Private AMA: MMKK

Private AMA: MMKK
audio-thumbnail
Recording
0:00
/5160.733

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


I. Introduction

A. Cuzzo hosts AMA with AlphaKEK AI team, an AI lab with various products.

B. Cuzzo is impressed by the team's credentials in astrophysics, astronomy, and AI.

C. The more Cuzzo looks into the project, the more blown away he is by the team's qualifications.

II. Team Introductions

A. Vinny:

  • Has a background in traditional finance, fintech, and has been working in DeFi/Web3 since 2020.
  • Early builder of AI in DeFi.
  • Helped develop a machine learning platform to analyze crypto wallets.

B. Vladimir:

  • Founder of AlphaKEK with 9 years of AI research experience.
  • Has worked on AI projects across various industries.
  • Presented a generative AI network for computer graphics at the NVIDIA AI conference.
  • Has experience in astrophysical research and natural language processing for multi-messenger astronomy.
  • Was an early tester of OpenAI's embeddings.

III. AlphaKEK's Technology

A. Built a custom AI engine specifically designed for analyzing crypto data.

B. Developed unbiased, uncensored AI models, unlike the limitations of GPT, ChatGPT, and Claude.

C. Constructed a custom data infrastructure:

  • Collecting data from both Web2 and Web3 sources.
  • Building the "AlphaKEK Fractal" knowledge graph.
  • Analyzing data from a broader set of sources compared to competitors.

D. Focused on providing qualitative insights, not just quantitative data.

E. Took a long-term approach to building a robust infrastructure.

IV. Monetization and Business Model

A. Implemented a hybrid cloud solution for scalable and cost-efficient operations.

B. Offering B2B APIs to integrate AlphaKEK's technology into other projects.

C. Integrating with ecosystems and blockchains, such as the Arweave integration for their technical content.

D. Compared to PAAL AI, which focuses on community bots, AlphaKEK aims to be the primary infrastructure provider for the Web3 space.

V. Future Plans

A. Short-term:

  • Integrating with various ecosystems and blockchains to onboard more users and educate communities.
  • Scaling their API offerings to support other projects and tools in the market.

B. Long-term:

  • Developing the "Alpha AGI" autonomous agent capable of multi-step reasoning and sentiment analysis to provide comprehensive crypto insights.
  • Becoming the default AI infrastructure for the Web3 ecosystem.

VI. Conclusion

A. Cuzzo is impressed by AlphaKEK's potential and looks forward to the project's progress.

B. Cuzzo plans to host AlphaKEK again on open VC for a more in-depth discussion and demonstration.

C. Cuzzo encourages the audience to explore AlphaKEK's offerings and stay tuned for updates.

Transcript


Note: The transcript below has been truncated for for greater legibility.

All right. Good morning. Good morning. Good morning. And welcome to MMKK. I am Cuzzo. It's really early here for us. In any case, today's AMA, I have a feeling we might just be hosting the team that can actually be considered quote unquote experts in AI language model bots. However, you want to define that. Of course, the project we're talking about is AlphaKEK AI.

First of all, I'll say this, that I don't know where you guys have been hiding. When I started looking to the project, I devoted a little bit of time yesterday and I'm looking at it. I'm like, okay, project launched last year, doxxed team with credentials and contributions to NVIDIA and OpenAI, like actual contributions in the field of astrophysics and astronomy. So I was blown away by those credentials. And so that made me really interested in get on here today and talk to you guys about exactly what you guys are doing. I simply put AlphaKEK, from what I understand it, it's an AI lab with a host of products. And of course, we'll get to all those products. But I'm dying to at least get the proper introduction of the team. Vlad and Vinny are here. Guys, if you don't mind, we would actually like to have a detailed introduction of your background, your credentials, and your contributions to AI space. If you don't mind, let's start us off with that.

Vinny: Yeah, absolutely. I think Vinny can start because I could do this like for hours. Yeah, Vladimir is the star of the show in terms of AI, so I'll keep mine a bit short. But yeah, I'm Vinny and I manage marketing and business development at AlphaKEK AI. And yeah, thank you for hosting us. It's a very well curated, very private group. And thanks for waking up early. I'm actually based in Asia. So thanks for the flexibility. Like you mentioned, Cuzzo. We're an AI lab. And we basically built a custom AI engine that's specific for crypto data. We basically built a custom AI engine that's specific for crypto data. And our team is composed of obviously Vladimir on this call, as well as four other core contributors, including myself. And yeah, we're essentially a culmination of his life's work as a data scientist and AI engineer. Myself, I've been in TradFi for about 17 years, fintech as well. I started working in DeFi and Web3 in 2020 during DeFi summer. And I consider myself an early builder of AI in DeFi, pre-ChatGPT. I helped develop this platform that uses machine learning to essentially analyze wallets.

