Binance Live AMA with Cryptokang
Disclaimer: The summary and transcript below were generated automatically from the full AMA audio. They may contain unintentional inaccuracies. Please refer to the original recording for verbatim quotes or affirmation of details. Notify us of any factual errors so we can correct the transcript. We make no guarantees about completeness or total accuracy.
Summary
About the Host
Cryptokang's Binance Live page boasts 57.6k followers and almost 9 million likes. His socials are below.
I. Introduction
- CEO Vladimir and marketing advisor Vinny join the AMA discussion
- Provide backgrounds to add context for their experienced team
II. Founders' Backgrounds
A. Vladimir
- Masters in cybersecurity and 7+ years AI research experience
- Previously led AI lab across domains like computer vision
- Passion brought AI and crypto skills together in founding AlphaKEK
B. Vinny
- Over 17 years in traditional finance, fintech sectors
- Immersed in crypto world past 4 years since 2017 bull run
- Philosophically followed bitcoin thinkers since 2010
- Dove into DeFi in 2020 sparking project interest
III. What is AlphaKEK?
A. Building AI tools for the crypto space
- Consumer products such as chatbots and analytics
- Business offerings including custom dashboards and bots
B. Problem they aim to solve
- Focus on unbiased crypto analysis delivery through AI
- Risk of dependency for others using external models/data
IV. Team Behind AlphaKEK
- In addition to founders, diverse technical contributors
- Blockchain developer, frontend designer, AI modelers
- Guidance from advisors aiding strategic progress
- Transparency on team backgrounds through public profiles
V. Current and Planned Features
A. Token audit functionality via language queries
B. Sentiment analysis and alerts under development
C. Goal of individually personalized alpha insights
D. Real-time metrics on popular tokens via on/off-chain fusion
E. Streamlining process to ease discovery of strategic opportunities
VI. Long Term Vision
A. Proactive AI agents requiring no direct user querying
B. Automated crypto data analysis for tailored updates
C. Central challenge of pacing fast moving AI technologies
D. Optimizing user strategies without personal effort requirements
VII. Testing Functionality
A. Telegram bot available at no cost to experiment basic features
B. Planned collaborative experiments like meme coin sessions
C. Demonstrating insights and strategy optimizing capabilities
VIII. Discussion Takeaways
- Overview of technical achievements and advancement roadmap
- Welcome for returning with future progress reporting
- Appreciation for interest in strengths of assisted strategies
- Potential for individualized market insights attracts curiosity
IX. Closing/Summary
- Founders expressed passion for supporting crypto community
- Goals of personalization and automation seem transformative
- Continued development generating much anticipation
- Future looks promising for AlphaKEK and traders it seeks to aid
Transcript
SPEAKER A
Hey, how's everybody doing? Long time no see on Ninance. The camera is working. Amazing. How you doing, guys? Today we host two people from a project that looks very interesting. The name is Alpha Keck. As you may already know if you're in my discord, they call themselves our web three copilot. They say they offer cutting edge AI tools for crypto and beyond. I'm very curious to know what that entails. And I know some of you are curious too, because you sent me questions in the DMs. I'll call them soon and we get started. I want to make sure a lot of the gangsters are here because this can be a good one. Welcome. Let me ping the discord one more time. You guys doing good? Yeah. No music today. More serious vibe. All right. Yo, chicas. Long time. Welcome, guys. I'll call them in a few minutes. Just waiting for some more cancers to come in. Let's get at least like four or 500 viewers and we start. We can make it good. It might coffee here. Get a drink, guys. Get a snack. But, yes, Alphake AI is the project we host today. A lot to be said. I've looked into their documents this Ama. We could make it last three or 4 hours if we really went in detail. But the person joining the stream today are Vladimir, the CEO, and Vinny, the marketing advisors. As you can see, I have super big roadmap. And there's even more to it. No quiz, guys. You can ask your questions, though, and good questions. If you ask really good questions, you win something in the discord. Let me see if they're available. Now I set up the call and we get started, guys. Let me know for the sound levels, guys. Hey, guys, thank you for joining us. Can you hear me? I can hear you guys, too. Amazing. All right, you're on the screen as well. Amazing. Okay, welcome on the stream, guys. Very happy to have you. Let me put the volume higher for Vinny. Nice background. I just noticed. All right, so, yeah, I was introducing you guys. Alpha ki, it seems super interesting. I was telling them from, based on the information I could gather, just doing my own research. We could go in detail and make it last for hours and hours. But let's begin with first things first. Right? And how about giving for my viewers an introduction on you guys? Who are you and how did you start your crypto journey?
SPEAKER B
Yeah, so I will start. Hi, everyone. Hi, I'm Vladimir. I'm a founder and lead of AI at AlphaKEK AI. So basically, I'm an AI researcher with seven or so years of experience. And originally I have masters in cybersecurity, so I have some hands on experience with crypto as well. And since my teenage years, I was a fan of crypto, and I was working on computer science and data analysis for all the different tasks from computer vision to astrophysics. And now I'm putting my two favorite things together. Yeah. And we have urgent AI and crypto.
