DePIN Economics: Token Incentives and Sustainable Business Models

[00:00:00]hey everyone if you have been listening to Empire you know that Santi and I are fed up with unaffordable fees and frustrating transaction speeds that make the onchain experience basically unusable so the arbitrum team reached out and they showed us the platform they showed us what you can do on arbitrum whatever you're doing you can experience frictionless transactions at lightning speed on arbitrum so head over to portal. arbitrum doio and check it out what's up everyone before we jump into today's episode I'm excited to share Empire's first first ever security partner Harpy is the best tool to prevent your wallet from theft in real time Harpy is not just a security solution they are a peace of mind solution but don't just take our word for it Harpy is the only wallet security solution that protected 100% of its users from attacks like The Ledger one and Q4 which was an offchain signature attack to learn more about Harpy click the link in the show notes or visit Harpy doio Empire what's up everyone before we jump into the episode epis little plug for digital assets Summit coming up in London March 18th to 20th tickets are pacing so far ahead of schedule that we had to decrease the discount codes so instead of Empire 20 it is now Empire 10 head over to the website digital asset Summit Das London March 18th to 20th use code Empire 10 and get 10% off your ticket see you in London all right everyone welcome back to Empire we have Kyle Sani back on the show we have uh Demitri uh previously at 1K and coin fund now over at archetype dimetri Kyle welcome to the show guys thanks for having us so the basis of this conversation is about deepin uh this is our first kind of full episode on deepin that we've done I have a feeling it's going to be the first of many the backstory of this episode came came about when Demitri tweeted out um I'm extremely bearish on deepin and happy to debate anyone about it on a podcast I figured I had to hit up Kyle to be on the other side of this debate um as they've been kind of leading the charge I think Kyle you guys were the First Investors to really go big on deep pin with I think it was helium then HIV mapper I think you guys have done render I'm not sure if you've done any of the file storage stuff I I think R weave maybe as well some of that kind of stuff but um Kyle maybe you can t us up I've heard uh deepin described like the why behind dpin as you can build infrastructure you know 10 to 100 times faster it can be more cost effective um you can be more attuned to like hyper local market needs you can take these Idol resources and deploy them there's a bunch of different whys what is the like why of deepin in your mind in terms of why it's so exciting I mean to me what's exciting about deepin is you can give people Equity ownership or Equity like ownership in building out something Collective uh that is bigger than any one person um all of the examples you listed and and as well as many others a lot of these things don't work at small scale if there's one person driving cars around then like uber doesn't work right you you need thousands or millions of people or whatever if there's one Telecom Tower then I got that's not a network that right that that's that's not useful um and so it turns out there's there's large there's a lot of different kinds of problems that um can be solved if you can incentivize people to work together um who those people don't know each other but they need to collectively produce some sort of um economically productive asset and you can not only pay those people on an ongoing basis for the Opex of using the service but you can also actually incentivize those people uh with some sort of equity like ownership for being involved early if you are the first person to put up a helium Tower you are taking more risk than the person than the 100,000th person who puts up a helium Tower um and so you should be compensated for taking on that risk right and and dein actually provides a mechanism to economically incentivize people for taking that risk before we go too deep down the rabbit hole maybe you could tell us just about dpin as a whole in the context of using tokens as a coordination mechanism and how what deepin is in terms of like using tokens to incentivize kind of the crowd sourcing and building of real world physical infrastructure maybe we could expand on that sure so so the basic idea in deepin is um you know you need to build out the supply side of a network we we'll use helium here since this kind of the the first of it of its kind um so if you're going to have a telecom Network you need to have Towers hotpots everywhere um and uh you know you could go online and you could create hotpots and sell in people and say hey plug it in and whenever the demand starts coming by uh you'll start getting paid you know per bite of data um the the problem with doing that is you know if you're the first guy to buy a hotspot well like that no no customer of that network is going to be interested um because there's not enough hotspots no one knows what the exact threshold is at which the number of hotspots is sufficient maybe it's 100,000 maybe it's 400,000 hotspots whatever some large number of hotspots with some degree of geographic distribution is required before you can actually call it a functionally useful Telecom network of any form um and so the idea with with dpin is you can go to those people and you can say look we're going to give you uh tokens in in the case of heing was h&t tokens for putting up hotspots um and you can Define the formulas that that specify how those tokens are handed out uh in the case of helium back in 2019 when the network first launched it was a very very simple model one that in hindsight was not optimal um but the model was just it was just X number of tokens are minted per day and based on however many hotspots are out there you get you know x divided by however many hotspots are present um that was a very simple model but what that incentivized was people who believed in the idea early to go buy hotspots some people bought 10 20 30 40 50 hotspots and they put them up you know in different places um the beauty of this model is that it allows you to build up the supply side of the network to get to that minimum threshold scale um and and if you end up getting to that threshold scale then you uh probably have a mechanism for uh giving more reward to the people who believed earlier again guy number four took on more risk than guy number 5,040 and that guy took on more risk than guy number 84,000 um and so you know it's a very elegant way to uh distribute risk and reward to the people who actually do the work of building out the network got it the idea of helium because you've touched on twice now is there are these big carriers the AT&T's and Verizon of the world they basically have nobody's able to compete with them because it's too expensive um too cost prohibitive and the scale that they've reached is uh basically nobody can break through and the only way to do this is by a very Bottoms Up go to market strategy where in instead of you know another company coming to Market raising $500 million building these hot you know these towers all around the world instead their thesis is look nobody can can compete on that model instead what if hundreds of thousands and inevitably millions of people are actually creating these micro cell towers in their home essentially is that the helium thesis that that's correct yes okay so expand out into so that's the uh Wireless basically I'm what I'm trying to get out with the this is what is the market today so that's the wireless version of dpin maybe you could tell us about the whole market today whether it's compute Wireless how do