Vladimir: Yeah, so I'm Vladimir Sotnikov, the founder of AlphaKEK. I'm an AI researcher with 10 years of experience now. I have trained my first machine learning model like, oh no, nine years of experience, yes, sorry. So I have nine years of experience. I have trained my first model in April 2015 actually. it was a simple like logistic regression model that was detecting number that was recognizing numbers on black and white images and over the last seven years I was working on various like AI projects in different industries I was working on large-scale projects in different industries I was working on large-scale computer code analysis scraping terabytes of data from github and building AI models on top of this figuring out how programmers write their code then I was involved in a lot of AI graphics research I have presented. My generative AI network for computer graphics on NVIDIA AI conference in 2021, basically before Stable Diffusion, daily and all this stuff. It was generating both images and videos um you can check the video by the way on our uh website and after that i was working yeah a lot of uh on astrophysical research including uh both like images image recognition from for um telescope arrays basically like detecting air shower cosmic ray data and i was working a lot on natural language processing for astrophysics i have like i basically created one of the very first language models for multi-messenger astronomy and while working, while leading the team that was working on this, I was one of the first private beta testers of OpenAI's embeddings, so basically I helped them, evaluating them and they even mentioned this in their blog post.

Cuzzo: I saw, you're one of those people that probably early in this space with AI interactions or research or contributions, you know more about probably this stuff than any of us do. As Vinny was talking about, there's so much LARP going on in this space. So many people claim to use so many different language models, right? LLMs. And they're biased at some point, not really original creation. There are some huge projects in space that are attempting to do something. But the more I look into what you guys have, and I think it just might be the case of discovery on investors part because the research and contribution in this field by experts in genius quote-unquote i mean you know um i guess i can call you that uh and but applied it to a crypto space and what degens do on an investment side of the things and just waiting for catalysts for something like this to blow up i hopefully that's the case i am kind of glad that this is the first project uh early morning that i am looking into with you guys together um uh uncorrupted in my mind and ability to kind of comprehend what's what's really going on with this project um there are a couple of things that I saw on the website, especially with the GPU servers, okay? You guys already know what I speak of and how many LARPs and mega caps that get launched with that language on their websites. And whether they do it or not, that's entirely different. You guys have actually been doing it for a while. And this is one of the reasons I feel comfortable speaking about this. I'm like, wait a second. They've been around a block for a while and this is one of the reasons I feel comfortable speaking about this I'm like wait a second they've been around a block for a while and you guys attended the conferences as I saw that i saw the articles I saw the your field of research and then you have a business partner here Vinny who's doing your marketing and and now you guys starting to bring it out so I think what we should should focus on what you guys are expert in, what problems you guys trying to solve, right? And how are you monetizing that so your investors can gain value? I think if we can concentrate this AMA on that, I think it will shed a lot of light on your project and get you going where you need to go. But I mean, at the end of the day, $3.5 million market cap, it's nothing small. A lot of projects can't even get up there. So it's going well. So why don't we talk about what problems are you guys solving? And how are you guys monetizing that? I think it's a similar conversation. Whoever wants to go first and say, okay, well, we have this. The other projects do not. People out there looking for this product, they're unhappy with what they had before. They're willing to pay for it. And this is what we could provide, whether it's custom data, custom AI models, LLMs, or however you want to define that whoever wants to go.

Vinny: Sure, I can take that. The first problem, and maybe I can talk from a personal perspective, the problem that AlphaKEK solved for me when I was just a regular community holder, was that it was the only platform that had an easily accessible, uncensored, unbiased model for me to actually generate the content that I needed. I was copywriting for a major traditional finance company based in Singapore. And I was writing a lot of niche financial topics. And I wanted to actually get alpha from 4chan, from /biz/, because I frequent there pretty often. A little bit of our lore, that's actually how Vladimir and I met. We have a lot of friends that are admins of /biz/, and it was a very easy place to start AlphaKEK and ingest that data because there's an API. And that branding ties into that as well.

But the problem that I faced was I couldn't find an uncensored model to actually discuss the financial topics, crypto, that I wanted. Claude-1, Claude-2, and Claude-3 would reject. So I think the major problem in this space is that many AI projects are simply GPT wrappers. And there are just heavy limitations when you're using existing AI tools. So not only in Web3, but in Web2 as well. GPT-4, Claude-3 will outright reject a lot of the input about finance, let alone crypto. And a lot of products, again, Web2 and Web3, whether it's like chatbots, analytics dashboards, telegram bots, alpha bots, alert bots, they're built on these. And it's unclear how AI is actually incorporated into these products. And we've talked to many Web3 projects who are utilizing these popular APIs, and they've said that the results are very generic or worse.

So we realized there was a gap in the market for custom infra for crypto, and we just quietly the last eight months developing this infrastructure, which we hope is going to be a new AI infrastructure standard for Web3. So we already mentioned the unbiased model, the data engine, which we call Fractal and you know we've integrated that with a lot of tradfi and crypto data sources from mainstream sources like CNBC, Bloomberg, CoinDesk to very obscure.

