Latest / Crypto Coin Show / DePIN Infrastructure & AI taking over Blockchain? W/ Sami Kassab | Blockchain Interviews
Transcript
- 0:00I'm Ashin Addison from block West Capital for Investment Pitch Media, and today on
- 0:04the Cryptocorn Show we have Sammy Kasab, the research analyst from Missari. Sammy,
- 0:09welcome to the show and thank you for taking the time. Yeah,
- 0:12thank you for having me excited chat. Definitely, there's a lot to dive
- 0:16into so far this year in crypto that I think a lot of people need
- 0:22a research analyst to be able to follow along with all of the updates with
- 0:27AI merging into blockchain, Bitcoin ecosystem growing beyond just to transfer a value,
- 0:34and so many other small industry use cases which are getting closer to mainstream ad
- 0:38option and I hope with blockchain technology. I would love to start with just
- 0:43hearing a little bit on your background and what your focus has been with research
- 0:48and leading up to Massari, and then we'll sort of dive into current events.
- 0:52Yeah sure, so I'll keep my background pretty brief. Studied aerospace engineering
- 0:57in college. Worked as design engineer in the airspace industry for about four years
- 1:03when that consisted on working on the design development of different gas turban engines,
- 1:10propulsion engines, helicopter engines, and then I also did a year working on
- 1:15missile and defense systems, so pretty cool stuff, pretty cool hardware to work
- 1:19on. But around twenty is when COVID hit. Just like everybody else,
- 1:25I had a lot more free time on my hands. I was previously pretty
- 1:29interested in investing. I thought I knew how markets worked until twenty twenty,
- 1:34but that just completely blew me out of the water because you know, the
- 1:38economy shutting down. Meanwhile S and P five hundreds just going exponential, right,
- 1:44So that really led me down the rabbit hole trying to figure out,
- 1:47Okay, what is monetary policy, what is fiscal policy, what is the
- 1:49federals? What is their goal their objective? In the conclusion that our generation
- 1:56was doomed, but I found bitcoin. Bitcoin gave me, gave me hope,
- 2:01and then from there, you know, discovered the world of other crypto
- 2:07assets. And I got the opportunity to kind of work as like a contractor
- 2:13in Massari's community hub about two years ago, and so I was writing part
- 2:20time, and then I got the opportunity to go full time in the space
- 2:23and I haven't looked back since. Incredible, and I always appreciate when somebody
- 2:30learns more about traditional markets and then they see they understand bitcoin more. No,
- 2:36I think a lot of people that are outside of the industry they have
- 2:38a hard time understanding bitcoin because they don't understand what is the opposite, not
- 2:44the exact opposite, but what is the traditional markets, how does that work
- 2:47and how does bitcoin differ from that? And I feel like that's the entrance
- 2:51point which then opens a floodgate of all of the other use cases and sub
- 2:54industries and everything else. And so I'm curious you. You know, you
- 3:00had that that bitcoin event, you learned about it, and then you started
- 3:05getting into all of the deeper weeds of all the other parts of web three
- 3:08and AI. Where are your main focuses for your research and what you're most
- 3:14interested in? Yeah, first of all, I completely agree with you.
- 3:19I think everybody entering the crypto space should start with bitcoin. I feel like
- 3:23that's a foundation, like an understanding that everybody should have. But so,
- 3:30in addition to covering bitcoin at Massarry, I also cover a lot of infrastructure
- 3:36in general, So that includes like your storage networks, your compute networks.
