Arkstream Capital: Merging AI and DePIN in Web3’S Tapestry
Last Friday, we hosted a space meeting with thought leaders from Ionet, Ator, Myshell, BSCNews, and MH Ventures. The session centered on the convergence of AI and decentralized networks (depin) within the Web3 ecosystem, examining the roles these technologies will play in shaping its future. We discussed the implications for AI models, crypto AI, AI projects, and applications, alongside strategies to navigate challenges like centralization and data security. This meeting was a deep dive into envisioning a collaborative, decentralized future. Here’s a recap of the compelling insights shared.
Chung @Investment Director of Arkstream Capital
How do you think best for web 3 and AI or web 3 and DePIN in should be combined and what are the narratives that excite you guys?
Garrison @CSO/CMO of IO.NET
Yes, I’ll try to give you a brief. I think that we will look back in 6 months and realize that compute DePin and many different types of hardware DePin are very good ways to launch blockchains, right? I think that the traditional way of launching a blockchain by bootstrapping your validator set by only inflating a token is a really tricky model. And I is not a little bit less sustainable than using a deepen, whether it’s a compute deepen or otherwise, to bring external revenue from outside of the protocol to incentivize no growth.
The second is I think that DePin, especially compute DePin, are really important in the web 3 space because in the, our race to launch a bunch of blockchains to compete with each other, we have started to over rely on centrally hosted validators. And so you have some really interesting dynamics where your sequences, your validators, your app chains, your approvers are all hosted by to recentralize cloud providers and it’s necessary for rapid growth and development. But at some point, we’ll wanna get back to being through decentralized, and that’s somewhere where deepens plus blockchains together can help achieve that.
Zengyi Qin @Co-Founder of MyShell
Yeah, I wanna talk about the AI creator economy. So, you know, they’re, there are millions of AI model developers and the app creators. And so their app after created, they face a very long way to monetize their AI applications. Cuz they need to invest in the model training and model deployment. So, but I think the web 3 can come inside into the investment side. For example, at MyShell, we convert those AI models and AI applications into tradable assets so that basically they can be traded in the secondary market. So we can bring the web 3 users to invest and support the developments of these AI applications and help the app creators to monetize their apps. And we have seen many interesting applications on the platform created by those fantastic creators like the AI language tutor or the resume reviewer or other entertainment oriented like a virtual boyfriend and virtual girlfriend, a wide variety of AI applications is coming out. So in order to support their development, the web 3 has a very critical role in it.
Neuratic @Co-Founder of Ator
I think now where we’re in the cycle, it needs no introduction and slightly more controversial statement. But over the past few years, I personally feel that web 3 has lost its way a little bit in terms of building things that normal people actually care about and with deep in and the ability to allow anyone to contribute to a network, I think people get naturally excited at that prospect. And absolutely, as garrison from Ionet into that building sort of a crypto currency around the use case where the ebb and flow of tokens naturally tends with real world demand as opposed to, you know, building sort of a new blockchain around one and hoping the use cases build on top.
But that makes a lot of sense for the long term. And you mentioned AI as well. I mean, what excites me also is the power of combining AI with these other top narratives of deep in privacy. And so absolutely, there’s no introduction at this stage. It’s, it’s a well known term and for very good reason. I do think we’re just at the foothills of what deep in can do.
KDOT @General Partner of MH Ventures
So the question being, you know, how do we think web 3 and AI should be combined? I think we’re already seeing that like, you know, people or devs that are building AI related tech or Lmms or and stuff like that, they’re having to use GPUs for the infrastructure. So, you know, projects like ion net are gonna be perfect for in this industry. As, you know, devs come into to web 3 and I look into get onto the scope of AI and having to think, okay, the provider is going to, I own it. And as they become, and as I said earlier, becoming more decentralized, it just makes sense.
