#495 Next-Gen Gaming: AI Agents, NFTs, and the Blockchain Frontier

[00:00:00] 

Mehmet: Hello and welcome back to any episode of the CT O Show at Meac today. I'm very pleased. Joining me, Greg Marlin, CTO of ZKCandy. Greg, the way I love to do it is I keep it to my guests to introduce themselves. So tell me a little bit [00:01:00] more about you, your background, and. But you're currently up to just as a, you know, kind of a teaser for the audience.

We're gonna talk a lot about gaming today, AI automation and Web3 NFT. So Greg, I don't want to take much from your, uh, time. The floor is you. So, um, and thank you again for being with me here today. 

Greg: Thank you. Um, yeah, it's great to be here and, uh, thank you for the invitation. Um, so I'm, I'm a longtime technologist.

Uh, so. I've, uh, founded companies in SaaS, uh, in crypto, um, and, um, ai, uh, as well. And currently I'm focused on, uh, being CTO of Zeki Candy, as, as you mentioned. So that's, uh, a, uh, layer two, uh, blockchain network, uh, where we obviously, we, we run the chain. Uh, it's a layer two built with the ZK sync ZK stack, um, to on, on, uh, Ethereum as the layer one.

And we also [00:02:00] built, uh, you know, uh, a kind of unique, uh, experience for both developers and gamers, um, to, you know, really experience a, um. A very seamless, um, onboarding and, and, and, you know, fun experience that, uh, leverages the best of what, what we can, uh, with, with blockchain. Um, in my spare time as well, I also, uh, you know, I founded a ceo.ai, which is kind of my, uh, personal AI venture.

Um, and I use it daily for, in my own work. So that, that, that helps me. I sort of have built it in stealth for the last year and a half. Um, and I also, uh, I found that I still maintain idex o, which is a, uh, was kind of the first like low, uh, low code SDK, uh, for, uh, Web3 and NFTs. Um, and it, uh, it actually, we've done some, uh, very new stuff for, uh, integration with zk Kendi.

Uh, and we have, uh, [00:03:00] actually a new client coming up that's, uh, minting 6,000, um, soul bound tokens on Zika candy for, for education. It's really, really kind of an interesting, uh, client, um, and, um, you know, inno, let's say SBT innovation happening on Zika candy. So, you know, I'm, I'm pretty busy, busy with all those things.

I think, uh, yeah, we're gonna be talking a lot about gaming and, and ai, so I look, look forward to, to talking about that. 

Mehmet: Great, and thank you again for being here with me today. Uh, Greg. Now let's start, you know, I like to do couple little bit of research before I, I, uh, do the actual recording with my guests.

So two things caught my eyes, so of course, you know. AI agents. I was like wondering what's happening in general, right? And found out that it's a growing market, I mean in general. And then I figured out that you think that the future in gaming experience lies in AI [00:04:00] agent. Now let's break this down. So first like let's you know, just for the audience, for the sake of, you know, just explaining things about what an AI agent is and then let's.

Give them, uh, Greg, if you don't mind an idea how, you know, you foresee this AI agent into the gaming, like, and, you know, feel free to explain it in details. 

Greg: Sure. So, you know, I think there's, um, there's sort of two ways people approach talking about it in the market, right? So, so one thing is, like, for me, like the core agent is, is made up of.

You know, initial prompts, right? So you have your user prompt, you have your system prompt, um, and you, you know, you have the model. Um, and you know, this is like the very basic, simple encapsulation of an agent. Um, you might add to that custom memories. So that is typically, uh, you know, uh, vector embedding.

So you, you have an [00:05:00] association with that agent and depending on how you, you set up the technology behind it, you can have, uh, you know, it, it, it just changes how it will leverage those memories. But you can associate with an agent. A whole set of custom memories separate from the memory of the LLM model that it's using, whether that be like a code from philanthropic or you know, oh one or oh three or whatever from, from uh, chat GPT or Open ai.