- 3:42I largely am attracted to any type of crypto network that has a hardware aspect
- 3:47to it, and I think that just goes back to my background in hardware,
- 3:53and so you know, when I say hardware networks, I'm talking of
- 3:58networks similar to like helium, so like decentralized wireless networks, decentralized sensor networks
- 4:04such as like hive map or demode. We can kind of get into this
- 4:08later on, but Amazon, we've kind of coined this term into deepen,
- 4:13which is decentralized physical infrastructure networks. And so outside of that, in addition
- 4:18to some application infrastructure like you mentioned, I've also been recently diving into the
- 4:25intersection of crypto and aiy interesting, and yeah, I want to dive into
- 4:31each of those a little bit. I'd love to start with with the physical
- 4:38aspect of it, because you don't often see projects, you know, when
- 4:42people first hear about bitcoin, they're just like, oh, it's just on
- 4:45my phone on the internet. There's no physical hardware. You don't need to
- 4:48have anything physical to use it, which is amazing, but it can also
- 4:53get confusing because when people normally have you know, paper bill or their credit
- 4:58card in their hand, they're like, the money's on this physical thing.
- 5:00And then it expands beyond that to these whole infrastructure networks. Which I'm personally
- 5:06fascinated by it because I do love when there is a hardware aspect to it,
- 5:11and I think people have a better understanding of it when they can actually
- 5:15touch it and feel it. Maybe you can give an example of one of
- 5:17these use cases or one of these companies that you're following, and how the
- 5:23physical hardware works with blockchain technology. Yeah. Sure, So first thing I'll
- 5:29start by saying is that nearly every crypto network has a physical aspect. For
- 5:34example, you need to run the nodes, right, but there's a difference
- 5:38between needing the nodes for security and then needing physical nodes to provide some other
- 5:45sort of service. And so that's kind of where I draw the line of
- 5:49like what protocopy, what projects should be considered a deep in project, and
- 5:55what should it be. And so I think a perfect example to kind of
- 6:00administrate the capabilities or like the value of a hardware based cryptoproject is Helium.
- 6:06Helium's vary widely known, but just to break it down, they're a decentralized
- 6:12wireless protocol. There are two separate networks. They have their IoT network and
- 6:17they have the five G network that they're currently building, which is their cellular
- 6:21network. So essentially the protocol's objective is to incentivize individuals to deploy these hotspots
- 6:30and basically by you know, you're distributing the capex of the network to the
- 6:36supply side with the intention of basically creating this like distributed mesh network in a
- 6:43way. And so Helium has actually created the largest IoT network in the world
- 6:49just by using these crypto incentives. And so this network can be used to
- 6:54actually transfer data over. And so when you think about a future potentially where
- 7:00IoT is actually a thing and you have these smart cities that all have these
- 7:04sensors embedded in them, they need this cheap, cost effective IoT network to
- 7:12transfer data over. And that's kind of the vision that Helium is going after.
- 7:16But really, how blockchain technology plays into this is like blockchain is really
- 7:21just the like coordination and incentivization layer, and that's key, and that's really
- 7:30what needs to be understood about these these networks and that's really where they shine
- 7:33rights incentivizing people to complete specific acts. Yeah, I know, I think
- 7:39the incentivization is key, especially because there's such hard challenges to overcome logistically to
- 7:46build this up to something so big In a decentralized way. I've been hearing
- 7:53about mesh networks for for years now, and on paper it sounds amazing.
- 7:57You know, there's there's a lot of issues with traditional uh internet or cell
- 8:01phone providers where you know, centralization or the costs and monopolies, oligarchies.
- 8:07How can we build a network for the people? Right? But I feel
- 8:11like there's a lot of hurdles to overcome, and uh, I think starlink
- 8:16might be like a hybrid where it's like, oh, you can sort of
- 8:20it's cross border, but it's still not you know, decentralized or owned by
- 8:24the people. Where do you actually see the reality of this setting in for
- 8:28people that you know are sort of tech savvy, they don't have to be
- 8:33Web three geniuses, but like they could use potentially an alternative to uh,
- 8:39these centralized internet providers. Is that actually going to become a reality in you
- 8:43know, the near future. Yeah. I think we're We're we're not there
- 8:50yet, but we're getting close. So for example, Helium, the court
- 8:58team which is called Noble Labs, they are the ones who are actually developing
- 9:01the network. They've actually created a mobile carrier, so they've partnered with T
- 9:07Mobile to provide a service to individuals where so you can like go on your
- 9:15phone and you can actually see like you can be connected to the Helium network.