SPEAKER A
It sounds very promising. What about you, Vinny?
SPEAKER C
Yeah, so I'm Vinny. I handle marketing and business development at Alpha Tech AI, and I have about 17 years of traditional finance and fintech, about four years in web three. So about the same amount of years as in AI as well. My start in crypto started probably in 2017 cycle with all the iCOs. That was a big boom back in the day.
SPEAKER A
Yeah, big one.
SPEAKER C
Yeah. Then ended, I think, like early 2018 with the bang. But philosophically, I would say this is going to reveal my age. But it started back in, I think, 2010. I was a huge gold bug, so I was into sound money, and I was an anarcho capitalist, which is philosophically similar to a lot of the bitcoin OGs. Right? Max Kaiser, Eric Voorhees. But I think what a lot of people don't realize is Max Kaiser, Eric Voorhees, da Vinci, Jeremy, these bitcoin OGs, they were gold and silver bugs back in the day. So I was basically following them in like 2010, 2011, watching their YouTube channels, and they were talking about silver. Max Kaiser, he had a show on Russia today and he was saying, buy silver, crash. JpMorgan.
SPEAKER A
Yeah, it was huge back in the. Yes, because that's around the time I got into bitcoin as well. I remember it was huge. It blasts from the past you just brought right now.
SPEAKER C
Yeah. Back then he had an iranian news network. He was on a news channel called Press TV in Iran. And then he switched over to Russia today in 2012. And then on episode two, he started talking about bitcoin for the first time. So that was my foray into the markets. But I think it makes a lot of sense because the two communities, they share a lot of the same views on government intrusion, personal freedoms, inflation, all the risks in the banking system that's so centralized and these webs of counterparties. Right. And in 2012, I started working for Peter Schiff, funny enough, the Economist, smaller, the world's biggest bitcoin bear. I was helping trading and launching some of his products. And then, long story short, in DFI 2020, well, DFI Summer 2020, I really got into how DFI changed finance, using code and algorithm and removing the counterparties, creating tools that were a lot more fair in terms of interest rates, in terms of risk. And yeah, I was a total convert after that. And then 2021, it was a breath of fresh air just being able to do financial services without going through a bank uniswap. Right?
SPEAKER A
It seems to think about it like ten years ago. Inconceivable.
SPEAKER C
Yeah, because you had to go through a custodian, all these gatekeepers, and who knows what they're doing with your money, right? So, yeah, long story short, in 2021, I started working in DFI. I was an early adopter to AI through this tool that we built on this platform called a token recommendation engine. It basically looks at your wallet and it recommends different tokens based on a bunch of other wallets out there. Hundreds of thousands of wallets through machine learning. And then in 2022, I worked on a personal project which was a music generator using AI. So it actually looks at your NFTs and it creates one of one songs from your NFTs.
SPEAKER A
That's amazing.
SPEAKER C
And then eight months ago, I met Vladimir through my favorite source of alpha four, Chan's Biz.
SPEAKER A
I mean, it's not what it used to be, but it's definitely up there still.
SPEAKER C
Yeah, it's quite spammy now, unfortunately, but that's inevitable. But back then you had like, amazing Alpha, I found sheep. But you knew January 2021 through one of the first posts on there. It was amazing. Yeah.
SPEAKER A
Actually, I have to be honest.
SPEAKER C
You'Re a chain link marine.
SPEAKER A
Yeah, it used to be like a lot more. Yes. Now I just hold and pray. But yes, that's how I discovered it.
SPEAKER C
Yes, I met Vladimir because we both knew some of the admins of the biz telegram group and he just announced Alpha Keck's launch in the message and I joined in love the work, became a holder, started contributing and I couldn't really figure out how small the community was based on Vladimir's background. You know, his. Just the age brackets.
SPEAKER A
Yeah.
SPEAKER C
Long story short, we became friends and yeah, now I'm doing marketing guys, really.
SPEAKER A
Solid background and intro, and it's crazy how things happen in crypto. You meet people randomly and like you guys, and now you're working on a project together. The project, which is Alpha Cac AI, as you guys mentioned. And how about we dive a little bit into what is Alpha Cac actually AI to begin with? Can you describe it for us?
SPEAKER C
Yeah. So in the most simple terms, it's a AI engine that powers different tools for consumers as well as businesses. So on the consumer side, you have chat bots, research tools. On the business side, you have custom analytics board dashboards, custom analytics, custom chat bots based on their Persona. So these are crypto projects. They could be blockchains, ecosystems, influencers. And yeah, we have several clients on the B two B side. We actually just announced our airweave partnership. So we built a custom analytics board dashboard for airweave, and we ingested a lot of their public and private data, which Vladimir can discuss a little bit later. But yeah, essentially there's a lot of AI projects in crypto that are providing alpha. I think that's the main use case. People want Alpha. They want to be curated content of what to buy and what to sell. We'll get into this sort of issue. But what we did was we decided to build infrastructure first tailored for crypto, because a lot of these underlying technologies built on OpenAI's API or other major LLMs, don't provide the right output for crypto. And yeah, we can get into a little bit of that, but there's a lot of key tech that we built that's very customized over the last six months that drove a lot of these products that we're actually starting to see new products pop up organically based on the data that we're seeing, and we're finding new use cases for this output, sentiment analysis, token audits, data visualizations, which we'll get into. That's in a nutshell.