you break apart this uh the dpin industry uh sure so at the highest level there's probably two major categories dvin or virtual infrastructure networks and dpan or physical infrastructure infrastructure networks so virtual infrastructure would primarily be compute resources so things that are in a computer so storage bandwidth and compute being kind of the three you know physical Primitives there um and then the the major networks here filecoin live Pier render R weave I there there's a bunch aosh there's a bunch of them uh focusing on different kind of sub you know uh slices of that and then there's dpin which is physical infrastructure um which is really pioneered by by helium and then more recently you've got teams like Hive mapper you've got teams like demo you've got teams like weather XM uh there's probably couple few dozen more that I'm not thinking of off the top of my head um but these are and obviously with the major difference with those is the D pins versus the D vins D pins you you're putting Hardware somewhere in the world that that that Hardware is intrinsically very GPS sensitive um or or G right so like you don't just need a thousand helium hotspots in Austin you want 100,000 all over the United States with with some distribution with within that if you're with doing Hive map or you're mapping roads obviously you need people driving everywhere um with d vins generally speaking G GPS sensitivity is much lower um than with d pins um there is some degree of GPS sensitivity that that is a real thing um if you're building a CDN for example like latency is a thing you don't want to be going from Virginia to Texas you want to be going from Dallas to Austin not from Virginia to Austin um so you have some degree of of latency sensitivity but generally speaking is is substantially less so um that that's like the broadest um way to segment the market and then you can kind of slice and dice from there but that that's probably the most important segmentation all right dri take the other side of this I know that you've since that tweet you've gotten a little more excited about it um you've talked to a lot of good Founders and stuff like that but take the other side of this why were you so bearish on on dpin back in November yeah um happy to give some concerns I do want to preface by saying archetype at archetype we are exploring the space we are talking to Founders ideally what I would want you know at the end of this podcast is for a Founder to reach out and tell me you know here's why you're wrong here's why we think the timing is now um you know strong opinions weekly held um and I think the last you know month or two has been quite interesting because I have been talking with a lot of mission driven Founders it's been refreshing to actually talk with these folks in in the deep end space an interesting through line that I see when you actually compare crypto to the internet is that the narratives often reflect the cohort of builders that actually see an opportunity and you could look at this with D5 for example saying you know that was the financial engineers and you know nfts or the artists and musicians and dii I think are the biology phds and you know I think dpin are like the electrical engineering phds so like the internet I think there is everything for everyone in crypto and I think there has been more of a Critic IAL mass of Builders actually coming in today than there were you know a few years ago when uh when this started um I think the biggest concern to address and also just to not you know at the I I I would hope that a benefit here is also you know for Founders to uh to think about you know like like what are the concerns that they should be thinking about when you know when building these things out um and you know for for users for investors to really know which you know what are the right questions to be asking you know because I think narratives are healthy to an extent um but I think we should be looking at both sides to to uh to make sure we're all going in the right direction um that said I think one concern that I've had is more on the supply side where you know Kyle talks about you're able to use tokens to bootstrap the supply side I think that's true uh I also think that the key word is bootstrap and I think a big issue for me where we are in this cycle is to what extent can you penetrate the personas necessary to have a supply side be valuable at scale whether that is some data that you're provisioning or a product platform service a lot of the deepend projects today often have a pattern by which they use tokens to boot the supply side and and then they either get some valuable data or or or some platform service that they then turn around and they sell that to more of a B2B demand side and I think when you look at a lot of these networks when you start to think about the sheer amount of scale you know dollars that need to be committed I worry that we're really not there yet in the adoption curve of crypto in general because I think we're we're we're at at the stage roughly where it's the innovators maybe it's a portion of the early adopters maybe early majority but I think we're we're actually still quite early and I don't think you know the the the late majority they know or care about tokens as much as you know everyone on this call and probably the listeners do so I think it's largely a timing issue you know we've SE I I ordered Amazon Fresh last week and and and you know I thought about web band and and you know like the story of that company raising you know an asinine amount of money just at the the dawn of the internet and and that failing spectacularly and and they they weren't wrong you know I think a big issue there was internet penetration uh and and and I think there there can be some analogies today where I think the the awareness with tokens the on andof frames are are not quite there yet to really have a lot of these net networks get to the scale where they have Network effects and they're able to offer something actually valuable to a demand site your main concern was around the demand side adoption the supply side it's two things yeah I mean we uh we could talk about the demand ey later um as well I think that's also a big concern the argument is on the supply side the ability to expand beyond the early innovators and early adopters is going to be very difficult because I still think we're at the stage of the market where very few people in the world uh find tokens valuable as a mechanism for CTO Equity Kyle thoughts um yeah so the demand side you at least on the networks I think about um demand side they don't have to know what tokens are or interface with them at all even if the protocols require that in in protocol um for example take um Helia mobile uh consumers who are signing up for Helia mobile you go put in your credit card now and you know you switch your Verizon phone number over um Nova Labs which is the The Entity that runs hel mobile U they actually then take your $20 a month they turn around they buy h&t data credits on the blockchain and they they do all that stuff um you could say that centralized but like the whole point is there's a single carrier the consumer has a centralized you know entity that they face so I think that that's like a a perfectly reasonable abstraction on the demand side um I'll take another example which is Hive mapper um Hive mapper whereas it you know it's good to contrast helium The Hive mapper helium is a direct consumer product um in the case of Hive mapper no one's buying data from from The Hive mapper Dow um Hive mapper Inc has uh signed I believe three four five different commercial contracts they have not disclosed to any of those customers are um but they have them in actually they're uh what they're doing now is those