So in terms of the narrative that excites me, I think both narratives being DePin and AI are super exciting, especially right now. I think we’re just at the infancy stage of what is to come. And as we’ve seen with every other narrative, as we, you know, being in the past, it just become bigger and bigger and the tech just gets better and bad. But yeah, I do. I just think that right now we’re just gonna see a influx of more developers coming in looking to build infrastructure. And that’s what’s gonna be super exciting to see how it unfolds in blockchain. So I just don’t think people are really seeing the bigger picture, so there’s a lot more to come.
Chung @Investment Director of Arkstream Capital
Absolutely. And just to follow on like cam, like wanna to know, so as an an an as an investor, I think you’re one of the very few in the panel that is a VC. So yeah, so just after that, yes, Sean need you can go after that. So I wonder from a business perspective, if like, what are the criteria you’re considering when you’re evaluating a AI project or the pin projects? Like what are the kind of considerations that you’re taking at right now when you’re kind of seeing these projects, for instance, out on that or ator, I know you’re the back of our dot net. So also curious on what kind of attracts you on that ongoing into them and also maybe other deep in projects, how do you evaluate them?
KDOT @General Partner of MH Ventures
Yeah, appreciate that. I’ll make it quick. I think the biggest thing is when evaluating any project in a certain criterias, and especially in deep in is like how we’re changing the aspect of the current web, the model that it is, let’s say in web 2, and what they’re solving and what type of scale it is, or is it on a minute scale or is it a bigger scale that we believe that, you know, they actually gonna be the a top ecosystem within web 3 and iron at being this GPU provider where, you know, actually big corporates are look able to, you know, grab clusters. That’s gonna be huge in there. That’s gonna be, they’re gonna be, you know, able to scale. They’re not looking for the individual users on everyday basis, actually looking for developers that on needing to have thousands of clusters or GPUs, I should say, to use. So when we’re looking at investments, we’re looking at a higher scale of if they can change the landscape of what they’re trying to build.
Jonny @Chief Editor of BSCNews
Yeah, absolutely. I think almost just to zoom out a little bit, one thing is that 2021 definitely told me, and I obviously told us a number of things, but I think a big one is that you aren’t just slap and emerging narrative onto web 3 and expected to do well and perform over this sort of mid to long term. So I think looking at crypto AI, web 3 and AI together, it’s really the projects that do leverage blockchain shaped advantages and map them onto AI that will do well. So I think, you know, to hawk on about iron, that again, I think is a really good case of, you know, you’re taking excess capacity and you’re giving it at more attractive rates to parties that really need it. Effectively, you’re marrying web 3 decentralized technology with the real sort of web to AI problem. And that potentially can create a really interesting flywheel. So I think for sustainability in crypto and AI or with the within, you know, crypto and AI projects, the value ad needs to be very tangible and you need to somehow leverage blockchain shaped advantages onto what is at the moment a predominantly web to sector.
Chung @Investment Director of Arkstream Capital
if we’re thinking about educating or onboarding users who are new to the crypto space and getting mass adoption in some sense, maybe in the next, maybe few years or in the next cycle? Like what do you guys see is the kind of, what is your plan to educate on onboard new users to come to the crypto or AI space. Like for instance, ator, like do you have any few points on how to get true mass adoption for users to come to this space in your project.
Neuratic @Co-Founder of Ator
The one thing ATOR has and has really started growing, especially in recent months, is an active community that are hungry to run nodes in the network and contribute and educate others. And I think developing that kind of culture, as difficult as it might be, is definitely a very powerful first step when you start to see an environment where people can kind of onboard others, teach others, you know, take in their exceptions or their case and turn that into education for more people. That’s when you really spark a movement, especially within this web 3 space.
If we take a step back and think to the broader question, okay, how do you onboard not just web 3 and crypto users into your protocol, but actually they’re sort of crypto curious or completely non crypto native into web 3 and into deep in.