SPEAKER A
I mean, there's so much you can apply to when you get down to it. Before we dive deeper into what Alpha cake is, I think one of the most important things, and one thing that drives a project to success is knowing who's behind the team, who's behind the project, who's leading it. And you've introduced yourself. Vladimir introduced himself, the CEO. Could you tell us a little bit who's behind the team as a whole? Who is the team? Who is behind the project? And before we move into more details, yeah, sure.
SPEAKER B
If I knew Vinnie would be like telling so many details about himself, I would brag a bit more, too. So, yeah, as Viny mentioned, yeah, I was the lead developer of the lab that was mentioned in open AI blog post together. During private beta testing of their tools. I presented on an Nvidia GPU technology conference twice in 2021, 2022. Basically, I was doing some work for analyzing large scales for physical data. And actually the discoveries they've made there, they brought me basically to Alpha. I realized I can try to re up all this stuff for crypto data now, and that's what we are doing. So yeah, we have a team that is constantly expanding. So yeah, we have. Aside from us too, we have Dimitro Sotic, who is not my relative, but we have like similar surnames. He's our Webfree developer who has experience with a lot of various projects and ecosystems, from Ethereum to Cardano, and who's our web free stuff and integrated onto other chains. We have one more, Vladimir. Vladimir, who is our front end developer who has a huge expertise in data visualization. And for now he was mostly showing us just skills related to website building. But actually we are working on some really cool data visualization apps that we will show people soon. And we have Valerie Sotnikov, who is my relative. I have a brother who is also like AI researcher, but he's working not with the text, but with multimodal AI, with image models. And we are working on some specific stuff just for crypto community that we will reveal soon, I think. And we also have multiple advisors that are just helping us to reshape the alpha gap to make it the best possible product.
SPEAKER A
Seems like an all star team, to be honest. I can see you have all your Linkedins, your Twitters, your telegram, all linked on the website for everyone's peace of mind, if they want to deeper on who's behind the team and where did they come from. We spoke a bit, Vinnie mentioned how you guys came together a bit. We can maybe go into more detail about it later. Right now I wanted to ask you, because you both mentioned there's a lot of AI buzz right now in crypto and other what's alpha cake AI trying to solve? And what makes you different than other AI projects in this space? Can we know that?
SPEAKER B
Yes. Regarding what problems do we solve? Vinnie briefly covered it. That's what we can broadly name finding the alpha, and we can formulate it as figuring out insights from public information earlier than anyone else could and make decisions based on that, execute trades or whatever. And regarding what, and we provide tools to do this. We have reports, we already have data visualization tools that we improve, and we have tools and both available as a bot and a web app. And what's really different about us compared to most of the other startups in the field is that we have some of things that no one else has at once, both of them at a time. So first of all, we have our custom model infrastructure, we have data infrastructure. It's actually quite simple. So if you are building AI, you need models and data, right. And if you are using someone else's models or someone else's data, eventually you are doomed. You will be replaced. That's our credit from the very beginning. Both of them are our own of the data. Both models, right?
SPEAKER A
100% yours. You haven't borrowed from open AI or anyone else?
SPEAKER B
No. Of course, we are based on open source models. Right now, we don't depend on anyone, any other service.
SPEAKER A
Forgive me, I didn't want to interrupt.
SPEAKER B
So, yeah, as vinyl also mentioned, we have terms of data infrastructure. We have our custom AI engine that we call alpha cat fractal. And the idea behind this agent is to basically create a special tool designed for the AI, to be used by AI, not by humans, that basically ingests all the various crypto related data.
SPEAKER B
Dex trades to even for chump form threads, all the various stuff that we can find in very different shapes, charts, trades, smart contract code news, and build on the fly, a special knowledge graph that has all the connections between these points and this being generated by AI. So we have basically like two models. So one is a large that builds all this stuff on the fly, getting you and new data points, like in real time, even right now. And another one is a smaller model that just retrieves parts of data from this knowledge graph and feeds it to the user. And the user uses our apps.
SPEAKER A
Where the data comes from, or is it the same for every member paying member, for example?
SPEAKER B
So for now, there are two scenarios where it could be different. So first of all, our functionality is token gated. So if you own some specific amount of AIQ tokens, you unlock more sources of information that are being used at the same time. And that's where our engine really shines, because since we have all this data already preprocessed in a very efficient way, we can use much more data sources at once per one response that any competitor could do. Basically, we can use very usual chat, GPT or cohero, or whoever else will show you like five sources. We will do like 25 or 40. And more sources means more accurate data, and more accurate data means more, like, more comprehensive data means more accurate response. And that's what the alpha is about, to have more precise, more relevant information.