customers are paying Hive mapper Inc um Hive mapper Inc is then turning around and uh uh interfacing with the hive mapper Dow and Licensing the data from The Hive mapper Dow which is the The Entity that actually owns all of the intellectual property of those images hi mapper Inc then on top of that data actually they're doing a bunch of AI processing stuff so stuff like which way is the stop sign facing is it 35 miles an hour or 30 all kinds of stuff like that so they're actually at they're they're acting as like a value added reseller to The Hive mapper right they're doing all of this this additional compute and work um and in that case they're facing whatever these big companies are that are buying buying this data um those companies don't care at all about tokens those companies are never interfacing with the tokens at all um so in the case of uh so so the demand side I'm not worried about I think de I think Demetri your concern is more on the supply side of saying all right there's this initial group of folks who cares about tokens and they'll you know they'll drive their car to get get these tokens but eventually you know maybe there's only 50,000 people in the world who care about these tokens right yeah yeah I do agree on the demand side you for sure need to abstract that but yeah I mean here's another thought experiment uh you can maybe say you know if you're doing whe uh if you're collecting weather data you can say you can ask the question you know how many weather stations are around the world do we need if you're mapping roads you could say how many unique kilometers do we need if you're building out you know a Wi-Fi hotspot Network you can ask how many um pieces of Hardware do we need and even better you can denominate it in cost in dollars of infrastructure and again you know like this is an example which has separate issues but in an example where the supply side is actually spending there's some Capital outlay to actually buy a resource rather than plugging latent Supply into into a network but you can say in some example you know say you need $10 million $100 million to to to bootstrap that supply side and you could say then okay it costs $500 per piece of Hardware to buy then you get a number and then you say oh we need this many people to buy these pieces of Hardware or uh around the world and performance are in action and we believe that we can incentivize them with a token which you can then say you know some portion of the supply um uh we can give out that equates to that aggregate dollar spend um the worry is that the the the number of people there is going to be quite high and more than the amount of people who actually care about receiving these token setups today um yeah I mean I mean it's it's hard to generalize um let's take again I just use hel Hive mappers since we have the most data on them um Hive MPP is particularly interesting because there's a large class of people who drive around all day um the UPS drivers your Amazon Whole Foods delivery people obviously Uber and lift drivers uh I don't know how many of those people there are but there's obviously a lot um is clearly the correct answer um and those people drive you know 10x or 50x more than you I do per day um and so for them the math on hey should I buy a dash cam um is very different than for you or I to buy dash cam um uh so I think about you know that that that's a very interesting group to go after uh moreover those are particularly interesting groups because they have their own communities there's group you know online forums for lift drivers and Uber drivers and stuff and they can share this and they can say look I bought this thing I made x amount of Honey tokens right and you can see this thing going viral very quickly in those communities I don't think you need that many Uber and lift drivers to cover the entire United States um with with pretty good pretty good coverage um that's just like one example in the case of helium uh it's hard there's there's nothing there's nothing obvious like uber and lift drivers to go after um but we anecdotal we have some data there were about a million hotspots iot hotpots deployed um the guess is like that represents 3 to 400,000 unique people um and then on mobile now you know I'm guessing there's probably 20 to 50,000 um you know in unique people who have done mobile hotspots so we have some data on that today um it's unclear is that enough or how how do you get more um however one thing we do know is that um token prices uh well in the case of helium we know that they have this consumer helium mobile thing and a lot of people want to sign up for the $20 a month plan and a lot of those people trying try and understand what is this thing why is the price so low how does it work and a lot of those people get excited to go buy buy a hotspot um so you have kind of a natural marketing like weirdly in the case of helium the demand side marketing for you know helium mobile also is the marketing for building out the supply side um which is very unique that you have that kind of natural uh Synergy um I don't want to trying to extrapolate these specific patterns to other dpin networks because I think that the nature of supply and demand for these networks will be will be pretty different but you know I think we have pretty good indications so far that um you can get to at least enough scale to get the flywheel going um uh we've seen that with theum and I think now with hi mapper and fingers crossed you know demo and weather XM and a bunch of these other guys will hopefully figured out as well Eyes Wide Open though that adding custom Hardware that's not latent Hardware it definitely just makes your it's it's much harder to do um both because of the uh capex requirements Hardware manufacturing is obviously difficult and all and all that stuff so definitely you're uh a lot less room for error in execution Kyle You' mentioned this flywheel a couple times what is the deepin flywheel yeah I mean the de-pin flywheel would be you get let's let's say the threshold scale for a wireless network is whatever 100,000 you know hotspots uh like if if you can get to that scale to 100,000 now presumably okay customers are excited they're onboarding they're they're paying for for this data that now is public information because the transactions are going to go on chain you're going to be able to see the revenue in real time um that's going to cause potential you know buyer number 100 101,000 to say oh I no longer have to speculate uh on like is this service useful is anyone willing to pay for it right you don't have definitive proof of that and you can see the revenue in real time so that that makes it easier to to underwrite the investment as being person 100,000 one um obviously as as that person then uh you know adds their supply to the network presumably that's making the network a little bit better presumably that will that incentivize the next customer to say oh great the network quality is now at a new higher bar that I I like it even more and I'm more willing to pay for it and so you get this natural effect where more demand leads to more Supply which creates more demand yeah it feels like there's kind of two flywheels going on here a little bit you've got um higher infrastructure utilization leads to lower unit costs which leads to higher unit which leads to more Network effects which leads to higher infrastructure utilization and then on the other side it's like you you actually do have a token flywheel which is like more token liquidity higher token value bigger Network effects and that all kind of ties together into like demand to supply to coverage to ux um I think uh I forget the name of I'm blanking on the name of this fir I think you guys might have backed this firm e EV I'm forgetting the name of what