So it really depends. And, you know, each model have their strengths. They, they're, they have a, you know, a per perspective. They might have a different memory or training set. They might be better at or worse at different things. And then when you give it custom memory, you're kind of giving it, uh, a better, a better way, a quicker and better way to access those memories.

And quite possibly the only way that it had memories. 'cause most often there's something very specific, such as, let's say. Your code or the, the gaming [00:06:00] code, like how the game works, how the game you just created yesterday works is not going to be, definitely not going to be in the, in the memory set of code or, or, or, mm-hmm.

Or, or chat gt. So, so memories can be important. Um, and then, uh, also when you get into tool use, right? So there's a, there's a whole nother, uh, set of, of, of, of, um. Feature set that you can add to an agent, um, related to what tools it can use. And, and here you're training it. Uh, okay. When you, when I give you this instruction, I want you to output exactly this, and then you connect that to the tool, right?

So it's like, uh, I want you to, uh, send me, send an email. On Gmail, right? So I'm going to prompt you something and then that's gonna, you know, and most often it's in a workflow, so it's actually not you, it's another tool is sending an API call to your agent, your agent's doing a bunch of [00:07:00] stuff and then pushing it out.

So you're teaching it to write a Gmail. So you know, within one little agent can be either just a very simple system prompt and a user prompt. And when you make that user prompt variable, you know, it will give you some answer in a chat flow. This is like probably the most basic, an AI agent can be right to what I just described, where it has its own custom memories and it has its own, uh, you know, tool set that it uses.

But the reality is that, um, one AI agent on its own is even with all its powers, which by the way, the, what I just described has a lot of power. Um, most, and you know, even a year ago I was almost always working in workflows. So you can imagine what I've done at the last year mm-hmm. And the advances that we've had.

So, so even very early on, like a year ago, a lot of the power was how do you combine AI agents together, work them together, create, you know, these, these [00:08:00] flows. And there was a lot of, uh, few startups that were only focused on that. Um, and now the, the, uh. The, the, the trick and why I, why I said it is that some people will say, and I, I, I can, I can see where they, where they go about it.

And, and I think it, it makes sense in some contexts, but it's confusing for users, right? So some people will say, you know what, all those agents working together, I'm not going to tell about it, anybody about it. 'cause I don't want to, you know, confuse them. So I'm going to combine them together. I'm going to make a one simple prompt layer in the front.

Right. And I'm gonna call that an agent. Right. But it's not an agent. It's like a hundred agents working under the covers. Right. But but to the user, it's one agent. Right. So it's, that's always a little tricky, is like what you mean by agent. Do you mean this like full application of multiple agents [00:09:00] working together, or do you really mean, you know, all the way back down to that very simple agent that we talked about with, with system prompt and user prompt.

Right. So it really is, uh, it depends, right? And in the, in the context of, um, gaming, like gaming agents, uh, probably we're talking a little bit like the last one, right? Because. Um, if we are saying, well, we want to give you a companion, right? Who's going to help you in your, your gaming experience? You just, you kind of wanna think about about it, like your little pet that goes along with you in the game, right?

Mm-hmm. Like in a, in a World of Warcraft kind of scenario or whatever, you'll get these little pets and they come along with you. Or, you know, depending on your class of fighter, you might actually, might actually have a, like an animal that, that, that fights with you, you know? So. There, you're going to probably have under the cover a bunch of different agents doing a bunch of different things and, and it's all going [00:10:00] to be kind of encapsulated with your one companion that you know.

Right. So is is that kind of a good ex explanation or, 

Mehmet: yeah, yeah, of course. Of course. And they have like a couple of follow up questions on that. So. Uh, you kind of touched it, uh, touched base on it. Like for me, like as a gamer, like, what, what would be you think, you know, the, the biggest, I would say change from a player's, perspec player's experience perspective.

Greg: So, so, um, so the, the good thing, the, the good thing about tech is that, uh, like in gaming is that they already have, like the, the users fairly well trained.

