes, this is a custom built solution without the you know, issues of sensor sensor ability and, you know, all those issues that come from, you know, chat GPT and the publicly available, I guess, like, you know, Models that are provided by, you know, these web two

Dan: 29:54

companies, right? So that's N A Q A I E N Q A I E N Q

Lawrence: 30:00

E N Q. Yeah. I've actually been watching this one and it's pumped insanely over the last like six months or so. Yeah. So they're a

Eunn: 30:06

technology partner. They're like, I'm pretty, I'm pretty close with their founders. Cause like, they're the ones that, you know, gave I guess the, The idea of like, you know, building a crypto project around Lush our first kind of part of our launch was we, we airdropped 5 percent of our supply to their token holders. So as part of like the, the, the process of bootstrapping community and you can like read about that, I guess on Twitter. So yeah, we're, we're kind of heavily involved with each other.

Lawrence: 30:31

Yep. Cool. Anyone else you guys are partnering with or is it just, is it just primarily them right now?

Eunn: 30:36

Just for now, I think yeah, I, I, I think, I think when, when it comes to any partnership or any kind of like, what a partnership, I guess, I think there needs to be like actually make sense where I don't want to just partner with projects to put out like a PR to like say a partner or something. Right. So I think from a technological standpoint, they're, they're the only one that makes sense so far. Sure.

Lawrence: 30:55

Yeah, I think also like where this project is interesting is, you know, It's actually a good story for like helping fight human trafficking. If you think about it, like I actually was, because let's say there's someone, you know, I'm, I, I, a girl wants to do this. They want to like a real person wants to be an only fans model or wants to do some form of modeling, but they don't want to fully commit. Like you have to right now, which is involves like basically. Being nude on camera and performing acts that you may not want to do. You know, this platform may give you a way to do that. You know, maybe you don't make as much money, but you're able to do that on the side. And, and not fully expose yourself. Like, you know, so I think like this, this could actually have a lot of, it's like a good story from that viewpoint too. And if there's a way that that marketing could take, could take place, like, I think that you guys would really win the hearts and

Eunn: 31:52

minds of people. I think, I think that's a very good point. To be honest, like this, this industry as a whole, like. It's very, it's like, it's brutal, man. Cause like, I mean, it's, it's a very brutal industry. Like, I mean, I don't think it's something that like, I think it's a net positive for society. If AI can essentially displace these, you know, these, these girls from, you know, doing this as a living, right. Perhaps, you know, perhaps the future is they might still be able to monetize off their sexuality. But, you know, they clone the AI, the AI just, you know, doing the background for them. Right. Cause like, I think society better off if we can one like I'll compete these like human girls that actually gives them more incentive, you know, to like reallocate their labor and attention to more productive kind of occupations in society. Right. As opposed to, you know, taking selfies and nudes, like as, as like a full time occupation. Right. So I think, yeah, I think I'm on point with you there. Also posted a link of our model, by the way, if you guys want to check it out. Sure. Yeah,

Lawrence: 32:54

yeah. I saw that. All right. So I want to go a little more into this. So you have the phase one, which is the chat, right? Yes. That's coming out soon. And then you have the phase two, which is kind of like the decentralized marketplace for the models, right? So like, let's say you know, I have a product and I want to market it. Everyone, I think we all know that sex sells and having models sell your product is probably going to, you know, include increased traffic driven to it. Yes. Right. We all see this at conferences, right? Like, you know, we, I was just at East Denver and the booths with the pretty girls had the most people at them. Of course. Yes. This is like basic human stuff here. So. Like, is that kind of the dream there is to, is to create a platform where like people can rent them out, you know, these AI models and put them in like social media ads and stuff like