SPEAKER A
Amazing. No, go on.
SPEAKER B
Yeah, we can describe this thing in depth, if you have time, or if.
SPEAKER A
You guys have time, I'd love to have a little, because I was curious to know why you think the crypto market analysis, the alpha, is what's going to be like the main driver. The main thing people are going to ask from an AI.
SPEAKER B
Okay, yeah, okay, sure, let's continue. We could do this. So, yeah, we have knowledge graph, we have two models, and the key thing about these two models, aside from just that, we have knowledge graph and it allows us to retrieve more data. The thing is, these two models are fine tuned to be as unbiased when it comes to financial decisions as possible. Because if you try to use any publicly available model like your public API, GPT 3.5, GPT, four cloud, et cetera, et cetera, you could try to search it online and see that on some tests they have their biases derived from the training data. We fine tune our models to be as neutral as possible when it comes to making financial decisions, and so they won't try to lean towards some specific decision and only look at the data when answering the question. At the same time, they fine tune to always reply to the user, because that's another problem that you could have when you use all these publicly available models, they are not suited for this task. Basically they will tell you, like consult financial expert, they won't give you their answer outright. And that's the thing that our models never do. They will always give you a question and they will never try to add their own opinion unless you will explicitly ask for. So for this, for the data infrastructure, there is also one more thing that we have and other projects don't, is that we not just retrieve the current info from the web or from the blockchain, when the user asks questions, we save it for sure for the future. So our graph is like growing nonstop now. Website, we have, like I mentioned, that we have over 30,000 data points, but actually it's over 40,000 now. We just didn't have time to update this. We are getting more and more data. And the key thing is when the users ask more new questions, we are getting more and more data that will be reused for future analysis. That's for example, how a new visualization like sentiment analysis tool works, that Winnie mentioned initially when user was retrieving some data, getting some answers to their questions, we were adding all new pieces of data to our knowledge graph without any specific plans what to do with it. And now we have all this data available in the retrospect. And now we can run sentiment analysis for some specific submarkets. For example, we can check what's the sentiment for binance tokens, what's the sentiment, for example for meme coins? What's the sentiment for bitcoin ordinals or some specific projects and see how it's changing over time. And our idea, like, for the next steps is to add also alerts, follow this stuff and make it interactive. So you could fetch multiple different trends and see if there are any correlations between them and get more insights available just to you based on this.
SPEAKER A
When you explain it like this, it's a bit scary, because it's scary for people that are not going to adopt such a tool. It's going to be impossible to keep up. Let's say, when you guys are fully operational with everything everyone is using, you guys, or a tool like yours, it's going to be impossible to trade without this advantage, without this assistance. I see here on the premium tier you offer also unlimited AI chat and so many things. You guys can handle many users already. How would it work?
SPEAKER B
Yeah, so that's one of the tricks, as I've said, we have like two models, one big one and one smaller one. So big one is used for data ingestion. So basically, when we have any new kinds of data points, as I've said from Dex, traits from newspapers whatsoever, the bigger model takes this data, preprocesses it, and then injects it into the knowledge graph. That's the part that is compute intensive, but it should be done only once. So once it's done, yeah. Users use this data for really cheap, and that's our trick. So basically, we have a big model that gets the data and injects us into alpha cat fractal, and then we have a smaller model that takes all this already precomputed data, these knowledge subgraphs, and generates answers for the users based on this stuff. And that's like 100 times faster than ingestion.
SPEAKER A
Even if you have like 50,000 users, people can have the data real time responsive.
SPEAKER B
Because these parts, we are like asynchronous and independent from each other.
SPEAKER A
That's amazing. Thank you. I understand you a lot better now because you guys are on Ethereum, and I was signing in, as you was explaining, and I was trying to see if I could see any of the free reports, but I also saw you guys integrated the binance chain, the BSC, into the new token audit function for the bot.
SPEAKER B
Right.
SPEAKER A
Could you also tell us more about that and on why the choice for the Ethereum chain to begin with?
SPEAKER C
Vladimir, can I lead this R1? Yeah. To go back to where we started from point A, which is the infrastructure, the underlying infrastructure that creates good output. The token audit use case was actually a surprise. It started off with, mentioned the unbiased model, the knowledge graph that allowed us to process all this information, the massive data source that we're ingesting across web two and web three, off chain data. Now that all that data is there and we're making connections, the output is surprising us. And one of the things that we noticed was it's really good at token audits. So a token audit is essentially using a function. Well, if you go into our telegram bot or chat, you go ask. You can ask for an extensive rug or scam check using natural language. So you could actually say something like, ask, is bobo secure? Give me a detailed explanation. It'll tell you something. But because it's natural language, you can actually have fun with it. You can say, is bobo secure? Tell me in the form of a haiku. And I actually did this earlier. It basically said, the haiku is bobo smart contract is secure. Anti whale measures are in place. Warning, blacklist could block. And then I did another one, which is, is bobo secure? In shakespearean iambic pentameter. So it said, in the realm of Bobo, the smart contract's code is open source, yet bears caution it possesses anti whale measures. A blacklist feature lurks, a worrisome trait. Security audits find the contract fraught with risk, thus tread with caution. Error thou invest. Essentially, there is a token audit tool, and we can integrate with different blockchains. And BSc was actually announced two days ago.