On, uh, these kind of things. So, you know, gamers are used to having macros and things like that where they can run, you know, different things to, to um, you know, kind of change their gaming experience and the, um, and, and then of course, you know, NPCs, right? So, so like non-player [00:11:00] characters are very, you know, well, uh, you know, well established in, in a lot of games.

So I think it's really just kind of, uh, leveraging. The, um, possibilities of what AI agents like, more intelligent NPCs, more intelligent companions, what they can bring to the table, um, and probably create some new gaming experiences for, for us, like we're, we are, um, you know, we, we started first by, by kind of making it really easy for.

Um, for newbies, not gaming newbies, but crypto newbies to, to be onboarded into the, um, Zika candy experience. So, you know, before we talk, you know, anything about ai, we, we, we have a way for people to simply create their wallet and use it, uh, via an email. So you, you just log in with an email. By the way, there's not even a password [00:12:00] because we're using one time password system, right?

So, you know, you put in your email, you get an OTP, you log into the game and that, and that works for any game that, you know, implements our kind of passport, SDK in, um, in their game. So, uh, you know, you, you, you log in with your email, you don't, you know, connect it to your email. You have a wallet. And, uh, the wallet is able to, you know, you can purchase your gems in the game or, you know, you can earn your NFTs, et cetera, right?

So it's really about like, what, what does the blockchain give you that, um, that you don't get in a typical gaming experience? And we start, of course, with the most basic thing that everyone, you know, uh, talked about is just that, okay, you, you can own your game in-game assets, right? I can't just suddenly turn off your account.

And you, you don't have access to your, you know, your all, all, all your life's work in that game. Right? Right. If you earn, uh, you know, a, a legendary, uh, [00:13:00] sword or something, you can transfer it. You can sell it. We're not standing in your way. Right. And, um, so, so of course you just wanna, you know, start with those basic layers.

And then what we found, uh, and what we have coming up, so is, um. Uh, we, we have like kind of a, a, a, a partnership aspect, but we are, we're, um, we're starting with the concept of, of rating, you know, like the most social aspect of a lot of these, uh, games was how do we all work together on a common goal to to, you know, defeat some monster or, or whatever.

And we have kind of, uh, some AI things coming with that where you, you can include your AI in a raid. And, um, you know, basically we call it raid quest, right? So you have these raids, uh, that you can, you know, spontaneously create together, and, and my AI can work with your AI to go and help accomplish those, those, uh, quests.

And, you know, there's kind ways to earn and, and that kind of thing [00:14:00] with that. Um, and we're actually, uh, so I think you probably noticed, I, I authored a standard. Uh, it's an accepted standard, uh, in, uh, ERC. So ERC 7 6 6 2 defines what a AI agent NFT is. And these are, these, these rate requests will start as like 2000 ERC 7 6 6 2.

Uh, NFTs, uh, ai, AI agent NFTs. Uh, and you know, we're gonna be programming some really exciting features into those NFTs and how you, you know, how you use them and how you can, you know, actually, uh, build on them. And I think I touched on some of that earlier. You know, giving your AI agent custom memories, maybe custom skills that are, that are specific, sorry.

Specifically good at, uh, you know, different, um. You know, aspects of what makes a good rater, um, and, uh, you know, then you can also trade skills and that kind of thing. So is is really gonna be fun. And that's gonna come out probably in the next six to [00:15:00] eight weeks. 

Mehmet: Great. Actually, I was about, you know, I wanted to follow up about the standard, but before, you know, going back to that, um, of course I'm, I'm a big believer in ai, but, you know, uh, I'm sure like, like other verticals where people would have some concerns about the ai.

Uh, what you tell someone who tell you now, Greg, like, Hey, like, uh, what would happen if. You know, the AI as we know, like everyone knows about the hallucination for example, or for example, unpredictability. Uh, like how you answer people who might question, you know, for example, the risk of breaking the game design or the rules or I don't know, like all these things.