Eunn: 33:38

you were saying, I'm glad you mentioned kind of the, the sexuality component is that it's, that's like, that's actually very important because only fans is simply the first vertical that we're targeting, right? This, this technology can be, you know Apply to across all types of sectors or many more sectors. And the actual, the way I think about it is that the product itself that we're selling is digitalized, like sexuality. You either like, it doesn't matter if you're an Instagram girl, high end fashion model, porn star, wherever the, the, the product that you're selling is essentially the same. It's just digitalized sexuality. You can either monetize that directly by selling content via a platform or medium, such as Onlyfans. But the actual larger market for digitalized sexuality is indirect monetization. What that means is that even the advertising industry as a whole, right? Like it's simply monetizing sexuality. Like even like, for example, like, you know, the, the yoga pants, right? The model who's, you know, you know, showing off the yoga pants on Instagram, right? The, the game streamer, like the Twitch streamer. Who's wearing provocative clothing. You know, do you really think the guys are tuning in, you know, to watch or play Fortnite? No, it's just, they want to, you know, just like watch her in skimpy like clothes. Right. But the point is like, OnlyFans is simply the first vertical that we wanted to like, right. Especially when you're starting on a business where you want to focus on one niche, dominate that, expand from there. Right. So we want to kind of focus on dominating the, the AI OnlyFans vertical. And then once we kind of establish our name as like the, the industry leader in the space. With there's, there's kind of a much larger market hit play, which is like the trillion dollar marketing social media marketing, digital marketing industry. Right now you're talking about Instagram endorsement deals, right? There's no reason why an AI model can't take that Jim shark affiliate, you know, link or affiliate sponsorship that a human girl is taking currently on Instagram, right? Even when, when the technology is not there yet, but in maybe it's 12 months. There's no reason why like a real time generative, you know, Twitch streamer can't take the place of that girl who's, you know, essentially like semi nude, you know, whatever in cosplay clothes, you know, on Twitch talking to guys, right? So there's much many, there's many, many more kind of. Use cases that this technology can apply to only fans is simply, I guess, the easiest to understand kind of industry to start with. And that's why we chose it.

Lawrence: 36:06

Right. Yup.

Dan: 36:09

Yeah, for sure. Yeah. I mean, I'll, I'll take maybe like a bit of a counterpoint here. I think inevitably to have a really good platform, you're going to need to work with like real people essentially. And probably like mostly real girls. Cause I don't think, I think all the models. At least that I've seen on like Civet AI, it's all kind of based on a person. And when I was messing around with Stable Diffusion for purposes like related to this, a lot of images are just like get very distorted and they're, they're not as refined. And yeah, I think it's. It's also video, I think is like a key component of this. So I think, yeah, I mean, I think the tech is, is kind of early on the video front. And I guess, yeah, open AI has Sora, but I also think the infrastructure is going to be like very difficult to manage because there's like token limits. At least on like the predominant platforms right now. And I think like. Yeah, I think the cost definitely will scale because certainly people are might message more than like 20 messages and like Three hour period or something. So I think like that that is a definite risk as well So I think there's like several interesting situations here. And I personally believe that you may want to appeal more to the models and like, how can you serve them? Because at this point, like they are entrepreneurs, essentially. I mean, they're basically running businesses and they have to basically decide where to allocate their time and their money. And they actually do wield a lot of power because they have massive followings. So I think they, they need to be catered to maybe more than you think, in my opinion, I think like they, they need to have a, have a marketplace that's competing that hat that takes lower fees from them and allows them to like, autonomously sell products and things like that. That's my personal belief at right at this moment right now. I think those models still will play a critical element in this like AI agent future. Yeah.

Eunn: 38:22

So I think it's a, I think it's, it's good. You mentioned civil AI, cause it seems like, you know, you have at least like a, you know, decent understanding of like, you know, the AI in general, cause you said you played around the stable diffusion, right? I don't know, like to what extent, but I guess, So I guess the, the good way to think about the Lush network is essentially like a blockchain based version of Sibit AI. The, for example, like the, the process like you generate content again, I don't know how, you know, in depth you got into like, you know, actually, you know, prompt engineering as they call it, but it's actually very complex to, like you said, create consistent, you know, People without, you know, distorted limbs across all types of environments and, and, you know, whatnot. Right. So the whole point of lush network is to not provide, not only just provide actually the marketplace for, for this to house these AI models, but also to essentially abstract away the complexities of prompt engineering. Right. And so, you know, so that, you know, the average person can, you know, without going too much in depth into actually studying, you know, how to prompt stable division, be able to generate somewhat you know, consistent content, whatever. Right. So that's one thing, I guess, I guess to answer your question of the importance of, of humans it's, it's, it's interesting. It's it's an interesting question. Whether how important role existing human influencers will will have in the role of A. I think we are at a current. I think the technology today is not perfected to a point where a fully virtual person can be 100 percent indistinguishable from the real person, especially on the on the, you know, Aspect of like video creation, right? Most of the high quality video that we create uses some form of like AI trained like deep fake instead of purely generative, right? So I guess I think humans will play a role like, I guess in the foreseeable future, maybe like, you know, for the next two or three years. But I would say that the long, in the long term, our thesis is that. Purely AI, AI influencers and models will eventually replace humans. Not because like, you know, they're like they're like better or whatever. It just simply because as the technology evolves exponentially. The A. I. People become more and more indistinguishable from real people so that you won't even be able to tell who's a human, who's an A. I. Online. Right? And like the monetization route in that route. Yes, it takes time to build the following whatever. So I think from from like a monetization route. It might make sense to, you know, for example, we partner with humans, you know, in the short to near term, because like you said, they already have that, you know, audience with them, right? But I think in the long term, I'm talking about like three to five years down the line, and it makes sense to start working on it right now is to, you know, start building up the following and the social media presence of these AI girls, right? Because that takes time to like harvest the yields of those efforts. So I think there's there's two approaches. And that doesn't necessarily mean that we can't, you know, target both in parallel. Right.