That’s when it gets into a more interesting challenge. And one of the ways we at ATOR have of going to tackle that and lower that barrier to participation has actually been through hardware. So we’ve taken some inspiration by the immense success of helium in the previous cycle, getting close to a million different distinct hardware units out as a way to onboard users. Indeed, the co founder of Helium, Sean Kerry, recently joined our advisory board and next month, we ship out the first sort of units 0 of our own hardware. So you don’t need our hardware to run a node in the network. But if you’re a non technical user and you wanna sort of plug and play solution that’s there for you. And we think if you combine something with a very nice, easy to use hardware device and pretty simple narratives, you know, this is a crypto mining that you can do that you don’t need tens of thousands of hundreds of thousands of dollars of investment. This is a network you can contribute to, even from your own home or even as a single individual. You combine those simple narratives with a easy to use sort of onboarding process, that’s certainly a great way to get that ball going. And in general, releasing documentation education aimed at different skill levels, but both from enterprise customers all the way down to the non technical individual. Well, that’s, you know, no better start than that, attacking the market and growing and deep. And that’s particularly true for us at 8 or because it truly is the case that anyone with a home instead connection can contribute to this deepen and earn rewards.
KDOT @General Partner of MH Ventures
So from a fun perspective, like, you know, for us, I would say is mainly that we’re gonna invest into projects that we feel are going to, you know, open up jobs. So that means that people are gonna come into the space and that could be from a death perspective or what ever role it may be. So opening up, you know, opening more jobs for people is one. And then two, the tech that’s being built is gonna be use by users. And then talking about the tech, you know, cool stuff, I feel like AI is gonna be a big thing that actually Bing brings in a lot of new people. And I say that because if you look from a user’s perspective, you know, a lot of the stuff that’s being built now that I’ve seen and stuff that’s gonna make people, you know, able to easier to trade, you know, talking about. So if you look at like, if you look at, I’ll just track back a little bit. If you look at like ChatGPT, the influx of users that they had when they release their first model was insane and they onboarded so many people. So now people have this vision of what AI can be, and people are still learning of that, how advanced it can get. But now we’re building that equivalent inside of web 3 and on the blockchain, which now we’re able to bring in more people that are, you know, don’t know web 3, but they’re gonna be able to use models that is gonna be able to explain how to do things and they can command stuff on the AI model. So let’s say for example, they wanna create a new wallet, but they’re scare to, you know, have their private keys or they don’t know how to make a meta mask or anything like that. They’re able to probably tell this model to do that. And then they can also tell a model to buy and sell tokens or sell tokens at a certain time or a certain action. Let’s say for example, Chung, you had, you know, they had your wallet and your wallet was tracked and said when Chung buys his token, dump on him, right? And that could be a command. So I think things like that are gonna make it easier for, you know, people that are want to invest a bit scared, but something like this would allow them to because it just able to have parameters of I’m looking from a user perspective, but I feel like that’s what opens the door to a lot of more people. I wanna get into the industry.
Chung @Investment Director of Arkstream Capital
So MyShell, I also wonder, do you have any views on that when it comes to mass adoption, how do you think yu, your project can kind of onboard users pre and new to the crypto space? What is your thoughts on that?
Zengyi Qin @Co-Founder of MyShell
Yeah, so my overall understanding is that if you wanna attract a lot of web 2 users into web 3, you first wanna build some really useful and interesting applications to attract them. So that’s why MyShell is focus on building, helping the creators to build many useful applications like the it, yeah, tutors and other resume. We will President May reviewer and the virtual interview and also all kinds of multi modality applications such as the video generation and image generation and in image editing. So all kinds of utility are something that both web 3 and web 2 users are looking for. So building this can attract the users onto the platform.
And also the another perspective is to build up the company’s reputation in the develop. For example, we build an open source many foundational models. For example, our open voice cloning model was ranked global top 1 on Github and that attracted a lot of users to use MyShell’s product. And also believe that MyShell is a technology oriented company and can contribute to the open source community. So that’s from the UH2 perspective, from the product side and also the technology side. Do something utility first and then protocol second. That’s not.
Chung @Investment Director of Arkstream Capital
and husky, like, do you have anything you wanna share with us in terms of how to, how do you out on that attract newcomers or new users to this space? I know you guys are generating, getting a lot of hype, except, especially first within the web 3 communities, but on the wider mass market, I wanna to also take your view on how you guys getting the user to come in.