So what kind of, um, countermeasures. Uh, are there, if there is any, or if there's even a possibility for something like this to happen, uh, in, in, in that, uh, space? 

Greg: Yeah. Well, for sure. Uh, I, I, [00:16:00] um, I originally got into ai, so I created my first AI company in like 2012. But I actually, I registered my AI domains back in 2010.

So like, I already knew what was coming. Right. And, and actually the real, the re the really, uh, the reason I got into it is I said, you know what I see, I saw the future already and I said, people are not going to put enough safety into them. Right. So I want to stay current with ai, always to know. How to turn it off, right?

If it goes out of control, right? And so far, I, I don't like, my, my threat level is, is not there. Like, I think people are waking up now to the risks that I saw in 2010. So that's a good thing. The more people that are aware, it's the better. Uh, of course the risk to, uh, employment numbers is, is always gonna be there.

Uh, I think it can be mitigated, but regarding, regarding the, uh, risk to gaming specifically, you know, [00:17:00] it's not healthcare, you know, it's not, it is not as high risk as that. Um, but uh, even in gaming, I think, you know, and, and how, you know, my answer to gaming, how you can kind of, uh, apply it to other industries that are much higher risk.

Um, so the number one is, is that, uh, of course for, for the, for the problem of hallucination, I've always found the best, uh, answer to a AI that hallucinates is another AI to correct it. Right? So here's actually the most basic AI workflow is when you have, uh, an an AI that's going to output, you can stick another ai, train it on.

All the ways that, that are typical hallucinations or key areas that you definitely don't want it to hallucinate at. Right? And uh, and then you ask the second agent to, you know, review the output of the first agent and correct anything that would be a [00:18:00] hallucination, right? So that's always like, you can put one AI together with another AI and you're gonna get a, a better result.

Um, but even there like. You know, LLMs are, are, are really good at a lot of things. Uh mm-hmm. Uh, but in order to get it to do exactly what you want, you, you know, that's where things like tool use come in or, or, or, you know, you need to, you need to massage it a little bit and then you need to put guardrails around it.

Um, and you know, how you, how you program. When you're, when you're building AI on the backend, right? However you're running your backend, uh, microservices or, or, or services, um, you have a lot of flexibility that you don't get in a chat window, right? So, so I mean, most of my day when I'm, when I'm programming ai, it's on the backend.

You know, I don't, because it's for myself. I, I don't. Really need to, to think much more about the user than, than about [00:19:00] myself as the user. But, um, you know, that's where a lot of the, so when you, you know, a lot of the guardrails come into, into that is like, how, how are you, uh, massaging the outputs, the inputs and the outputs, um, in, in, in the, um, perspective of how you're going, you know, how the end user's going to interact with it and, and see it.

So I think it, it would've be like. In terms of, you know, keeping it from breaking the rules of a game, that would be pretty simple. Um, and in the you, but you know, of course in the concept of, you know, spontaneous quests that players can create together, it's a little bit harder, right? Because you have to anticipate all the different things and, you know, but you can give yourself, you know, uh, ways to, to fix it.

Uh, a adaptively. In the case of something like, you know, healthcare, where you'd be worried that the AI agent might completely mess up the A, the diagnosis, and b, the [00:20:00] prescription. Right. I think you, you'd be, it's, you're, you're probably having to put it in a lot of guardrails and I think the, uh, yeah, I think it'll, that, that will slow down the development time and people who rush that will probably get caught, you know, so we'll see.

Mehmet: Yeah. Now talking about like, you know, we, we talked about the agents when you mentioned, uh, blockchain and FT multiple times. So you, so you mentioned about the, the ERC, uh, 7, 6, 6 2, and, you know, so what, what I'm, I'm curious about, and maybe for the audience, like, first, like what's the, what's the significance of standardizing, you know, the AI agents on the chain, and what do you think, you know, like in the general.

