Coinomi Connect with Fluence

Okay, just my check, my check. Hello, hello, okay, the microphone's working. And I think we got fluids also with us. Can you try to seek fluids just to make sure you're audible? Wait a minute. Hey, hello. Hello. Yes, loud and clear. How are you? Awesome. So you've got a hundred fluids. Thanks for having me. Yeah, nice to have you. Okay, so I'm just gonna give one or two more minutes to wait for more people to join in. And also to set up a bit of things in the background, but yeah, we'll officially begin in one or two minutes. Okay, I think we're good to go fluids. I'm gonna call you fluids unless otherwise you have a different name. It want me to call you. It's my name is Yvgeny. I'm not sure if you're if we are on that. Oh, Yvgeny. Okay, yeah. Okay, Yvgeny. Okay, let's just officially begin. Okay, hello and good day to everyone. That is currently tuning in here at Koi No Mi Connect. This time, I mean, this episode we're gonna be having fluids. And from fluids, we have Yvgeny from their team how we're doing today and welcome to the show. Hey, hi, everyone. I'm excited to be here. Thanks for your notation. Go founder of fluids and maybe to be here, answer all questions. Talk about fluids. It's awesome to have you here, Yvgeny, because this is quite an interesting project. And if I'm not mistaken, this is the first of its kind that I've had here at Koi No Mi Connect as, you know, the centralized CPU and GPU cloud platform. And of course, we're gonna be knowing more about that as we go through with this show. But before before anything else, Yvgeny, can you tell us more about the team behind fluids? Yeah, we are now pretty small team. I mean, there is a pre-co-founders. We are together building fluids since 2017. And it's been a long journey. You know, we had sort of team scaling from three people to 40 handbags. So right now, as AIAH changed the dynamics of you know, how you hire people or you hire agent or you scale operations with agents. We are trying to keep the team small. But you know, be experts in different areas and scale our resources with AI and LMS as much as possible. Yeah, and since 2017, gosh, time flies. And actually, you're about to be a decade old next year then. So you've been here for quite a while. And this is Oh yeah, it's been a long, long journey. It's been a different world 10 years ago. And you know, AI back then is definitely not as big as we have it right now. I mean, I think even five years ago, barely AI has touched a lot of the, you know, normal users here in the world. But now, nearly everyone is using AI. And unfortunately, AI has caused a lot of trends in the market. I would say negative for the most people that are using, you know, GPUs and stuff like that, increasing the price of, you know, storage and yeah, sort of stuff. But you know, what you guys, AI compute is now such a huge focus. So what do you think made this the right moment for fluids to go deeper into AI infrastructure? Well, yeah, it's actually funny fact that when we started, we started sort of the project from the experimentation around intersection of AI and blockchain. And actually, like exactly nine years ago, there was a big topic of, you know, machine learning and first big models and big data and this stuff. And we were thinking how blockchain can help AI. And then we spent a lot of time, you know, iterating the server products for WebG audience. And now we sort of last year around last year we focused back mainly on AI market because, you know, basically all the time we've been building some solutions around how do we run computing a decentralized way, how do we run verifiable compute, how do we run, how do we secure compute with blockchains. We build a network of different data centers globally, compute providers, you know, experimented with different designs and solutions. Some of this didn't work, some of them were better, you know, some of it depended on the state of crypto market, which we mainly serve as customers. And because crypto market, as you know, is very volatile. When the bull market, a lot of people build in, they want services, they want infrastructure. When there's a beer market, like a lot of people gone, they just, you know, out of money for their own operations. So, we, when we found quite a lot of traction with AI customers because right now AI market is so big, it's kind of bigger than crypto already. And it was just organic for us to focus on serving AI. I guess that makes you one of the, you know, the first ones to be developing that tech, especially during that time. You barely have any competitors back then, but I think the reason why back that AI is not that, not that popular is because not a lot of people knows what AI computing is, actually is. So we're just AI in general. I mean AI artificial intelligence is one thing, but the actual AI that's getting popular right now, it's a bit different from what we, you know, what we have right now kind of jumbled up the words right there. But I guess when people hear the word decentralized compute and immediately, immediately people think it's complicated. I mean here in the crypto space, decentralization is a word that we just toss