Lawrence: 41:21

No, I think it, that, that does make sense. And I think the chat's a good way to start because you don't, I mean, you don't need an insane amount of compute for that. Right. But yeah, I think longterm, yeah, I think, you know, to Dan's point, like it's going to take a lot of money and, you know, compute to, to do a lot of like, especially video, you know, and then that is so new but it's going to get there eventually. So I think you guys are in a good place to, if you build this now, you know, over the next ensuing years, you know, when the technology gets better. Yeah. When be

Eunn: 41:53

able to capitalize on when, when, like I said, when last chat goes live and you can even take a couple of like even on the Twitter that I showed you of our AI model, you can take a look at, I guess the, the video examples that we have right now. And you can, you know, judge the quality there. Right. But I think even in those instances, like I said, we're. Those aren't purely generative. They're like AI, we're training a model based off like our AI girl, right? We're using AI to train the AI essentially. And then we're going to, we're going to essentially like use that as like an S like AI assisted defect to create like video. That's, that's the best approach with, you know, today's technology. But even using that approach, it's like the results are still quite good. I think.

Lawrence: 42:33

Yeah. Another question I had was like the influence, right? So, like, let's say, you know, this Jenny model, she has a Twitter account. She has. You know, social media accounts. How, how, how is that being promoted? How, you know, are you generating the Twitter posts yourself or do you have a model doing that? How is she gaining

Eunn: 42:51

followers? Yeah. So right now the primary kind of go to market strategy of our model, Jenny is to basically like drive attention to our, like our crypto project, right? That is in kind of contrast to the actual marketing strategy of driving fans to our platform. Because When it comes to that, like that you're talking about, like traditional kind of Instagram and TikTok marketing. Jenny's purpose right now is essentially like, we call it like Operation TTT, what that stands for is like Twitter thirst trap. It's like an inside joke where you're like, you have this like girl kind of flirting with these like crypto bros online. And then they, you know, they like use through that funnel that, you know, get to find our project. Right. But to answer your question, the the content generation, like the, the content that you see on a page is generated, like, is, is like from our technology, right? The actual kind of engagement we've tried, like, quote, unquote, AI assisted, you know, Twitter engagement, bots, whatever. We just found that nothing beats like the high quality engagement of like a human being. Just like managing these accounts. So I mean, like, but even in this instance, right, we have like a VA that's managing like five of these accounts for like, you know, like 5 an hour. So from a cost based perspective it's still, you know, worth it to, you know, run it using a human. Sure.

Dan: 44:08

Yeah. I think maybe like the last question for me, I'm just curious. So like on the token page when, when did the vesting schedule

Eunn: 44:17

start? The vesting schedule started. TGE, which is like January 22nd. So that was like two months ago. So yeah, to discuss Yeah, so we like the token went live two months ago. So we're still kind of relatively new. We have, for example, regarding the tokenomics, we have tokens allocated for future VC rounds, which we do plan to raise like a more official round during Q2. Up to this point, it's just been funded through like the, the trading tax. So we have like, even, but even with that, we have about like, I think 700 K at this point. So that can fund us for like the foreseeable future. We also have tokens in Q2, one of our plans to also see the trading market on Solana because that's obviously where like most of the retail attention is. So we're just going to, we have tokens allocated for that. We're going to bridge some over onto Solana and I'll see the trading pool there. Maybe in the future we're like, I don't know, see to another trading pool on base or whatever. Right. So we, we want our token to be kind of cross chain to tap into as many kind of. Network I guess blockchain network community as possible.