- 9:20But essentially how this works is that when you're not in range of like
- 9:24a Helium node, you actually will then leverage the T mobile network. And
- 9:30so that's really how I see this progressing. It's going to be like a
- 9:37combination of leveraging like traditional infrastructure with decentralized infrastructure infrastructure until the decentralized infrastructure actually
- 9:46is grows large up or stable enough to actually be able to accommodate a lot
- 9:52of these users. But really, like the key reason I see why people
- 9:58would use these decentralized infrast structure networks just comes down to cost. Right,
- 10:01most people don't really care about decentralization at the end of the day. But
- 10:05what's great about these decentralized networks is they actually have like superior unit economics,
- 10:13so they reduce the capital expenditure, they reduce the operational expenditure, and so
- 10:16through that they're able to deliver lower costs or more cost savings to the end
- 10:22user. And so that's kind of what I see driving adoption at the end
- 10:26of the day. Yeah, I think at the end of the day,
- 10:31for regular people, they just look at, Okay, what's what's a better
- 10:35deal for me, And as long as it's not too much more complex than
- 10:39I'm in for the cheaper option. And from what I understand with the Helium
- 10:45and some of these other decentralized networks is in a way you can contribute,
- 10:48whether it's your bandwidth or some kind of time effort that you're actually earning,
- 10:54So that's reducing your costs. You can actually earn tokens or something else.
- 11:00Yeah, exactly. A large portion of these decentralized networks when you think about
- 11:03like decentralized storage computer bandwidth, it's leveraging idle resources or resources that you'll already
- 11:09have that you're looking to monetize. Right, So, first of will he
- 11:15up a gaming computer with a couple of high end GPUs, let's say like
- 11:20RTX thirty ninety. If you're at work and those GPUs are just sitting idly
- 11:26not being put to use, then you're you're kind of missing out on an
- 11:30opportunity where there are some decentralized compute network are emerging to where you can contribute
- 11:35those GPUs and it's a very passive activity and you're earning without really putting in
- 11:46any additional effort. Right, So that is like a key aspect of a
- 11:48lot of these infrastructure networks, and that is part of the reason why they're
- 11:54able to achieve like lower costs than centralized Yeah, great point, Sammy,
- 12:01And I feel like there's huge inefficiencies across most industries with technology, you know,
- 12:07pivoting slightly thinking about automotive industry and cars and how cars are idle for
- 12:16you know, eighty to ninety percent of their life just sitting in your garage,
- 12:18you're sitting in the parking lot. I feel like tapping into a web
- 12:20three network or somewhere where they can monetize that would create a huge more much
- 12:28more efficient for many IoT devices. And I feel like over the next coming
- 12:33years that's going to become a reality. Yeah. I mean, something that
- 12:37Metallic said a long time ago was that if you remove these centralized intermediar areas,
- 12:45you just drastically increase efficiency of a lot of these networks. And so
- 12:48if you think about something like Uber, if you remove to them, you're
- 12:52removing this intermediary that's just taking a cut, right, So if you placed
- 12:56Uber with like a dese centralized protocol and you allow these like decentralized front ends
- 13:03to emerge on top, similar to what we see and like decentralized social where
- 13:07you have like the decentralized social graph layer and then a bunch of like decentralized
- 13:11front ends emerge. That's kind of how we see it playing out for like
- 13:16you know, real world networks, and these are actually being built. So
- 13:20there are decentralized rideshare protocols, just to name a few. There's one called
- 13:24Teleport, there's one I believe it's called Drive, and then there's also decentralized
- 13:30food delivery protocols being developed as well, and I think there I think that's
- 13:35called there's one called Nauch, and so that is really the vision at one
- 13:37point is to turn these centralized web two apps into protocols to increase efficiency,
- 13:46lower costs on the end user, and really just increase fairness overall. Great
- 13:52points, Sammy, and I'm going to look into those companies as well.