Garrison @CSO/CMO of IO.NET
Yeah, the best way for any technology, whether it’s crypto or something else to drive adoption, I is just to deliver value and be useful, right? It took consumer AI a very long time to become popular until ChatGPT and rendering images and videos actually became usable. It’s the same way for something like a deep in, right? If we can’t actually deliver real value for node operators and providers and if we can’t deliver real value for companies, whether AI or not through a compute network, then what are we doing, right? So, so the way that we drive adoption is basically by providing users and companies with the product that they come to expect. And today, we focus very much on the supply side, right, building out a way for node operators and graphics cards owners to monetize. And in the near future, and especially through partnerships today, we’re providing that capacity in a way that’s usable for AI companies. Everyone’s happy.
Chung @Investment Director of Arkstream Capital
do you have any advice for them in terms of navigating the AI and deep in landscape effectively. Like, do you have any advisor for them? Maybe they’re trying to, I’m not sure, some opportunities in a more practical sense.
KDOT @General Partner of MH Ventures
Yeah, hundred percent. So from a DePin aspect, but there’s a lot of overlap. But from a deep in aspect, you know, for retail investors or people that are looking in to get into the industry, there’s so many opportunities that are gonna require like near enough zero cost.
You know, the thing about deep into the cool thing about is that the protocols actually incentivizing the users to help validate the ecosystem, right? So we can use, for example, Hive Mapper, you know, you’re able to go around and with this webcam that they provided as a software to go around and help build out this map, a decentralized map. So there’s so many different things within in the in this industry that you can take part in and actually get rewarded and earn real money in return. And that would just broaden your horizon on educating yourself of how this ecosystem works. And that’s from, you know, from a deep in level, but then also just from a web 3 level. So if you’re looking to get in and you know, yeah, you don’t wanna take too much risk, then I think this is a great opportunity. You’re definitely still early. Everyone’s still building out the infrastructure. And then from an AI perspective, you know, you just use that layer on top as well to navigate through. Cuz. So it’s a lot of cool stuff that I’ve been seeing that’s being built that’s gonna help a lot of retailers just, you know, ease of use for trading and anything else. So yeah, that’s what I would say as per advice is, you know, there’s a lot of opportunities that are gonna be a lot of like really low risk as well.
Chung @Investment Director of Arkstream Capital
MyShell, are there any key short term milestones that you think listeners should look forward to? Like what are the kind of very short term plans or some of the plans that you guys are having in mind for the maybe next one to three months, listeners should be very excited about like, maybe you can start with Husky.
Garrison @CSO/CMO of IO.NET
Yeah, so we our currently pre token, our token launches at the end of April with the token launch, like PE, most people are saying here, it unlocks a bunch of opportunities to get involved with crypto, economic security and shared incentives for network growth.
Many of you may have noticed that Io dot net also operates a gen AI platform that is 100% run on the deep end called BC eight on AI. And if you’re paying attention, you would notice that BC eight inference infrastructure is now running on App DOC. And so there are some really interesting things that we’re doing at the intersection of using modular move with scaling AI inferences.
Chung @Investment Director of Arkstream Capital
Understand. So, so basically, hum for in terms of how I understand token knowledge is very soon. So if I’m looking for our jobs, I expect to get engage and also notice what you guys are doing. And also there is this BAC platform on the modular blockchain partnering with apple. So that’s something that I should be paying attention to. And for a tall, what are the key short term plans that you guys are having in the next maybe couple of weeks?
Neuratic @Co-Founder of Ator
There you go. Yeah, I was gonna say short term and one to three months. I think I’ll run out of breath going through that. So let’s just talk about the next couple weeks for a tour. We truly are on the foothills of a really exciting few weeks in a really exciting month with our long awaited hardware that we first sort of distributed to some community members at East Denver will finally start shipping out for those who pre ordered their units in December.