Lawrence: 45:16

Okay,

Dan: 45:19

cool Anything else you wanted to say? Yun from your side on behalf of lush

Eunn: 45:32

Yeah, I would say that I guess just keep out, keep an eye out for us. The, the next kind of forward looking event is the lush chat. Like I said, we're just going live early access in about a week and then public access a few weeks later. So I guess that will be a great way, I guess, for, for any listener, any participant, you know test out our, our platform itself as well as, you know, to, I guess, Get a good kind of glimpse of like the, the quality of our like AI content. Right. And I guess, yeah, that's, I guess a good way to, I guess, you know, scope out, you know, the quality of our product and, you know, judge for yourself if you see there's like, you know, there's like a value proposition there, right.

Dan: 46:15

Okay. And then I guess users can contact you through discord primarily.

Eunn: 46:19

I mean, I'm, I'm pretty responsive on all of our platforms, Twitter, discord, Telegram is probably the best way to do it actually. And I mean, yeah, like I said, I'll, I'll like follow up with you guys after the podcast, give you guys early access, you know, test out the product. And if you guys like it, you know, you're more than welcome to, you know, share a referral link or whatever to, to your listeners and your audience, you know, to, you know, get them, you know, to test it out as well.

Lawrence: 46:46

Nice. Cool. Well, you, and thank you for, for joining the podcast. This was a very interesting project. I'm very you know, I think it's a cool kind of intersection between AI and crypto happening right now. So, you know you know, thanks for coming on and talking about this. Yeah.

Eunn: 47:01

Thank you guys for, for having me. Thank you. You're welcome. All right. Take care.

Lawrence: 47:07

Yep. Take care.

Dan: 47:11

Alrighty. Cool. So what are you thinking

Lawrence: 47:14

on that?

Dan: 47:15

Yeah, obviously this is all not financial advice. You know, we're trying to basically, yeah, investigate anyone who's building encrypt there right now. And there's a lot, I'd say there's a lot of projects that kind of go very. I'd say like Hyper DGen route where they have the Gitbook and it's like very kind of philosophical documentation and yeah, I, I personally feel like I, I don't know if I can like trust this platform fully to like put my funds into it. Yeah, I'm just like a little sus on just, it's very limited in tech. Yeah.

Lawrence: 47:56

I also think they're going to have, I mean, they need a lot of money for this. I mean, ultimately, I mean, if, and I think it's going to be hard for them to raise, like what investor is going to put their name on this and have it be marketed like a big, you know, it's like, Oh, I invested in lush, you know, creating AI pornography essentially, or, you know, AI models. I don't know. It's like kind of hard to. Do that. And then also make other serious investments on, you know, too. So,

Dan: 48:22

yeah, I'm also like unconvinced. I mean, like the images that he claims is like from their technology, it looks like they could have just been directly ripped from like any old. Model. Like when you go, I mean, it's, it looks very kind of similar, I guess, of that caliber. So it's, it's kind

Lawrence: 48:40

of I mean, they could also be real people. Like, how do you know they're, they're not just like Googling something and claiming that it's there, you know, they're a model. You know, you know, they could just be Googling someone copy. Yeah. I don't know.

Dan: 48:52

Yeah. I mean yeah, it's kind of like a hard, it's striking a hard balance. I think, unfortunately with crypto, there is such a difficult regulatory environment where like, it would be great if people could build out in the open and not have to worry too much. I mean, there's, there's definitely elements of certain crypto projects that expose people to rugging elements. And, you know, sometimes people are just trying to build in earnest and like, they're trying to just get funding so they can build, but it's hard to mechanically separate projects that are going to rug people versus like very honest actors. So yeah, my, my personal opinion, I, I like. I guess a little concerned about this project. So, yeah, it

Lawrence: 49:41

seems, I agree. It's a little bit, I just think it's, it's very hard to do this. I think the chat could be interesting. Like, because you, you can definitely train them. You can definitely have a model that takes context of previous conversations and treat it like you're, it sounds like you're talking to a real person, but that's, that has no real value in a token. Right. I mean, that could just be part of only fans or something

Dan: 50:07

like, yeah, I think that was also what was concerning is like. I didn't really, I didn't really fully understand like the token model. And I think ultimately not being close to the creators and like understanding like the, what are the creators doing? I think it's even in like a fully AI context where you're maybe trying to replace the creators. It's still good to know, I don't know, both sides of the marketplace. Like what, right. What it is that they need or they want. And yeah, I mean, I'm seeing like some other. Like when I just searched like decentralized AI or decentralized only fans, like only one dot app shows up and yeah, I mean, I could see like moving gradually into web three with like. NFT models where you're like, you know, giving special access into the, that offers access into the broader Web3 ecosystem, like, you know, rather than being super gated with only fans. I think that's like an interesting value proposition. And then obviously there's the AI element, so you can like offload the work. So I think there is a marketplace in this realm. But I guess it just remains to be seen if, if like Lush is going to be able to execute and garner the funding and infrastructure to do all this. Right. Yeah.