they stand for but uh EV EV3 EV3 yeah they had a great report on on deepin they laid out this fly pretty well um what what what are your thoughts on actually both of you guys because Demetri I know you've been doing research the last uh several months on this and and Kyle you guys have been investing heavily the last couple of years on B Toc versus B2B for dpin so it sounds like Hive mappers business model they map all the roads then you've got that data and right now I think it's Google is probably the world leader in this um but it's very hard to scale that so what Hive mapper can probably do is they can sell all of this for maybe let's call it 10 times cheaper to to companies who need the the road data um but healing them on the other side is so that's a B2B model helium on the other side is more B Toc right they've got there's now consumers signing up for their uh you know helium cell phone plan what is your takeaway Kyle in terms of B to C versus B2B in terms of the best model for dpin I I don't think there is a a best it's going to be it's going to be dependent on the nature of of the demand side there is no consumer demand side for Hive mapper that that that won't ever be a thing um they only sell Enterprises helium is the exact opposite they're Focus like the reason reason to build out a global 5G Telecom Network is for Consumer use um there's no other demand uh Market that's big enough to to Warrant the investment um to build out a you know Nationwide scale would would helium be better off though not being the one to actually sell the be the front end for the consumer like in in in theory helium could basically just provide the network and other people could build on top of helium right uh yes uh I think so there's been this idea in uh if you're live in Telecom land for the last 10 years it's been out there called this idea of kind what's called carrier offload and the idea is that there are specific places where carriers either don't have coverage or they have poor coverage like let's say a football stadium is kind of like the quintal example uh and there can be agreements negotiated between let's say Verizon and whatever the guy who owns the football stadium to have some sort of carrier offload function there um so this idea has been floating around for a while um it really hasn't taken off in any meaningful way um my expectation is that uh the helium network will probably become the first Network where you know real Nation carriers both in the United States as well as in other countries I think will sign carrier offload deals um with with helium uh because it will be in their interest to do so um so yeah that that can happen uh it hasn't happened yet definitely part of part of the game plan the further you dig into uh Telecom nerd land the more you you kind of learn about carrier offload carrier offload is when you go to like a you know the giant Stadium or something and it just says Verizon Wi-Fi access right it's a way to basically deploy to divert reduce congestion basically of the network yeah it's not in that specific instance it's not clear to me if Verizon actually put up some infrastructure in the stadium or if the stadium did it themselves there's going to be different configurations of that you can also Imagine train stations some subway stations I mean even Parks like for music festivals and stuff like there's a lot of these places where you get heightened congestion and where carriers are are generally you know not providing great service Demetri are you uh optimistic or pessimistic on uh I guess both both of those B2B and and B Toc dpin I look at it from the perspective you know if you're building a business even outside of dein there's pros and cons to each if you're doing B2B the pro is that it's quite easy to find pmf because there's a relatively small number of customers that you need to go to and it's pretty straightforward you you go to a business you ask them what they want they tell you and then you turn around and and and you build that the difficult it's that easy yeah um but to generalize Beyond One customer is hard because there's often quite unique demand need for for for one say like Enterprise um on the other hand with consumer it's very difficult to understand what the market need is because you need to do a lot of quite in-depth market research and and often people tell you what they think they want and it's not actually the thing that they want um but then if you manage to hit that then it's very easy to scale and generalize uh it feels like that the order of operations for for um for deepin uh networks that have been moderately successful on on the demand side and and and I define success more as some tangible willingness to pay more so than the number of like the amount of Supply you have on um they have tended to start out with B2B first and I think that's actually um uh quite nice because then you know if you have some finite uh token incentives that you can use to actually bootstrap a supply site it would help if you exactly what you're trying to to to achieve uh that's kind of my my sense about it um there there's obviously uh concerns with like General like Market size I think for um for a lot of these uh uh projects if you're just targeting uh uh uh B2B um but I think you cannot run a sustainable token economy if there is no Fiat inflow and you kind of gravitate to where there's willingness to pay and it seems like when you look at a lot of these projects it tends to be more on the BW side today Kyle in um I was re rereading your guys's post proof of physical work that I think tar published in let's call it Q2 2022 and there were five main benefits of actually of the crypto element of this because someone listening maybe who's not in cryptand might say well yeah there's a lot of businesses like this right um you know more servers equals lower latency equals more servers lower latency demand drives Revenue drives R&D drives better ux drives demand like this is a normal model right this is a web two model so but you guys laid out these five benefits of using crypto I think it was integration with defi rails frictionless payments Collective ownership credible neutrality and Rapid scale I think was the fifth one what are I mean looking back you know two years later how would you stack rank those in in terms of order of importance basically and is there anything that you would take off that list or anything that you would add I think the fact that you can go Global very quickly is a is a big deal um no one who's ever I mean if you're building on any business in which you're scaling out physical infrastructure like going global is extraordinarily difficult I mean dealing with all the legal entities and hiring people and dealing with all the time zones just like that's a huge huge logistical challenge um and Deepa Network kind of the beautiful part of them is they self-organized you write the rules of the protocol you enforce those cryptographically uh and then you just tell everyone else in the world like hey guys if you're interested sign up um and so the ability to to scale globally is is uh a pretty huge deal um and that that's moreover that's also enabled by having global payment cheap global payment rails um sending 20 cent $1$ three payments to people in Africa to people in India to people in China uh is is not doable um and that's kind of the whole point of the vision of these of these networks is they should have uh facilitate large numbers of very small transactions um doing that in in the US yeah maybe it's possible like inside of Square's Ledger or inside of venmo's Ledger or Apple's Ledger doing that around the world like forget about it that's that's not going to happen um so crypto really is the you know in terms of scaling the network both logistically and