Vladimir: Exactly. So one of the key things is that blockchain data is immutable, while Web2 data slowly dies. Websites are getting more and more closed to scraping, etc. So one of the reasons we started collecting data early on was to be able to backtest and backtrack it in retrospect for future analysis. For example, we could try to detect people who were consistently front-running the news based on their wallet activity.

Another thing is that we're slowly making our services available via B2B APIs. The idea is to use our consumer-facing apps as a storefront for our technology and also sell the API to other projects that could leverage it.

Cuzzo: I totally share that sentiment that if you're going to be considered an expert in a field or a flagship product or a project for that fact, you're going to have to take the long route and not cut any corners because when they're actually time to comes to monetizing it and delivering the product, especially when you guys were talking about in the field of blockchain data, how to monetize that because that is one of the big things, right? We have people actively looking at, you mentioned the word smart money. There is a huge field of how to analyze that data, what's really important, and then how do you apply that to your daily investments or trading, things of that nature. So the field is vast. I mean, from an investor's point of view, when a project can have products of this caliber and multiple products for that fact, monetizing it is going to be the easy part once you can actually demonstrate and deliver that product that everybody wants. Another thing we've been seeing lately is the Telegram chats moderation, how projects are incorporating that and everybody's willing to pay for it because it helps bring the project in the best of its light. You know, some of those things, I mean, I'm probably only going to mention a couple of them and Vinny, you were talking about the wallet trackings, et cetera. I also saw the volume comparison right here in the chat about the base volume and Ethereum volume. There's a lot of things that are going on. And so I guess what begs the question, you guys have been working at this for a while. Would it be fair question to ask, what is ready as far as what other projects can approach you guys to incorporate something for themselves whether it's via partnerships or not has anything uh that you guys selling it as a product to companies or projects right now?

Vladimir: Yeah, so basically we have two options. First is that some other projects integrate their own, project-specific data into our infrastructure and gets all our services automatically available for them.

For example, we recently integrated with Arweave, a deep Web3 protocol focused on decentralized storage. They have various websites and blogs with a lot of technical content, and we ingested all that data into our AlphaKEK Fractal. Now, their developers can use our question answering, sentiment analysis, and other features to quickly familiarize themselves with the Airweave ecosystem.

The second option is that we sell our services as APIs to other projects. We're gradually making our infrastructure available, like the custom AI models and the data engine, so other teams can integrate it into their own products and applications.

Vinny: Yeah, there are some things that PAAL AI has done well, like aggressive business development and marketing to onboard a lot of projects and communities. But we don't really consider that our primary market.

What we're more focused on is being an infrastructure provider, kind of like the OpenAI of Web3. We want to power the AI engines behind different blockchain projects, ecosystems, and products. So we see ourselves as more of a B2B play, where we can quickly integrate our API and our custom models into whatever the project needs - whether that's a community engagement bot, a research dashboard, or something more advanced like an autonomous agent.

The key is that our technology is built from the ground up for crypto, with unbiased, uncensored models that can handle finance and Web3 topics that typical LLMs struggle with. So we're aiming to become the default AI infrastructure for the Web3 space, rather than just another consumer-facing bot or tool.

Cuzzo: This AMA has become quite technical, which is understandable given the nature of your project. I think it would be great to host you guys again on an open VC, where we can showcase your capabilities to a larger audience.

In the short term, what can we expect from AlphaKEK? And what is your long-term vision for the project?

Vinny: In the short term, we're focused on integrating with various ecosystems and blockchains. These projects are often looking to onboard more users and educate their communities, and our AI engine can be a powerful tool for that.

We're also looking to scale our API offering. There are a lot of existing bots and tools in the market that are built on top of generic LLM APIs like OpenAI. We've already had success integrating with one marketing bot that was able to significantly improve the quality and contextuality of their output by switching to our API.

Vladimir: Longer-term, we're really excited about the potential of our autonomous agent, codenamed Alpha AGI. This is an AI system that can perform multi-step reasoning and detect changes in data sentiment to provide more comprehensive and up-to-date insights. It's a step beyond the typical chatbot or alert bot functionality.

We believe this type of advanced AI engine, built specifically for crypto, will be a key differentiator and infrastructure piece for the Web3 space going forward. And we're working hard to make Alpha AGI a reality in the coming months and years.

Cuzzo: This all sounds very promising. I'm really excited to see what the future holds for AlphaKEK. I think this project has a lot of potential, and I look forward to following your progress. We'll be sure to host you again on an open VC for a more in-depth discussion and demonstration of your capabilities.

Vinny: Thank you so much for having us and for your interest in the project. We're looking forward to the continued journey and to showcasing more of our capabilities in the future.

Cuzzo: Absolutely. Thank you, Vinny and Vladimir, for taking the time to join us today. I encourage everyone to check out the AlphaKEK website, join their Telegram group, and stay tuned for more updates from this impressive team.


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