Lawrence: 51:29

That's, that's the thing. I think funding is going to be tough for them to get, to be honest. Yeah. I just don't see how they're like. Like, you know I don't know, electric capital, isn't going to invest in something like this, like input their name on, you know, it's like who, or I don't know, they're going to have to work hard to sell to investors. And I mean, but you know, it's risky, like it's not a high valuation really. We're still early in the market. It could be, it could be something. But yeah, there's tons of risks.

Dan: 52:03

Yeah. And I have seen a non projects like get funding. I mean, it does happen. And I, I do think, yeah, they, they will put in maybe less than, you know, they would, if, if, You were, you're D and non equivalent, but yeah, it's definitely happened. And yeah, unfortunately rugging like can still happen. I'm. I still, I don't think being a non like insulates you from legal consequences. Like there are ways. No, it doesn't. Find you basically. Like if you screw up and people like they're, I don't care how a non you think you are, there are so many methods and tools at the disposal of the legal system that I, you know. Yeah. Eventually you will.

Lawrence: 52:52

You're going to get found by someone, man. Someone's going to rat you out. It's only a matter of time. It, the thing is like, I mean, it's ultimately if. If a government wants to find you, they will, they'll, they'll leverage a real person to, to rat you out. Yeah. And it tends to, that's tends to be how it works. I always get a little skeptical of anonymous teams who don't want to be docs because I'm like, what's, what's the reason why there's got to be either. If you really believed in what you were working on and you really like, why wouldn't you want everyone to know who you are building it,

Dan: 53:28

right? Yeah, exactly. I mean, yeah, yeah. Even bear chain, like. I think Beartain might be in a class of its own where, like, it did garner so much attention and, like, insane amount of fundraising. They are anonymous, but, like, You, you can still see them and everything. And that, that was actually another question I wanted to ask is like, has he met his team? Like, does he know who his team members are? I think that's like also interesting to know, just like, you know, and are you going out in public and like meeting people still? Like I did see the face of someone at bear chain at East Denver. You know, I didn't take a picture or anything, but At least that team is like out there and take a picture. I

Lawrence: 54:11

found them, you know, like docs and shit. Yeah. Yeah. Starting all drama. Yeah, there, I, I do think that their partnership with NQAI is interesting though. And to be honest, I, I would like to bring them on if I could try to like get his, cause he says he knows the founders. So I feel like maybe I could ask him, like, can we get a founder on here? I think that, cause that project is, I mean, I haven't done deep research, but I need to be, because the token has done really well. Cool. Lately, so

Dan: 54:37

yeah, I like, I like what I'm seeing or what I've at least heard. Very yeah, that was my first time hearing of that. And Q and yeah, I mean, it sounds like they're almost like an infrastructure layer for basically any sort of AI chat marketplace. So that, yeah, I think could, could be interesting as well to, to see what they're up to and yeah, it seems somewhat like a similar vibe where it's anonymous, their docs are maybe a little bit more filled out. You have a white paper at least. And it seems like a little higher budget. So they, they might be. Yeah, the doing things as well. So,

Lawrence: 55:21

yep. So in short, in conclusion this isn't financial advice. You know, you, you could lose all your money do your research. I personally wouldn't put too much into this. You know, if you want to speculate, don't throw your whole portfolio into this or even a large amount, I would say, but you know, just that, you know, You know, it's, it's, it is, it is an interesting idea and, but whether they can execute is the real question here and whether they can raise money for it.

Dan: 55:48

Yeah. Yeah. At the very least, I think we're, it's good to shine a light on things on maybe the darker side of crypto and just seeing what's, what people are doing and like allowing people to judge for themselves for more protective POV. So,

Lawrence: 56:05

all right. I think that's all for today. Right.

Dan: 56:07

Yeah, I think so. Yeah. Thanks to everyone for watching and we will catch y'all next time. Yep. Take care of everyone. Peace.