in terms of facilitating those payments um I think that's really the big unlock um I think the other stuff is is important but those are probably the two that I would highlight Demetri what about you yeah I mean when we talk about scale and and and I think you know crypto is an enabler of that I think programmatic payments are are special um I think the interesting part is to me it's more so like the interesting part is having Laden Supply onto the network and programmatically enforcing that um uh rather than enforcing or incentivizing people to have some Capital outlay um I think the the the capex cost of buying infrastructure I think can be incentivized in other ways he like here's a strawman example um imagine you know Verizon or or you know extend you wanted to um have people purchase you know some some spot for uh for whatever reason what they can say is if you buy this piece of Hardware you will get a 20% discount off your bill for the next year or two that's a very different way of incentivizing the the the buildout of uh of infrastructure and the procurement of that um but those folks uh you know i' like I think a fairly large portion of the population would actually want that um I think it becomes more interesting when you look at more commoditized Hardware that's already online and more internet native because I think that's where having um programmable incentives uh becomes particularly more impactful all right I mentioned them in the pre-roll now I'm going to bring them up again it's arbitrum Santi and I are really fed up with these high fees and we're really excited to have teamed up with arbitrum for the next couple of months on Empire as the leading ethereum scaling solution arbitrum Now power hundreds of decentralized apps across defi hers nfts gaming and a whole lot more the team has showed us everything in the ecosystem both now and what's to come and we're really really excited about it arbitrum allows both daily users and developers to interact with ethereum at scale with low fees and faster transactions the way the team got me excited was through portal. arbitrum doio so my call to action to you is to get started by visiting portal. arbitrum doio go exper on chain like it was meant to be for a lot of Empire listeners your crypto is not just another number on a screen it's part of your future I know Santi and myself feel that way our security sponsor of this episode Harpy takes this responsibility seriously and is the only wallet security tool that Shields users from both onchain threats and sneaky off-chain signature attacks if you've ever been in that situation where you're moving quickly you approve something on chain you realize that the address might be a dubious address or you're really hoping that you could take that back Harpy has you covered Harpy can redirect your assets to your self-coded Vault ensuring they remain completely under your control safe and sound with Harpy's always on monitoring you're not just detecting threats you're actively blocking and recovering compromised assets from malicious transactions before they can even confirm onchain Harpy is the only wallet security solution that protected 100% of its users from attacks like The Ledger one and Q4 which was an offchain signal attack so if you're serious about protecting your crypto Investments it's time to make the switch secure your wallet for free at Harpy doio Empire that's Harpy h r p ie.

[00:35:38]iempire if you want it to be even easier just click the link in the show notes let me tell you guys the way I think about it maybe um and I'd love to hear why I'm right or wrong here I think the simplest way that I've tried to understand dpin is just on a capex Opex like if someone came to you and said I'm building Verizon but I can I found a way to do it with capex and Opex at 100 times cheaper that or or any bu business in the world but I could bring your capex and Opex down by 90% that's a good business right because then you can pass some of those savings you can throw some of that to the balance sheet and throw some of that onto savings for the customer which will grow which will help you guys eat it market share so I think on the capex front uh with centralized infrastructure you have massive upfront Capital requirements um which create huge barriers to entry and just make it really really really tough to continue doing R&D and developing um whereas in dpin you have uh the users are contributing the capital right they're contributing the the capital and oftentimes the labor as is the case with you know someone like Hive mapper right Google has to pay the drivers as opposed to Hive mapper um where the users are contributing both both capital and labor and then on the OPC side I mean how many people do you I've done you know when you call Verizon like how many people are what dealing with uh you know working in the bureaucratic system that is Verizon right there's probably 50,000 people on the operations team of Verizon as opposed to you know I don't know Kyle how many people helium has at their company but 50 100 200 like it is just REM a remarkably different Opex basically of running one of these businesses so I think it really does come down to just you can build similar type of products with you know 95% cheaper Opex and capex I'd be curious what you guys think of that as the argument so so I have an issue with uh capex costs um as the main uh uh driver um uh for for the reason being one I don't think the capex magically disappears you know it still gets distributed to a wide set of a supply site and because the supply site tends to be less corporate more retail I think it actually in some way caps the upside because you know educ cas scenario say you need a large investment to build out some infrastructure in the scale of T to 100 million uh dollars there I don't think are enough people to actually be able to have that Capital outlay um but the moat I think for traditional you know let's call them C pens I think is Wall Street because they're still able to finance these things via cheap debt financing so yes you can have some distributed Capital outlay but I don't think that is really effective at scale verse other methods of financing infrastructure buildout that's that's one and then two is that when uh when we talk about um unit economics and uh and and the cost of these things I think the costs often get reintroduced back into the system as you scale a project and this kind of goes to the demand side I think a lot of projects that are focusing on cost are focusing on the wrong thing because I think a large part of the demand side actually doesn't really care about costs relative to other factors like service quality uptime reliability um I I I have a friend who's the the head of strategy at CDM project that I won't name and and you know like he was leaving and and we're were catching up I I asked him what happened he said oh you know we didn't find pmf and and I said why and he's like oh you know our customers were the the the YouTubes of the world and you know we were going to them and saying we can produce your CDN cost by 80% and the customer said great that's a cost but it's not a cost center for us in the sense that you know we'd be happy to pay more if it meant actually offering a better service especially relative to our competitors so that's a fairly big disconnect um but yeah think those are some of the reasons why I push back against those two points I have I have a fun counter go so uh in the case of helium and Hive mapper I one of the things we look for in deepin Investments is what is the structural cost Arbitrage um that the team is trying to capitalize on in the case of helium it specifically is real estate um Verizon they have to either lease a a space from someone who owns a building to put a tower