Like view would make, um, the blockchain, um, gaming ecosystem like truly usable for both developers and players. 

Greg: So the key is that in order for like, um, you know, you, you put a lot of time into a prompt to get it, to get it right, [00:21:00] right. And, um, an NFT, you know, if, if you were to put everything in the metadata right then anyone can just look up the metadata.

They get the prompt to build the agent and off they go. So there's no kind of market benefit to you writing an AI agent as an NFT unless you're going to encrypt the prompts. Right? So it was, uh, you know, and, and so basically the standard was, I wrote that actually, uh, you know, OA year, a year and three months ago, I think I started writing that standard, right?

So, uh, it was still quite new AI agents at all. Uh, and so I, I, I said it was a, it'd be a good idea to write a standard so that, you know, marketplaces or anyone wanna sell it can follow the standard and, and know this is how I, I expect to, to get an AI agent, NFT that is encrypted. And you know, if people write to this pro, to this standard, then.

It is, [00:22:00] it's the same way, you know, it's the same kind of benefit of EER C 7 21 or E EER C 20, is that you, you know what to expect if someone says, I have EER C 20 token, you know, okay, it's going to have these, uh, you know, storage variables, functions, et cetera. And so you can, you can expect to add a ERC 20 to your.

To your exchange and you know, you're going to be expecting these functions, et cetera, right? So in the same way, uh, eer, C 7 6 6 2 defines how you should be able to expect the encryption to work so that someone can, you know, uh, have an NFT that is ERC 7 6 6 2, and you're expecting to be able to have that user, so only the owner of the NFT can decrypt the prompts.

So that's the beauty of it, is that, is that you, you can list your nft, your AI agent, NFT, on open, see someone can come to it, see, see the public [00:23:00] information, but they cannot access the underlying, uh, you know, secret sauce of that AI agent FT they have to buy it. Only once they buy it, now they can see the prompts.

So that's kind of a, it's creating a. The necessary technology to, to make AI agent NFTs uh, commercially feasible basically. And, um, yeah, and you know, I will be kind of extending that with, with grade Quest, but that's like the core basis and that's, you know, like that's when you're writing a standard, it's, it's like you, you, you try and, um, you try and be careful not to overly, uh, specify the implementation details.

While, um, you know, making, making enough of the, um, uh, you know, you know, what's the benefit? What is, what is the, what is new here, right? And what is new there is, is the kind of definition of how prompts are stored [00:24:00] in relation to an NFT and, and how they're encrypted. So that, uh, you know, you, you, you, they're not exposed.

Mehmet: Cool. Um, so, um, that, that's very, very, uh, you know, insightful. Uh, um, uh, Greg, now I wanna ask you something like I. So this is, you know, something com completely new, I believe, you know, because, you know, when I was, so how, how far do you think like we, we gonna need until we have like, uh, a industry adoption for, you know, whether it's on the AI agents, whether it's for the, you know, for, for the blockchain on, in, in the gaming?

Greg: Yeah, so I mean, I, I think the, the big studios that have nothing to do with blockchain will. Obviously be implementing ai. I think there's heard something, they're doing some cool stuff at, um, you know, who grew the theft auto or whatever. So of course they will, they will, you know, implement, [00:25:00] uh, more, um, kind of companion type user experiences.

Um, and, but I think, yeah, I think the, um, like, so the, the, um. The companion aspect is actually, uh, a really interesting point and a really interesting point. When you talk, when we're, when we're talking again about the AI agent NFTs, like, uh, if, if some service provider gives you a AI companion, do you own that companion?

Right? Do you have, do you have access? Like, do you have always on access to their memory? So these are actually very good applications for. Blockchain. Um, so, uh, encrypting the memory, uh, such that the memory of the companion is only available to the owner. And, you know, basically regulating that by and using an NFT is, is, uh, is a pretty [00:26:00] powerful technology solution for that.