SPEAKER A
Go ahead. My bad. You made me want to ask something. I don't want to interrupt real quick, but what are the capabilities of UAI in terms of how it can process speech? And like you said, it could render the result as a haiku or shakespearean. How would you compare it, let's say, to chajipt four or Grox ability?
SPEAKER B
If you are talking just about generating haiku in general, the intelligence of it.
SPEAKER A
I'm just trying to understand, is it more for trading, or is it like something that you adapted for trading? I'm not as well versed as you guys in the topic, and I use the excuses to make it easier for the viewers to not embarrass.
SPEAKER B
That's absolutely valid question. So, yeah, it's adapted for trading and crypto related questions. So, yeah, if you take the same kinds of information. So the haiku part and rhyming probably will work better for GPT four, right. If you are talking about specifically BC token audit, most likely GPT four will refuse to answer at all and say consultan expert. Yeah, but if we are talking about some isolated cases, yeah, GPT four, of course, will be better, but we are not trying to compete in terms of general reasoning capabilities. Of course they have much higher budget than us. We are working on really specific narrow cases to analyze crypto market a little bit of threat Phi if it's important for the context of how crypto doing.
SPEAKER A
Of course I didn't mean to be real, I was just curious because the way Vinnie said the high blew my mind that it can do that. It's crazy.
SPEAKER B
The case is that we are making our tools more and more complex under the hood, but on the surface it still will be just the ask commandment. So all the time we are adding more and more mechanisms that are making underlying processes of analyzing the data more and more sophisticated. But for user there will be no difference at all. You just ask the same kind of basic question, but you will get a better reply every week.
SPEAKER A
All right, so it learns that fast, week by week it will be better, smarter and no more and so on.
SPEAKER B
Right now that's our pace. We'll try to keep it amazing. So we have a lot of stuff in our roadmap. There are a lot of things that we could make it.
SPEAKER A
The tech seems amazing. I've been taking a look as you guys were talking on your website and on your examples. It's amazing team, amazing tech. You guys seem to are going to be a great tool for traders. But there is one thing, and this question is more for viny, I think what's your marketing and go to market strategy? Because your product is good, but people need to have eyes on it. What do you have in mind and what do you have planned, if I may ask?
SPEAKER C
Yeah, I guess you can separate it by B to B and B to C. For B to B. There is a pretty well established route to market for custom bots and analytics sites. Pal AI has done a great job doing this, serving basically creating custom bots for different projects and different blockchains based on their Persona and based on a custom knowledge set. So that's an easy route for us to take. We have a couple of institutional clients, enterprise clients already. The first one is actually a Mimi token and the reason why they came to us and in full transparency, I was a marketing lead there and I actually helped bring the project to Alpha Keck and it's a Sam Bankman freed meme token called SBS. And the project was trying to create a chat bot in the telegram that was trained on his Persona and his style as well as a Caroline Ellison bot. And this harkens back to the issue with a lot of AI projects in web three, which is the output just wasn't good enough. It was a very sort of generic response, or oftentimes when users would ask questions about finance, it would yield. Can't answer this question? Please diversify your holdings. Talk to an expert. Right. So that was one of the reasons why it was a good fit. And if there's projects that want to have a use case beyond just simple community management of asking about the project's knowledge base or docs or something, if they actually want to do crypto research. We mentioned the token audit earlier and later Alpha then I believe it starts with the unbiased model and the infrastructure that we built specifically for crypto. And the haiku thing earlier was bit tongue in cheek, but what it shows is that you can actually generate content about finance. And that's why I actually went to Alpha Tech originally to use the app, because I was working traditional finance, I was writing newsletters. I want to research biz as well, four Chan, but the tools I was using, Chachibt, wouldn't give me the output about those specific topics. Claude, for example, from anthropic, won't even give me answers about crypto. So I needed an uncensored, unbiased model to even do that work. So it sort of starts there, but to go back to the go to market strategy, yeah, there's enterprise clients, there's blockchains. We mentioned aeroweave ecosystems where we can offer them for Airweave. We created a custom news site, analytics site based on their data, and they were actually surprised after we ingested all this on chain and off chain data, they noticed that we had information in there that they even forgot about. They were like, wow, we haven't read this or thought about this in a long time. So we were able to pull out and mine a lot of pretty deep stuff on the B to C side, on the consumer side. Right now, the app is sort of this broad chat GPT use case. But as we start seeing the output come about and we start seeing other tools in crypto trend alphabets that are telling you which tokens are moving right now, right? Volume increasing, momentum is increasing. There's a lot of alpha hunters in the space. We can actually replicate those use cases, but with our engine instead, because we have a feeling that we're not really sure what infrastructure they're using, but we have a feeling that a lot of these projects are simply using major LLM APIs. That's not a knock on them. They went to market fast, and they're sort of geared to go to market, but we took the long route of going infrastructure first, which I think is the right way to do it. And now we have the luxury of creating these templates from this output based on different use cases. And we actually invite your audience to recommend different alert bots or alpha bots that they like, and let us replicate them pretty quickly.