on top of building or American Tower does the same thing and then they they lease from American Tower but that that that's an explicit Opex cost um Is For Real Estate um and then relatedly for the for the employees around it right so you have to hire a guy to drive to the Tower and check in on the Tower every now and again whereas in in helium if your hotspot goes offline uh you like it's your job to get it back online and there's no dollar coassociated with that so that you actually remove cost from the system in a very direct way um in the case of Hive mapper um obviously that would be the person is already driving around Google pays people today to drive around um obviously the correct way to capture that data is to capture while someone is already driving around so so those are actual fundamental uh arbitrages in cost structure um that and that that's definitely one of the key things we look for when we we do deep in investing um that's part of what you alluded to which is cost the other side dimetri you alluded to which is interesting is is service quality and and that's definitely real not everyone is strictly optimizing for the cheapest cost service quality is is a thing um Hive mapper here is probably the most interesting which is um Hive mapper especially in in major metros they're capturing um you know updated Maps 20 to 100x more frequently than Google is um because people are driving around those roads every day um not all customers need the the freshest Maps um but there are certain customers who really do um the most obvious example would be Zillow not that Zillow is a customer of hi mapper but you can very much imagine a world in the not to dist in future in which you're looking at homes on Zillow and you get there's an option it's like show me a photo of this house from the street as of two days ago right um You can imagine that for real estate agents and everything else so that that kind of customer like Google cannot serve that customer today because the images are too old if you have a photo of the house from nine months ago like that that's just weird right it needs to be fresh um so there's actually certain things that that Hive mapper unlocks that were not possible before yeah and and and freshness is a different value prop than cost which I think does work right so so novel just saying dpin don't strictly have to compete on cost in the case of HIV mapper it actually allow the dpin model allows them to offer something that Google is can never offer um which in this case is freshness so it's a different you know element of the service right I I I do think they will be I think the projects that focus on the net new rather than just competing on costs will likely have a greater chance of success you know even looking at it from I suppose a distributed systems perspective it's really difficult to imagine a decentralized network to be cheaper than a centralized one because of the redundancy and the latency overhead with consensus and and uh and and and verification um it feels like projects that focus on net new are doing something that is harder to replicate um I also wonder though is Google not doing this because they don't want to or because there hasn't been enough demand from from their customers you know I I I wonder what a thought experiment might be if they actually see hey we where you know our API business is starting to lose quite a bit of of of money to Google does like Google Maps makes a remarkable amount of money I think I think um I have a friend at Uber and they I think they paid 50 or six Uber alone paid 50 or 60 million I think this is pretty sure this is public 50 or 60 million dollars over the last two or three years to Google Just for their map service right if you look at um we we almost hired someone who worked at mapbox and they were saying Google basically as full Monopoly on the market mapbox is doing pretty well too but Google's Just Jack their prices up so much every year because no one I mean no one's basically able to compete with them on on the supply side of of mapping all this stuff so uh Google Maps is estimated to be more than 10 billion they don't break it out in in their filings but but there's a lot of analysts who've you know backed into that number various ways and general consensus is more than1 billion Google Maps is $10 billion yeah $10 billion annually um the estimates are I believe also Uber pays more than 100 million today to Google per year 100 million okay wow yeah that that's also out there well that's nuts what about filecoin so why when I think deepin like filecoin is kind of like the first uh the first big one basically um why has the filecoin game taken so long to play out um question I've thought a lot about uh multicoin is an investor in filecoin we I hated filecoin for a long time uh and then did 180 uh and you're back Onin train I well I was never on and off I I started off and now I'm on um wellcoin owns owns some fcoin and we've invested in one filecoin based team and looking to add more um the the core challenge for filecoin I think is twofold one it's technically really really hard to prove that you're storing content um it's just like technically a very very difficult concept um to do and the second I think really big imped to them has been uh it's kind of as Demitri was alluding to earlier cost alone is not enough um most customers don't care about their S3 bills they do care about AWS bill as a whole but S3 is actually a very small part of that um and so being cheaper alone is not enough they offer some other guarantees like content addressing and and some other um redundancy stuff so like they they do have elements uh that they they try and differentiate from AWS on but the reality is is that like it's been out for three years and yeah the demand is not what we want it to be and so that tells you that it's it's not differentiated enough um I suspect what's going to drive file coins growth in the next handful of years is not going to be Snapchat you know ripping and replacing Google Cloud for some part of their functions and switching to filecoin it's going to be the growth of Crypton native Services yeah um hi Hive mapper being like the most obvious example um farcaster would be another one I think these kinds of companies adius and these kinds of things will be the ones who are the primary demand drivers for basically anyone who's trying to unbundle AWS in a distributed way um and those guys will be the the first real customers at scale for someone like filecoin I think once enough of those customers are out there then I you can start to see the web 2 companies starting to switch but I I still think that's that's pretty far out yeah I think they they've done a good job on the supply side you know I think part of the issue also is around retrievability which then is not great if you're on the demand side what was that word that you just said retrievability yeah yeah actually retrieving the files from ipfs and I think that's where you know like yeah S3 works good enough for for uh for people and and I think people just often don't want to think about another problem and and and I think ky's point on more crypto native I think is probably a better way to go um because you might have more novel demand that that cannot be serviced by S3 that has other things that they're looking for um but yeah it's been somewhat disheartening to see uh the the kind of poultry numbers on the demand side with willingness to pay you know I think like like that's something that I I I hope people pay more attention to do you know because again it's it's it's fairly easy to uh to boost trp a a a supply site you know and and and I think the like the argument is also TBD on on on to what extent you know that's valuable you know going back to uh the supply side in