Um, and that's something that we'll, you know, we'll have already with rate, uh, these encrypted memories. So the, the, um. I think blockchain has a role to play, and there'll be a, uh, of course a, a big tug of war, uh, because, you know, maybe, or not, maybe definitely, uh, some service providers will say, well, you know, my part of my value is that I own all the memories of the companions.

Right. And I can, I can data mine and, and, you know, uh, use those memories to, um, train better and better companions, right? Or even worse, I can, um, I can use those memories to learn about, you know, human behavior and, you know, make games more addictive or, you know, make whatever more addictive or, you know, uh, uh, move populations in certain ways.

So I think that's, uh, always [00:27:00] gonna be the, um, the, the, the, you know, push and pull. And I think blockchain has a role to play in saying, uh, we're gonna give users more control, uh, more agency, if you wanna use that word, uh, over their AI experience. Um, and of course, like, it, it, it, I think everyone knows that at some point, probably soon.

Uh, you know, a five-year-old will start with an AI agent that, you know, basically follows them through life and, and helps them complete tasks in a better way. And there'll probably be quite a few providers of that. Um, and this will be one of the core issues, right? How do you, who, who controls the, the memories.

Mehmet: Great. Greg, one thing I maybe we didn't discuss much. So we, we talked about, you know, the, the, the player experience. What about the, you know, the developer experience, like how that looks like in, in, uh, in this, uh, Web3 gaming space and, um, you know how you held them at the [00:28:00] ZKCandy? 

Greg: Yeah, so that's, um, so we, we've done some really neat things in, um, first of all, we give them a, like a.

A hosted authentication solution so they don't have to worry about authentication in their game. They just, uh, apply our SDK. And so everyone, uh, every game that is on the network is, is taking advantage of these passports. So if, if you have a game and, uh. User could just go with their email, same email they use in any other game on Zika candy, and boom, they have their wallet.

Their wallet is there with the, you know, their hosted wallet. They can also have a connect, an external wallet. We don't like force them to use passport. Um, but they can, you know, and they can fund wallets back and forth seamlessly. Uh. They can transfer NFTs, et cetera, right? So, uh, but, but it means that, um, you know, the user doesn't have to have a, a, a different wallet experience on each game.

That's a tremendous friction point, right? You [00:29:00] would think if you're part of a network, you're playing on a, a, a blockchain network, that one game should be somehow compatible with the other, right? You're not hold holding 10 wallets from 10 different games. So, and of course, you know, if you're using external wallet, that, that, that's the same thing, right?

But, but here's, uh, just a, you know, another level of simplicity. Uh, but that, that also works then for sign in with Ethereum or anything. So they have the whole kind of, uh, sign in kit. They don't need to worry about authentication. So that's, that's number one. Number two is, okay, so you have this hosted wallet experience and you want to allow them to transact with it.

You, you've created your custom smart contract. It, it's not a standard one. You created it, you know, specifically for your game, and yet you can go to our passport and you just, uh, basically upload your a BI and, uh, it will, you know, parse that and give you methods now that you can use. Um, [00:30:00] so you, there's a, a set of standard methods and you say, okay, I'm going to use this player is using this method in the smart contract, you know, each smart contract has an ID that is generated when you upload your a BI.

Mm-hmm. And, uh, it's making, you know, a, a, a call to this ID and this method with these parameters and it just works. So they're getting this kind of hosted, uh, uh, you know, Web3 interaction kit. That every game on Zuki candy gets. So it's just like you go down the line. We're trying to simplify each piece.

And so when it comes to adding, you know, ai, we'll, we'll, we'll, we'll, again, you know, this is in the works. It's not, uh, live, but we'll have a, uh, a way of implementing, uh, any AI solution that we have available. You know, developers will just be able to interact with it in as, as a simple way as we can. Right.