- 13:56And there's one that I've been following I've had on the show a few times
- 13:58called Network as well. From what I understand, they're building AI IoT devices
- 14:05and you know, the the end vision is to have cars that are automated,
- 14:11can drive themselves, have their own bank account, can transfer to other
- 14:13cars, the gas stations, everything sort of communicates through their IoT devices.
- 14:20And I feel like we're getting there. But one of the main hurdles,
- 14:24which I think made huge achievements this year towards that direction, is AI technology.
- 14:31And been working on it for many years, but I think reality set
- 14:35in for people with with chat ept uh this at the beginning of this year,
- 14:39and I know they've also been researching AI and it looks like, you
- 14:43know, it's on an exponential curve to just keep growing faster and faster.
- 14:48Maybe you can talk about your research in AI, where you know the achievements
- 14:52that we've seen this year and the projections that you see moving forward. Yeah,
- 14:58like you said, chat GBT was really that wake up moment for a
- 15:01lot of people about how disruptive and how powerful AI really is going to be.
- 15:07And I think the reason for that is is because chat GBT was the
- 15:11first consumer facing AI product, right, And so I'm kind of waiting for
- 15:16Crypto to have that moment as well, where we have this this a real
- 15:20consumer facing product where people can kind of realize, Okay, I see what
- 15:26I see what's being built in this industry. That's real value being created here.
- 15:31So that's just kind of a side note. But yeah, so ever
- 15:35since chat ebt really took off, there's just been this realization among pretty much
- 15:41every company about the need to adopt in AI strategy. And what that's caused
- 15:48is just this this increased competition for the resources that are underpinning AI. So,
- 15:58actually, before I dive in to that, because that's one of the
- 16:00use cases that I see emerging between the intersection of crypto and AI. There
- 16:07are you know, a number of liquid AI and crypto assets right now.
- 16:15The majority of them I see as vaporware. I don't think they actually provide
- 16:21real value in a way, I kind of see them as being like a
- 16:22fun toy. While there are some projects who are building like necessary infrastructures,
- 16:30such as like decentralized marketplaces for AI agents or for machine learning models. I
- 16:36think the real use cases are still being developed, and that's what I was
- 16:41researching. I can basically group the use cases into two broad categories. You
- 16:48have one category where crypto can be used to address existing challenges in the AI
- 16:55sector, and so that can include alleviating the GPU shortage which I was touching
- 17:00on earlier. With this increased demand for everybody to train this mL model,
- 17:04we have this major GPU shortage, and so crypto can actually be used here.
- 17:11If you leverage these decentralized compute networks that tap into idle compute resources that
- 17:18are distributed all over the world, you can help mitigate that supply and demand
- 17:22mismatch by bringing a lot of GPUs online for others to leverage. So,
- 17:27like that's one use case there. Crypto can also be used to help combat
- 17:32against deep fakes. You can use it to build these networks that can incentivize
- 17:37collaboration, you can incentivize data collection. And then the second category is really
- 17:44how AI can help crypto, And in a way, I see AI as
- 17:49accelerating the adoption of crypto infrastructure and then also unlocking new possibilities with what you
- 17:56can do on chain. So I'm happy to die into wherever you think is
- 18:00interesting. Yeah, I know, we we touched on the first half there
- 18:04a little bit, and it sort of made me think about, you know,
- 18:07all these people that have these graphics cards, especially ones that they were
- 18:11using to mine bitcoin. You know, some trying to find some lucrative way
- 18:15to profit off of their graphics cards, and the competition, excuse me,
- 18:21the competition gets stronger the graphics cards, it makes it harder to actually mine.