a decentralized network being less efficient I think the file coin uh thesis was probably more so that the the network is Le less efficient but it's more of like an economics argument where you have enough Supply if you have just more Supply then that'll drive the market down at the same time it's becoming cheaper and cheaper to run a data center you know and and and and they're selling you know like these mini nuclear reactors now where where where you can use that to power you know your your data center with electricity for the next 50 plus years know so so so I think um Technologies don't operate in a vacuum as well you know and and I think that's been a struggle for them Demetri when you were researching the token side of dpin I don't know if you were able to go down that rabbit hole but there it seems like there's a couple different um approaches that we've seen there's like time based um like token launches there's supply kpis and demand side um token ISS I'm talking really about the token issuance I don't know if you have General thoughts around like the right model uh to do token issuance for a dpin project I haven't gone super deep into it I do think it is good to bake in some demand focused elasticity into your emissions rate because I think that you know every project you have a finite number of tokens and and I think you you need wasting token dilution basically yeah yeah if if if there is you know a high emission rate um but not a lot of Supply on the network um uh then that's you know centralization um uh you ideally want to find a way where you tailor your emission right to some kpis either Supply or demand um I think it's fine for um uh for projects to have some terminal emission rate um because I think it is good to to have effectively some additional ammunition there um I also think it's good to have more granularity I think High uh I'm not sure if it's high mapper or helium has been doing a clustering as a strategy um I think that is a good evolution in thinking you know it's basically I I I look at you know like these tokens as a very powerful incentive and I think teams need to think about how to be very targeted um and and how to preserve the supply because I think you know if you run out of of tokens after you know two years then that becomes quite difficult to to what happened with healing I mean you're basically operating on a spectrum of efficiency of your token and like your cap table with quotes around it and and growth mode right and helium went full growth mode um and had you had very strong incentives I remember uh Kyle you invited us uh down to the helium launch party in 2019 I want to say it was in Austin um and we sent two folks from blockworks and one of them there were some drinks that were had and one of them spent it's like 500 bucks to get this to get the Hotpot and they put it in their home actually in their New York apartment I think they made like 150k from this thing or something um so the uh extremely high incentives to deploy on day one the counter is that um probably leads to if you do that model like helium did it probably leads to what happened to helium which was there was a decent crash in the token price um but I don't know Kyle I'd love to get your your take on this as well like token issuance yeah I mean the the the the challenge heing faced was it was the first one and quite frankly we didn't think it was going to work we were like we really wanted it work neither do we I remember they called us they're like should we buy this thing I was like ah I don't I don't really know about that yeah we we were really scared it wasn't going to work and so uh the right move was to be very aggressive so that if those people out there exist who do believe like you can get them on the bandwagon and get them to go tell their friends and you know get get that out there the thing is the the world has changed now because helium has proven some success and now Hive maper and others uh is that now the everyone in the market who's watching has some reason to believe it is possible um to pull off something of this nature and so you you just don't need to be as aggressive on a go forward basis with with token emissions Hive mapper I'm most familiar with their algorithm for for how they hand out tokens I mean it's it's really very granular and very nuanced they have a a city- level multiplier so for example like if they need more people mapping in Soul versus Phoenix like they can increase the multiplier on Soul um versus Phoenix they can also do that uh time base if they need freshness in whatever downtown Tokyo they can say you know after X amount of fresh time ped this certain region of Roads gets their own localized multiplier um so like all of those are already built into the system and on top of all of that they recently added the ability I think they call them hotspots I think is the term they use um they were about in a blog post a few weeks ago but but now they they actually uh Hive maper rink can now go onto a map and say I want a fresh data point here and like drop a pin on the map and then they're creating a bounty they'll say I'll pay you 20 honey tokens whatever like please drive here um and then if you have the hive map app installed on your phone you can get any you're nearby you'll get a notification right to like go to this address or whatever um so you can be hyper hyper targeted with um token distribution um and so that that that degree of precision is obviously the the right model now again you it's harder to do that for a wireless network than it is for for mapping um but yeah you certainly want to be very thoughtful um in in how you design the algorithm that that hands out the tokens I'd love to ask one more or go ahead Demetri yeah so that like we can also learn things from outside of dpin that can potentially be applied for example the point system because for for points you don't need to ascribe what percentage of the network you are going to distribute right so so you could have Epoch where you have a point system that still serves as an incentivization and and accounting methodology for the supply side and they look at you know a leaderboard but they're not looking at what percent of the supply they're getting so so it it can be you know 1% of the supply per Epoch still based on a point system and my my sense is that might still be quite um attractive to actually um incentivize particular kinds of behavior for the supply set as well it's interesting I uh one more question Kyle for you on helium takeaways um what are your thoughts around open sourcing Hardware uh generally it's good for these networks to maximize scale and credible neutrality um if you're launching a new dpin with some sort of custom Hardware almost certainly the right answer on day one is to do it in house um you need to do this quality just control yeah control for quality control for Branding control for just the early Community you want the community to be really tight-knit um You probably quite frankly need the revenue to keep the lights on uh because like you you haven't you know made any money yet for the business so there's a whole bunch of reasons to do that uh and actually maybe the biggest one is uh designing any these algorithms that verify you know things in the in the real world um is much harder to do when you you cannot assume control over the hardware um so I remember when back when we were first doing our helium diligence way back in the day that was actually one of the the items we spent probably the most amount of time on in our original DD um so those are all very hard problem and that that problem is much harder in an open source environment um having said that if you really do want to build something that's truly Global in scale that's truly credibly neutral that doesn't