So really then it, it, it, it becomes less about the technology, um, experience and just all about the user experience. What are we doing to make the user [00:31:00] experience better? And eventually if you just, you know, layer on enough simplicity, I guess, uh, make it as, as easy as possible for, for users to interact with it, eventually you'll hit that, you know, one, uh, one feature that, or one key benefit that, that, that, you know.

Uh, sells the whole thing, you know. 

Mehmet: Yeah. Yeah, that's, that's really, um, you know, it's exciting, I would say. And building on that, you know, um, Greg, from, from your experience, if, if you want to give an advice maybe for founders and builders in, in this space, like what does, what do you think in, in the gaming mainly, of course, uh, leveraging AI and Web3 and blockchain, what do you think, you know.

Like these founders they would be focusing on today from, from, you know, product features perspective and vision perspective to have, uh, you know, success. 

Greg: Yeah. I think, [00:32:00] um, there's kind of two ways to go about it. I, my, my first advice would probably just to go with composability, right? So think about, right.

You know, try and build something as, as small as you can that you think. Uh, it's still media enough that it's hard enough for that. It, you know, someone doesn't wanna just build it themselves, right? So if you make something too small and, and you come to, to, uh, you know, A-A-C-T-O or, or, or, or, or, uh, you know, uh, even a CEO company and you say, Hey, I would like to integrate this in your game, and it's like.

Well, that's, uh, 10 lines of code. I'll just write those 10 lines of code. Why not? You know? Yeah. Uh, so it can't be too small, but you know, if you, if you find the right thing where it's like, you know what, uh, that's eight hours of work or more to build, uh, and this person is really specializing at it and they've already hooked it up to a few things that can be interesting.

Then you say, you know what, yeah, sure. I [00:33:00] could build that. But then I would have to maintain it. And, uh, and, and they are learning every day. They're talking to all the gaming companies, right? Something like that is useful. And then if you tell me, well, it's actually, you know, 80 hours to build, well, that's really, really good, kind of, uh, um, you know, uh, feature set.

So, you know, little plug and play things can be very, very useful. Mm-hmm. Um, and, you know, make the business model as, as easy to, um, take on as possible. I find, um. Uh, you know, the s the, the, the opposite of that is you say, uh, I think this is always, this always works. This is also why I took my, taking my time with CEO, that ai, even though it's, you know, a very sex name and, and a lot of markets, is that in AI specifically.

The obvious opportunities are going to be handled by everybody, right? So, and I think AI is still, is still like very, uh, [00:34:00] still very new. And I think, uh, the, the, the really, uh, big, uh, opportunity takes time to really see what it is. Um. So, you know, you know, and I'm not devoting it a hundred percent of the time, so, so it's different.

If I was a founder, okay, I'm gonna devote a hundred percent of my time to a problem, I would probably say this is all still the best, best advice. If you're not going to go with plug and play and you know you have to be fine, you're not going to be the big cheese, right? If you're just a little plug play thing, you can make a very good life at it.

And you know. Grow into something, but you're, you're always a little bit in danger of being, you know, uh, just sort of added as an additional feature to some platform. Right. Um, yeah. The, the, the other way to go about it is to say, to do a lot of research into an area and say, um, well I see, I have a good sense of where the [00:35:00] market will be one year from today.

Um. I'm going to build something for that, right? So that by the time I'm finished building the products, you know, the right market timing will be there. But you know, this is very, this is very hard to do, right? But at least like if you take a year to build something quietly and, um. Nobody, you know, you know, by the time you've built it, someone says, great, I'm gonna copy that.

And you're like, great, I'll see you in a year. But I will have moved the year forward since then. Right. So it's, it is like you wanna have a certain amount of, if you just built something in a weekend. Guess what, someone else can build it in a weekend too, right? So it is, it's a little, it's a little tricky to, especially if you want to be a founder who's not gonna raise $10 million and, you know, really blast the market with marketing, then you're always going to be susceptible to copycats.