SPEAKER A
I saw when I was joining your telegram that you can also make custom bots for communities. I was curious how that will work. I was going to ask you guys after the if it was possible to do one for us, because when you mentioned the SBF one, I'm assuming there's not enough data of him to make it have his mannerisms and the way he talks. There's not enough footage of him online, I'm assuming. But maybe for someone that has a community that has dozens of hours of content online every week, maybe it's easier. But yeah, of course I'm more interested in the B two C side because I would use it myself, but I didn't even consider that the B two B might be really big too, for such a tool. And it makes sense, to be honest. One question for a customer like me, would he be able to give me data only on the meme coins that are trending in a real time enough that I can act on it? So, like, analyzing the new tokens that got launched on exchange, for example, seeing which one gets the most volume, will that be fast enough for me to have a sizable edge compared to a non user?
SPEAKER B
Yeah, absolutely. And that's just a part of general direction that we are working on. So the issue with any conversational interface with AI chatbot is that when you ask questions, you need to formulate that you need to have some information on the topic already before to be able to ask a question. And that makes discovering entirely new stuff extremely hard. And one of the things that we are solving at the moment, and we think that's a really important thing for any kind of AI chatbots, not just for crypto, but for crypto too, is to make the system proactive. So that means that you don't need to ask questions. You either just say what you are curious about, or you don't say anything at all. And it should be the task for AI to figure out what's the useful information to show you. And that's the thing that we will be working on, as we are already working on, is to basically create AI generated alerts that will be tailored just for you. Instead of writing a message to AI, getting a reply, writing a message, getting a reply. It will be like you start the EI agent and it will send you what you need as often as needed.
SPEAKER A
That's a game changer.
SPEAKER B
Our ultimate goal. That's a game changer.
SPEAKER A
Honestly, that's the most annoying part, so to speak, is having to go back and forth, back and forth and redefining your query and changing everything. Of course you can't have everything, but now you guys are doing it. You don't even need to do that anymore. Which brings me to my next question, because meme coins for crypto is really fast and I could use an AI to keep up with such a fast moving scene. How do you guys keep up with crypto and AI that moves so fast? Both spheres move so, so fast. Huge leaps just in the past six months. How do you guys keep up with it? Do you have also an AI assistant or how do you do it?
SPEAKER B
First of all, we have AI reports and one of them is available on our website. It's basically a tiny dashboard for some specific topic that self updates every few hours. You can check one on our website and it contains some latest information. We have one for meme coins, we did a custom one for folks from airwave and we could make a custom one for you if you're interested about, for example, BC existing. So yeah, in terms of crypto, it actually gets easier and easier over time because as we are getting more holders, we get more and more people who are not just token holders. They are really interested in moving this project forward and getting our tools more better, more suitable for them. So we are getting more and more and more feedback and we have.
SPEAKER A
Quite.
SPEAKER B
Clear picture of what people want and we are basically implementing stuff they want step by step. And regarding AI, well, I have quite an experience like working with all this stuff. So I was like fine tuning, for example, for some personal research, GPT-2 when it just came out. So all this stuff is not so new for me and I know where to look. So it's indeed hard to keep up. But we are doing this and we are like for example, trying to think, not just to keep up, but to think ahead. For example, what would be the next pivotal moments for AI? This year we had one right now with Nvidia. So they are going hyperscale. So like gaming GPUs? No, just like selling one GPU? No, they sell server racks. One server rack is just one minimal product. You pay like multiple millions.
SPEAKER C
Yeah.
SPEAKER B
And you should keep this in mind that on one hand they are scaling another. For example, important trend is that more and more labs that used to be open sourced, they are getting like closed AI. Open AI was first. Mr. Oi was acquired partially by Microsoft. They are now closed source. As far as I know from talking some sources, the guys from meta still committed, for example, to make llama free open source. And that's, for example, that everyone in crypto AI field should keep in their mind. For example, if you're working on stuff, you should keep in mind that we will get like in a few months we will get a new future open source model, and most likely that will be easy. There are new prospects that maybe some chinese labs will save us, who knows, in terms of getting lower consortium models that we could retrain for crypto data. But otherwise, we should keep in mind that this is the general trend. The next foundational model from meta would be the last. And if you want to have your custom AI, you should figure out a way what to do. So Elon Musk just released their grok, so it was promised to be like open source, but if you take a look, so they don't have any code for training attached. You just have the weight of the model, like the model file itself. Okay. And the small tiny piece of code, just how to send prompts. For example, if you want to fine tune this model on your data, they don't provide code for that open source.
SPEAKER A
They said, the way they said it meant like you could eventually trade your own.
SPEAKER B
Yeah. So kind of, they showed us the piece of their stuff, but for example, they haven't released the version of the growth that they had in their chat. So this is like a basic version that they fine tuned. And this fine tuned version, it's not public, so they partially kept their promise, but still you need like a couple of millions of dollars.