Imminently. We’re really talking in the matter of those weeks to our community. And with that, very shortly after, we actually have the first version of Decentralized Network also going live. And that’s coupled with an education series, that’s coupled with all the tools needed to start running relays, and I’d like to think that it’s a chance for us to see the pent up power of our community by having them run relays again, and I’m really excited to see just how far we can take this first version of our network, just how many nodes we can get and how diverse they can be around the world. It’ll be a really exciting and hopefully fun time for anyone in our community, and we invite anyone who wants to get involved and be incentivized to do so to come check us out in the run up to that. Cuz very soon we’ll be running relays all around it.
Chung @Investment Director of Arkstream Capital
Actually. So basically, product is going to be out very soon. I’m gonna be shipped. Network is like going to be launch very soon then. So just some pay attention to that. And I remember when we were trying to buy ATOR OTC, the price is maybe 30% of the price level right now. And back then, you’re saying a lot of exciting stuff we’re gonna happen. So we bought on open market. It was a very decent investment for our funds. So thanks for that. And so for myself, like, are there anything interesting in the short term that we as listener should be looking forward to?
Zengyi Qin @Co-Founder of MyShell
Yeah, we’re building two foundational AI technologies. The first one is the text to app. So this AI model can create any app for you. So you just need to tell it in text what app you want. Then the model will build this app with the AI capabilities on my shelf and deliver the app to you. So that’s text to app. And whenever you have an idea, you can build app without do any coding. So that’s the foundational technology that is to come in the recent month, and we also have another AI foundational model that will come in the early April, which is our new large language model. And that language model has surpassed the META’s Lama two model and its inference speed is very fast and computational efficiency surpasses the lemon to also. So that is to very exciting foundational technology that Masha will release in the coming.
Chung @Investment Director of Arkstream Capital
So two models, one new product and other one is kind of an improved person. So two new stuff are happening. And just to follow on what is your plan for token launch? And before that, like, are there any opportunities business can kind of get by involved, involving it, maybe engaging in the models that you’re or products that you’re launching.
Zengyi Qin @Co-Founder of MyShell
Yeah, so now we already open the AI model and AI app staking mechanism. So if the web 3 users think that some apps on my shell is promising that you can stick in the app. Yeah, that is something that we already have. And for the future tokenomics and token launch, we’re actually internally designing it and working on it. So that’s not a very good time to release.
Chung @Investment Director of Arkstream Capital
Okay, absolutely. And so Johnny from bi seniors, like as a crypto media, I think you’re seeing lots of great stuff. Maybe some of the secret within the industry, maybe first hand. They also like, are there any alphas or any information, any advice that you will be giving to listeners in order to basically get some practical knowledge or practical reward in some sense in the AI and DP landscape.
Jonny @Chief Editor of BSCNews
So sorry, just to clarify, are you looking at like sort of educational resources or advice to projects building within the space?
Chung @Investment Director of Arkstream Capital
Practical stuff, of course, educational. So resources or some projects that we should be paying attention to those kind of stuff if you’re seeing anything interesting. But of course, if not, any general resources could also be quite worthwhile for us.
Jonny @Chief Editor of BSCNews
I think, you know, we do our absolute best to try and identify these exciting projects as they come through. I often say that we work fairly similarly to a VC, in the way making sort of bets on which projects are gonna be exciting and develop and gain traction over a sort of six to 12 month period. But instead of a VC where they’re playing with capital and other services, we’re playing with time resources and you know, who do we spend time researching and writing about.
So what I would say is, you know, to chill ourselves, so to speak, is to keep an eye on the kind of projects that we’re looking at funding rounds as well. They’re often a good sign of a project that’s very early on and has some fresh capital to build some exciting stuff. But once you know, you, you’ve, you’ve seen it, you’ve seen a project you think is interesting or has potential on BSCN, then go away and deep dive into the docs, right? The white paper, because there’s absolutely no substitute for getting in down and dirty and actually sort of picking projects apart. But my bit, if that makes sense.