around, but decentralized compute is something different. Yeah, it sounds complicated. So how would you actually explain fluids to someone who understands AI that we have right now, but knows nothing about decentralized infrastructure? Yeah. So actually like decentralized can mean bunch of things and it's you can, you can sort of deliver decentralization. We have different solutions, but we lately try to even get sort of distant from this decentralization term, because as you write, it's a bit a little bit confusing for wide audience. So it's just aggregation of resources in our case. So we have a lot of connections to hardware that's actually sitting in data centers globally in professional facilities, and we have water relationships, they all integrate it into our platform. So whenever you sign up and use our platform, you can deploy virtual machines like virtual servers. You can find GPUs. We can help you set up the GPU cluster for big workloads in very, very different locations. So like right now, in our case decentralization means this geographic distribution. The second thing is that as we are in crypto audience, we have, sorry, in crypto market, we have the blockchain part of the project, which is right now represented in terms of the token, the governance model, and some staking incentives. And the governance is decentralized. So there is an on-chain which can vote for certain decisions for protocol development, for treasury allocation, and so on and so forth. So right now, the state of the project is that we decentralize in terms of geographical distribution of the compute, and we decentralize in terms of governance and having the token on blockchain, I guess, a little bit of decentralization as well. Okay, we're going to be talking more about the token in a little bit later, but for now, I want to focus on the tech that you actually have right now. And actually, one of the benefits that you guys claim when we use fluids instead of, you know, compared to decentralized cloud platforms. So yeah, one of the claims is that you guys have is you can say about to 85% compared to, yeah, as I mentioned, centralized cloud platforms. So that is huge. Can you explain to us where does that cause difference actually come from, and how does it actually make sense for you guys and us potential customers? Yeah, that actually comes from, you know, just a pure server hardware economics, and we, like there's a little bit of obviously of marketing on, you know, these numbers and discounts, right? But then if you just compare some CPU server, some virtual server price on big cloud, convenient platforms like Amazon, Google, or Microsoft Azure, compare to our pricing, you would find that all pricing is much much cheaper. But it's just not because we do some, some entrepreneurial magic here, it's just because big clouds charge a big margin on the hardware. So because they have sort of monopoly in the compute market, they offer lots of services to enterprises, they log them in into multi-year contracts, they can charge quite a lot of money for pure hardware. And so if you compare apples to apples like Amazon EC2 instance price to the same server configuration and output for an output for more would be much much cheaper. But then if you compare, for example, to other bare metal or virtual server providers that are smaller than Amazon, obviously the price difference is going to be lower. But as we offer not only the CPU servers, we also offer GPUs, and for GPUs it's a little bit different world. Also Amazon and Google also offer them at highest prices. And we aggregate them from a bunch of data centers who are ready to random out with the minimal margin for themselves. So we can always find the best price for like that exists right now from the GPU hardware that stays idle in some data centers and offer it to you. So that's the nature. So the nature of this price difference is mainly because how this big clouds said their prices because of their monopolistic position. And that is why we definitely need decentralization not just in crypto, but especially in markets like these because you know they're big, the big players. So we'd call them are the ones manipulating a price. People have the illusion that they're the only choice that they have. So they're forced to just adapt to whatever price they put on all the services that they give to us. And unfortunately for a lot of companies, for a lot of people, those kind of services are crucial and important for their day to day life or work. So they have no choice but to follow that. But you know they're not, they don't have the they don't know that there's this kind of alternative in order for them to save money or to save their own resources. And I guess lots of information campaigns are needed. So people will be well informed about this kind of stuff that not just Google or Amazon or any other big players that can actually provide this kind of services are the only options that they have in terms of yeah, this sort of stuff. And something interesting that you guys have if I'm not mistaken are what you guys call is the GPU cluster options. And it's interesting because instead of simply buying capacity, customers connect.