SPEAKER A
Just put the entry back. I mean, it was already super expensive at the very beginning when the first, the chat, GPT-2 or 2.5, I forgot the ones that were circulating on fortune that could reply to you could have a chat already. To run those, you needed a pretty decent investment if you wanted to do it good, or like your Dali too. And now it's going to be possible, I think so. I'm glad there's people like you guys making it. So of course you tend to gain a lot from it too, but you make it so that people like me can access and have the advantage that the AI offers without having to put millions of dollars in, just like the hardware, because then there's the training and the science behind it, which I don't.
SPEAKER B
Have to be honest. Yeah, we are not going to just close our plans and just letting you know, when you see some other AI project, take a look and think, are they planning actually for the future this year or not? Because, for example, one trend that we previously briefly touched is like proactive AI AI agents. So Business Insider published a post, I don't remember one or two days ago, that GT five will probably came out roughly at the same time as llama three by meta this summer, and it will naturally support autonomous AI agents. And this will be like the next big thing. And that's also direction that we are already working on. We release some prototypes available for users to use before them. But yeah, I can't recommend or say something about our competitors just put forth. So there are some clear trends happening right now in this year, and a good rule of thumb how to determine whether those projects survive or not is just to take a look. Do they understand what's going on? And if they at least understand it, maybe they don't have a clear plan yet. Yet. But if they at least understand and acknowledge the current trends, not just trying to follow what's happening like this very mind, then probably they will survive because that's enough. But if not, that's a red flag for me.
SPEAKER A
The way things move, we're going to know very fast who's going to survive and who doesn't. The way you speak, it seems you guys are planning for the future and you're trying to not just catch up with AI every time, but where the AI comes, you're already there in terms of what you want to do with it. So I think that's a great advantage, as you said.
SPEAKER B
Yeah, thank you. So of course, we are not profits. There is a lot of uncertainty and we are flexible. We just keep in mind what are stuff that happens for sure and what could we do more than that? And we have some carcs up on our list.
SPEAKER A
Well, talking about the future for the last question for you guys, because if you look at the upcoming page in your website, there is so much to be excited about. What would you say out of all the upcoming features on your roadmap that are planned, what would you consider is the Keystone feature or the one you're most excited about or both? Up to you guys.
SPEAKER B
Maybe Vinny could name something and I could name something. What do you think?
SPEAKER C
Well, for me this is because it's front and center on my work, but working with ecosystems, massive blockchains, because as an AI project, integrating with them is actually very different than if you were a Defi protocol integrating with them. If you're simply just adding a token balance integration or a swapping integration, or earning or yield borrowing, it involves their data on chain data, data that they want to ingest into the knowledge base and find really good output from that. And it's a lot more collaborative, but I also think it's a lot easier, too. So, yeah, these conversations are going really well, and it's also very unique. It's a unique way to integrate. And even the token audit. Right. With BSC, it's like a great value add. It's obviously helpful for a lot of people, and it's something that anybody could use. So I'm most excited about working with ecosystems and different blockchains.
SPEAKER A
Yeah, I can understand why this would be big for you. Of course, many new eyes to your project. When would you say that? Is that coming for, or can we not know?
SPEAKER C
I mean, it started, we'd announced the air weave integration, but there's blockchains that have a lot of different grant applications. They're relatively easy to get to. Previously, when I worked in DFI, I had the luxury of integrating with about 13 of these blockchains because I managed a simple arrogator. So we were just arrogating different d five functions, and all I was doing was just reaching out to different blockchains, getting grants. So I got a grant from Moonbeam, from polygon, from near, and become good friends with a lot of these blockchains. So it's really just a matter of reaching out to them and speaking to them and figuring out what kind of integration they want. So that process has started already.
SPEAKER A
If the project is good, you're right. You just need an entry or an open door, and then the project speaks for itself, like the quality of the team and the work. But, yeah. Thank you, guys. I'm really excited about what your project can do and what it can do for me as a user. How much can it improve my life as someone who's. Because in crypto, you need finger on the pulse 24/7 even if you miss one day, if it's the wrong day, you miss out on an entire narrative change, for example, and you're stuck behind.
SPEAKER C
KING well, speaking of finger on the pulse, you mentioned that you were tracking meme coins earlier and you're trying to be fast, right? What are some ways that you're trying to keep the finger on the pulse when it comes to trending tokens? Just curious.
SPEAKER A
So you know how the information, like we said, starts on the alpha groups, then there's biz on fortune, then it goes on Twitter, then on the telegram groups. I basically like to see which one makes it through, if that makes sense. With the meme coins that start on biz and end on biz, those will never make anything. But as you see, a meme coin start gaining traction. Like, oh, this one I saw on biz, now it's on this private group, but they don't know about fortune, so it means there's already, like, word to mouth going. And the more you see it going through the pipeline, you can tell, like, wow, this one maybe could have, like, a two, three, four, maybe 8 million market cap. It's at, what, 80,000 and now possible. Of course, it doesn't work 100% of the time, but I think an AI could refine this way better than. Because I do it all myself.