- 18:26But hey, there's all of these other use cases where you know,
- 18:29this graphics card could be put towards helping protect this infrastructure network or this AI
- 18:36or other ways to monetize. And I feel like what you're talking about with
- 18:38AI is going to open up, you know, ten other doors for idle
- 18:44technology that maybe once was used to mind Bitcoin be able to monetize from other
- 18:48aspects. Yeah, exactly. And we actually see this happening when you look
- 18:53at large publicly listed crypto miners, and so a lot of them had high
- 19:00end GPUs that they were using to mine ethereum, and what they actually did
- 19:03was they pivoted towards creating this high performance computing side of the business. So,
- 19:11you know a lot of them are still aligning bitcoin, but now they're
- 19:15also these high performance compute providers, and a lot of them are looking into
- 19:19supplying the GPUs that they have to do some of these decentralized infrastructure networks or
- 19:26decentralized compute networks. And so, for example, a COSH is a decentralized
- 19:30cloud, and there are some miners who have actually come out and say,
- 19:36yeah, we are going to be providing some of our resources to this network.
- 19:40And so it is a viable use case, whether it's for decentralized rendering,
- 19:45decentralized machine learning, or you know, any other compute workload that emerges
- 19:51in the near future. I'm glad that you mentioned a COSH. They're sounding
- 19:56because I have spoken with them before. And I think infrastructure that's holding Web
- 20:02two together right now and including most of the crypto projects adapts, even nodes
- 20:08for ethereum and such, is hosted most of it on Amazon or some other
- 20:14centralized provider that you know, if anything happened or if they wanted, they
- 20:18could just flip the switch and most of the servers of the Internet would be
- 20:22shut off. And I feel like these decentralized networks that are able to store
- 20:27servers, applications in a decentralized way is the future. But I think what
- 20:33we went back to at the beginning about you know, at the end of
- 20:36the day, it's sort of about a cost benefit, you know, is
- 20:38it similarly difficult, but it costs a lot more. I'm not going to
- 20:41use it. I'm going to use I'm only going to use the decentralized one
- 20:45if it's cheaper or there's incentives over using Amazon. I'm curious on your research
- 20:51on that and where you actually see that switch to a decentralized infrastructure actually hosting
- 20:57not just Web three apps, but like the Internet. I feel like we're
- 21:00far away from that stone. Yeah, I'm gonna be honest, I think
- 21:04we're really far away from that because convincing a business or an organization to move
- 21:11their infrastructure off of this reliable Web two provider to this Web three provider is
- 21:19it's going to be a difficult, a difficult thing to convince them. For
- 21:26one. So one thing that you know these decentralized clouds struggle with right now
- 21:30is really reliability. So if you look at the existing providers they have or
- 21:37they can actually guarantee like ninety nine point nine nine percent up time, that's
- 21:41not really the case with these centralized providers because you are leveraging these these individuals
- 21:48who don't have these SLA agreements, right, so it it's not as reliable
- 21:53for running these critical workloads. But that's not to say that there can be,
- 21:57you know, methods in the future that arise that basically make these networks
- 22:04a lot more reliable. One thing I'd just like to mention here is that
- 22:10previously a lot of these distributed compute networks were only providing CPU resources, and
- 22:17this was a resource that was largely abundant in the existing market. Right.
- 22:21GPUs are now the like in demand resource, the resource that everybody needs,
- 22:26right, And I think these decentralized compute networks have really been handed like an
- 22:30opportunity on a golden platter to where if they can if they can basically source
- 22:37these GPUs offer up at a reasonable cost, then you could actually potentially see
- 22:44these web two machine learning researchers or small to mid sized companies actually jump through
- 22:51all the hurdles of crypto, go through all the onboarding process to leverage a
- 22:56service or a use case that they actually need. That's one critique that I
- 23:00can have I can make about crypto is that there really is in the use
- 23:03case that has spread wide to like general society. So I think decentralized PEW
- 23:11has the opportunity to actually attract a lot of individuals just by offering a service
- 23:17that the existing market doesn't have, and to me, that's going to be
- 23:21scarce GPUs at a reasonable cost, great insight, Sammy, and we've talked
- 23:26about a lot of different industries here that they have a great foundation built and
- 23:33nothing is you know, Rome wasn't built overnight. You know, all of
- 23:36this stuff is going to come to fruition, but it might take a while.