have a single centralized point of failure then obviously you want to to decentralize the hardware part over time and so helium started moving in that direction I think within one year of the network launching I forget which hip it was maybe was hip 20 or something was where they they started moving in the direction of having third party manufacturers um at this point I believe there's like seven or eight different manufacturers making hardware for for helium networks um so it's been a gradual process but but it's you know been pretty thoroughly decentralized at this point Demetri how do you think about active versus passive dpin basically in Medi land in crypto medial land all the crypto media folks are tring trying to figure out if you can use tokens talking about using Tok like basically building on there's a whole conversation happening about onchain media and um a few companies right now are experimenting with basically paying readers and like podcast listeners tokens um for you know you read this newsletter we'll give you one one token and I really hate that model because it's uh what you're doing is you're basically putting a dollar value on somebody's time so by Kyle opening up our newsletter and we give him two cents I'm basically saying okay Kyle took two minutes to read this newsletter and we gave him two cents Kyle's time is now worth one minute for one cent um and I just think it's a really bad model basically um but I it got me thinking about dpin in terms of like if you look at the successful folks Hive mapper and helium those are passive deepin networks um but I've seen some people play around with active deepin networks I'd be curious to get yeah Demetri or both of your takes honestly it's interesting you bucket Hive mapper into passive because I would think that would be active because they need to actually well you're already driving um yeah I don't think basically I don't think what media is trying to do what blockworks I don't think is is actually blockworks is not going to do this but um what some other media companies in crypto are doing is they're basically saying hey look if we give our users tokens we think we can let's say increase an open rate on a newsletter from 25 to 35% um I don't think or you know we can get 20% more podcast listeners I and I don't believe that works um in the same way that I don't think Hive mapper because people are getting paid are driving 20% more or taking a longer commute to work that's interesting I I mean I'd love to see the data for that I guess my mental model um particularly on this supply side it feels that passive is better because if your goal is to bring on latent Supply onto the network then you want to make it as frictionless as possible so downloading a browser extension you know with like grass you know to to uh to be a proxy um that feels very low friction you know installing uh Telo Hardware feels more difficult and it it feels like the more effort that it takes to actually set up the supply side the more more trouble you're going to have with actually scaling it you know I think it also probably depends on what the what the demand had actually wants um but I think passive just feels like lower friction yeah we we this is one of the primary things we we look for in our deep Investments is we passive it's very hard to have a structural cost Arbitrage if the person is actively spending their time um to fulfill the service howy helium you plug it in once and set it and forget it with Hive mapper it's again the intended use is set and forget um all cont interestingly there's been a handful of teams that have uh tried to do kind of Hive mapper but for drones the idea being like hey let's get updated mapping coverage of of uh drone drone whatever footage the problem with that um Market is the FAA says uh drones need to always have one person who's in direct plan of site controlling a drone um and so in order to have Regional you know mapping like where is the cost Arbitrage and deep in coming from someone still has to buy the Drone and they still have to spend the time to fly it every day or every week or whatever um whereas in helium And Hive mapper they said it and forget it and so you're no longer paying for someone's time yeah um that that's a very very substantial thing we look for nice um all right guys those are kind of the main things I wanted to to get your take on anything um I guess Kyle throw to you first and then Demetri I'll throw the same question to you but anything that you think is we're clearly missing from this conversation that you think is important as we think about deep in in context of you know 2024 in this cycle um I think we've done a good job covering it I mean it's just an area we're we're super excited about um the use cases are are generally easy to understand generally are being going after monopolies um and so everyone hates the monopolies so it's also like a good you know David versus Goliath kind of a story both helium And Hive mapper have even render and the others all all kind of have that that um framing to them so it's something you excited about something that you know hopefully you can tell your mom and grandma about to justify why you work in fake Internet money um they also have the non-speculative they have the this nice non-speculative uh real real world uh you know use case type of thing moving crypto into the meat space type of thing so yeah so great stories to tell to friends Skeptics um yeah we're we're super pumped about it nice nice and yeah I mean I think actually related to that you know it's it's it's really just making sure you know projects Founders uh you know investors think about both sides you know I think a lot of the conversation and narrative has been on the supply side and I I think there should be more focus on the demand side uh probably Beyond deepin um but to also realize that you know crypto is largely a generational phenomenon as well and and I think younger folks get it because they grew up as digital natives but to get to mainstream uh uh scale I for a deep in on either the supply side or uh uh or demand side you likely need to go beyond that cohort you know and and and and thly have that discipline in thinking about what is the Persona of my supply and demand side today where do I think it's going to be in in in three years how do I reduce the friction of of of both sides and are there any particular use cases where the demographics on either the supply or demand side skew younger because I think that like that actually is um a very important factor in and in uh and any crypto user that that we see today nice cool Jens I appreciate it Kyle Demitri thanks for coming on um and yeah I think this will be the first of many deepin conversations we have to see so I appreciate it awesome thanks guys thanks guys everyone Jason here thank you so much for watching today's episode wanted to take a quick second to thank today's title sponsor arbitrum we know you are tired of onchain experiences that have unaffordable fees and frustrating transaction speeds and that's why we partnered with arbitrum you can experience frictionless trades lightning speed and LaGree transactions all for pennies per transaction explore arbitrum expanding ecosystem at portal. arbitrum doio that's portal. arbitrum doio see you for the next episode everyone thank you so much for watching today's episode really hope you enjoyed it we wanted to take a second to just remind you about our upcoming digital asset Summit in London March 18th to 20th Santi and I got your back seats are limited if you heard it earlier in the podcast there's a little competition running at blockworks to see who can drive the most number of tickets so when you register for the digital asset Summit make sure you use our code see you in [Music] London