Um, and, you know, in, in today's market, you [00:36:00] know, you can build amazing thing. I'm always amazed what I build in, in like six hours using ai. I'm building stuff in six hours that would take a team of a hundred people, you know, six months to build not too long ago. Right. So it is, it is really tricky to figure out, like, you're not gonna be able to say, well, you know, we're advanced technologists.

That's our moat. You know, that's not gonna work in anybody's, no. Anyone who knows anything about what you can do with this technology is gonna say, well, I, I doubt that. You know? So it's, it's, it's really about kind of finding, I would say like. Uh, yeah, some something that you know, you know, and that you were able to build long enough, you know, infrastructure so that when you hit the market you can kind of capture a lot.

I think it goes back a little bit, this is more marketing advice. Uh, you know, if you find a, a, a core group of users and you know, that kind of initial beachhead, uh, target [00:37:00] audience and you kind of get enough market share. In that audience, you've built something very specific for them. You capture 'em enough of the market, you start to drive that word of mouth engine.

I think that's always a good strategy, is even today a good strategy. Um, so yeah, it's tough. It's, it's, it's, I would be, it's tough to be a founder. It's always tough to be founder in any market. Uh, but today where, you know, one of the biggest, uh, challenges always in founders, just copycats, right. Copycat, especially if they copy you and they raise a bunch of money and give the product away for free just to try and knock you outta the market.

There's gonna be a lot of that, I'm sure. Uh, yeah. Even more than, than before. 

Mehmet: Yeah, absolutely. Like yeah, copycat is always have been the problem there, Greg, as you know. Absolutely. Um, Greg, as we almost come to an end, like, um, any final thing that you want to share, maybe something I didn't ask you about and where people can get in touch with you?[00:38:00] 

Greg: Yeah, so I mean, uh, Zika Candy, you know, we're live on main net. We've been live for a while. Uh, we recently completed, uh, uh, an acquisition of of Chroma, and we've got some exciting things coming out, uh, for chroma. Uh, related to this raid quest. So I think number one, I think, uh, people should be looking for, uh, our announcements about raid Quest.

Um, and, um, uh, you know, in terms of, uh, uh, finding about Zika candy in general, uh, of course, uh, Zika candy.io is the main site. Um, and, uh, you know, we, you, you'll see our, kind of our explorer, our SDKs and everything to, to use it. And we have a, a, a number of games to play like Candy Defense and, and Pepe Kingdom that are like super popular.

So I think that's kind of the, the, the good way. To get attached with me personally. You know, I have a LinkedIn, uh, so that's probably the easiest way. Uh, and I'm pretty sure you can, uh, find me pretty easily on there. Um, yeah, and I, I, I, [00:39:00] you know, I, uh, if you have any questions about any of the, the stuff I'm working on, then I'm, I'm always, my door's always open.

Mehmet: Great. Thank you so much, Greg. Now I'll make the, you know, audience lives easy. I will, I will put the, uh, links in the show notes so they don't have to look right and left. Um, as we come to an end, Greg, I want to thank you again. You know, I learned a ton of things from you today, like, uh, about this really cool integration, you know, or like, let me call it this, uh, intersection of, of two, um, you know, exciting, actually, three exciting wars gaming.

AI and blockchain and Web3. So I, I, I learned a ton from you today, so thank you for sharing your, uh, your knowledge with us and your insights. I really appreciate this. And, um, this is for the audience. This is how I end my episodes. So guys, if you just discovered this podcast by luck, thank you for passing by.

I hope you enjoyed it. If you did, so please give me a favor and. Subscribe, share it with your friends and colleagues and if you are one of the [00:40:00] people who keep coming again and again, thank you very much for your support. Thank you for being loyal to the show. Thank you for taking the show to the, a new level this year, 2025.

Now we are Eric, this probably in July. We are recording in June. Um, you know, the podcast is doing really well this year. We've been on the top 200 charts in multiple countries at the same time, and this cannot be happening without my guests. So really appreciate this, and as I say, always stay tuned for a new episode very soon.

Thank you. Bye-bye.