- 23:38I'm curious on your thoughts on a more short term basis, say it
- 23:44through the next six to ten months, moving into twenty twenty four. Is
- 23:48there any thing in your research showing, you know, some breakthroughs in web
- 23:53three or AI and what we might see come out next that's usable. Yeah,
- 24:02that's that's a good question. So within AI, I see a few
- 24:07use cases that we could actually see in the near future, within around six
- 24:11to twelve months. The first one is, I'm not sure if you're familiar
- 24:15with AI agents. They're essentially these autonomous entities that which are basically like an
- 24:21LM, like a large language model communicating with a large language model. If
- 24:25there are these entities that can go out and complete these automated tasks for you,
- 24:29there's this narrative that these entities will need to use these decentralized payment and
- 24:36infrastructure rails just because you know, they're not humans. They can't go out
- 24:40and KYC and get a bait account to transact with one another. And so
- 24:45within the sex six to twelve months, I would actually predict that we see
- 24:49a lot of these AI agents actually start to move on chain to leverage some
- 24:53of these crypto infrastructure rails. I think what we'll also see is on chain,
- 25:00you'll start to see smart contracts gain the ability to query or inference AIA
- 25:06models, and that's going to just broadly expand the design scope, right because
- 25:11now you allow smart contracts the ability to make predictions and decisions based on real
- 25:17time on chain data rather than like a fixed set of rules. Right.
- 25:21So now these these these smart contracts actually become a lot more intelligent and they
- 25:27they can more compete with you know, Web two applications. And so that's
- 25:33two entities, or those are two verticals within AI that I kind of see
- 25:36taking off. And then you know, decentralized storage is another sector of them
- 25:41keeping a close eye on. I think decentralized storage is the closest Web three
- 25:45infrastructure vertical to being commercialized. I've spoken with a lot of storage providers,
- 25:51specifically in the filecoin ecosystem that are storing data for research institutions and universities,
- 25:57and so I think the trajectory there is looking optimistic, and you know,
- 26:03I think more small to medium sized businesses potentially will look into leveraging these d
- 26:08essentialized storage networks, largely due to you know, the increased efficiency, but
- 26:14also the decreased costs that they offer. Wow. Incredible, and I'm definitely
- 26:19looking forward to the day when smart contracts become even smarter. I know that
- 26:23there are a lot of limitations to them right now, but they are a
- 26:27beautiful thing, and incorporating AI and AI agents I feel like would be a
- 26:32huge leap forward. So I won't hold you to a date a date on
- 26:37the prediction, but I'm looking forward to when and if that happens. So
- 26:41thank you for the insights, Sammy. What is the best way? You
- 26:42know, we don't have a lot of time left, but is there a
- 26:45place on missary or personally that we can follow you and look for these updates
- 26:52and follow along with the latest trends in web three? Yeah, I would
- 26:56say that you can follow me on Twitter. I'm at Old Underscore or Samster
- 27:02or just type my name Sammy can Sam on Twitter and that should pop up.
- 27:04And I'll also say that to follow Massari's Twitter as well to keep up
- 27:08with some of the updates that will be releasing the summer, and that's at
- 27:12massari Crypto as well. Amazing, Thank you so much, Sammy. I
- 27:18really do appreciate the time, the insights into all of these innovative technologies.
- 27:23It's it's it's amazing, what's what's happening in our world right now? And
- 27:27I love that you're at the forefront of it and you just have so much
- 27:30knowledge on it. So would love to have you back on the show,
- 27:33Uh, sometime when there's more updates and there's more to speak about, So
- 27:37let's definitely follow up in the near future. Yeah, I would love to
- 27:41be back on and really enjoy chatting with you in general. Man,