SPEAKER A
Yes, a tool like yours, I think, because you could also analyze metrics in real time, like how much volume, how many transactions, how many buyers, how many whales there are, how many people. Like, insane, totally different world, and faster, too, than I could do it. So, yeah, very excited. Very excited.
SPEAKER C
Yeah. It sounds like you have, like, a qualitative strategy. So you're actually putting boots in the ground. You're looking through these forms, and you're looking for trends across different publications and platforms, right? So it sounds like a very qualitative one. Actually, the reason why we mentioned alpha like bots earlier is because a lot of the community actually go more quantitative. So they're looking at bots that just tell them numbers that are trending on deck. Screener. These bots are integrating on chain data. But what we're trying to do is that we're actually trying to incorporate all modalities through AI, because we all have our own methodologies of looking for alpha. For me, I look for Biz, I look for little ideas, then I'll validate that on chain or through Dex screener. But you're more qualitative. Totally fine. But if we can merge dex data, on chain data, smart contract code, off chain news, obscure forms, then essentially, you're able to connect very distant, related piece of information that would just take so much time and energy to piece yourself together. It's just not.
SPEAKER A
By the time you do it, you miss on the pump, to be honest. By the time, the time it takes for you to do it, I want to specify the way. That's how I do it for meme tokens, because I feel like the meme has to be funny and it has to be memeable. People want to have to see it rank high just for the lulls, just for anything, for trolling. So I think for meme coins, yeah. The way I do it is there of course, for a serious project be different, but for like a longer hold this, I'm talking for like a week hold to the most week, ten days hold. Like we're having now.
SPEAKER C
Another data source we want to integrate is these telegram caller groups, right? These alpha groups. If we integrate that, then essentially we're not replacing them, but we are democratizing them, we're distributing them and then connecting them associated to other data sources.
SPEAKER A
It's going to be interesting to see how it changes when you have a huge user base using your tips, how it moves, the token itself. I would be very curious to experience it. But yeah, I'm very excited. Guys, seriously, I would love to have you guys back on the show when you want to share more news or new developments. Or we can maybe do a test live once. I can have one of the tiers and we could do like a meme coin trading session or something. Maybe like a longer hold, see how it goes. We can do an experiment. Tell me.
SPEAKER B
Yeah, by the way, if you want to test it out, you don't have to be a holder. You can just open our telegram chat. And in this chat, the basic features of the bot are available to everyone.
SPEAKER A
Okay. And the telegram link is on your website. I'll share it on my discord too. It's Alpha keg underscore chat on Telegram. Right? Amazing. And yeah, we'll do some tests, we'll do some fun content because the way going back to the ICU thing, it can make content too. So I'm curious to know how I can apply it to my content making while offering good and real time Alpha. Very excited about that.
SPEAKER C
We'll add haikus to the roadmap later after this.
SPEAKER A
That stuck to me. It blew my mind. But yeah, guys, thanks for coming in. Is there anything else you wanted to share before we end this ama?
SPEAKER C
We'll see you in the telegram and yeah, come play with the bot.
SPEAKER A
All right. I'll be sharing the link in the next hour or so as we end the stream and we move on to discord.
SPEAKER C
Thanks, Kang.
SPEAKER A
Thank you, guys. Thanks for joining. Thanks, Vladimir and I hope to see you guys soon on the stream. Means we're both doing well.
SPEAKER C
Sounds great, man.
SPEAKER A
Next time, have a good one.
SPEAKER C
Take care.
SPEAKER A
Bye bye. All right, guys, I couldn't follow the chat much. Forgive me, because the video was taking the screen I saw the spammers. I couldn't do anything about the spammers. Sorry, guys. I know some of you got the value from the alpha, from the project. From having two knowledgeable guests about crypto and AI. Both. Thank you, Jimmy. Thank you, Kangal. Thank you, chicas. Thank you everyone. Thank you, Natalia. Thank you all the people that stayed. I have to catch up with messages on the discord too. I disabled the notifications during the know. So yeah, I'll close it now. I'll just remember guys, the project is that it was Jimmy, I think that shared the discord link or their Twitter and their telegram link. You can try their AI for free on their telegram. The basic tools. The whole team is docs. So feel free to do your own research and get your peace of mind. Roadmap very extensive as we saw. A lot to be excited about. And they're pretty much everywhere. So choose your community, guys and join them. But as far as the cancers go, I'll see you guys on discord in a bit. We'll share all the links, we'll ban all the spammers and we plan for the next stream. I love you guys. See you guys sooner. Sa.
About AlphaKEK AI
AlphaKEK is an AI lab that seeks to be an indispensable AI infrastructure layer for Web3, helping users better navigate the space and improve their financial decisions. Their B2B solutions currently power Web3 tools and applications using a next‑gen, unbiased AI model and data engine designed for crypto.
Website | Telegram | Discord | CoinGecko | CoinMarketCap
