The Rise of DePIN: Redefining the Compute Landscape

Transcript

this team is clearly raring to go um hi everybody um let's get stuck right in uh my name is Richard mad I'm uh as already said managing partner of fabric been around for a decade uh back in great Founders in this space uh unfortunately not enough in the deepin area but but um great projects like like near and ocean and and pad and so forth look at fabric we've always maintained that the kind of superpower of web 3 um is the ability to coordinate get people to do stuff and deploy their their activity and their resources um uh if you will uh with a token everybody achieves more team get it I made that up this morning you can

have it for free um so this is a space I um I particularly Love Actually even though we haven't made Investments um we're looking at how to of allocate and um Network bandwidth actually as far back as um last century um and we know that obviously the kind of granddaddy of use cases in this space is how proof of work is is operating we're proud of how that's that's uh in place proof of stake obviously now with the Fantastic success of the merge still in a year ago um and but you know that's fantastic but this is it couldn't be a more pivotal time to try and go to the next level because um as we shift into kind of high gear in

the move towards Ai and perhaps even AGI beyond that we will we're finding bottlenecks and those bottlenecks those scarce resources become incredibly expensive and and if we're going to avoid the centralization of power and profits and indeed possibly propaganda with you know companies of you know that are driven by capitalism we need to find a new business model one that is uh decentralized that has the same kind of economic uh impact as decentralized model and I think if we do that we can actually all create a system where the outcomes uh are actually far better than a kind of brittle stoved pipe centralized model it can be adaptive it

can be emergent it can serve the needs of individuals and communities and the World At Large so it's so what I'm saying is I think the stakes have never been higher so with all that being said I'm going to do let the panelists do a little introduction to themselves we're going to start with Rand from iotex and I want to say um he seems to be connecting uh sort of physical devices of pretty much every description what we'd like him to do given his Uber background is to deliver us decentralized Uber and and an open permissionless one and they could all be operating together and we can skirt around those puddles more effectively but overd Rin sure hi everyone so this is Rin so I'm the CEO co-founder of iotex so we started back in the year year of 2018 actually we are pretty much

focused on iot and blockchain you know on day one now after a few few years of iteration so now it's called deeping right now but like the underlying spirit is all the same so iotex is all about building like a deeping infrastructures you know for Builders right so in our reses so building a deeping projects should be as easy as you know building like a def5 or n projects small team with big heart can do like a impact for work so that's our fundamental thises um yeah so we have like a our own layer one like the option compute some identity stuff so definitely like a 2.0 like an entire road map is actually coming along so happy to talk with investors with Builders with people who's passionate about deeping thanks and now on to you Trevor almost as old as me um got a got

a story background making our metaverse dreams a reality with render fantastic hi everyone Trevor from render Network um for those of you who don't know render uh render was founded for by creat for creators really to enable fast affordable movie quality rendering and today thousands of Motion Graphics artists can tribute their consumer compute to run Ai and rendering jobs in parallel at a fraction of the time and cost thanks Trevor and and over to Ahmed at i.net who I've already heard has got more gpus than Zach at his command I think I've got about a million I got the

right number yeah sure tell us a little bit more my name is Ahmed I'm the CEO and founder of i.net um IET is the creator of the internet of gpus basically a new network where it started from a profound question what if accessing GPU compute power for AI is as simple and as seamless as accessing the internet U and from there um we built the io cloud and all the iio coin and the iio worker and all the multiple components that basically built a compute layer on top of multiple networks on the network uh on the crypto industry like uh render and others um and from there we basically built an infrastructure layer that allows anyone who want to run large scale AI workloads

to seemlessly deploy a cluster of 5,000 GPU or 10,000 GPU which you can't find anywhere on the planet that's ironette I'm very excited about it thank you and and last but not least Ariel um love to hear about um um mapping the world and the Honey Pot that you've created there for people all right so high MPP is a very simple idea um we are building a new Global fresh map um to compete with the likes of Google Maps to power both people in terms of their day-to-day navigation businesses right so Uber lift are obviously massive consumers of maps um and ultimately navigation for you know people just running around their city and so forth and so the the idea is

pretty simple is you buy this device from us and so you install it in your car and then you just do what you a lot of people do which is drive and so all of the imagery that's being collected here is then being processed so that we can actually understand okay there's a 35 m per hour speed limit sign there's three lanes over here there's road construction here that's all happening on this device through a series of map AI process and then we build that map and then we go and sell it to various businesses and organizations all across the world um we've mapped about 18% to the global Road Network in about 15 months so it's clearly working and the Beautiful part is unlike a ways right or many of the other predecessors where the

people who were really building the map never shared in the economics of it when you're building this map through this device you're actually earning our honey token and so you're sharing in the underlying economics of this map that we're all Building Together brilliant okay so um so we're going to going to start back at this end but then we're going to jump around from then keep it unexpected and um I'd love to kind of surface a couple of questions from the audience kind of towards the end if we can I got a special request from the panelist to make it interactive so we might ask for kind of two or three and then kind of maybe pick the best if that's allowed um so you know back to the key question here so we spoke about how this new business model driven at the heart of which is obviously tokens

and token economics um you know it's essential if we're going to fight this good fight against centralization and um but but with that there is a feature not a bug which is um its appeal to speculators uh so I want to start with the question of how do you harness the power of that speculation but also on board folks who are really going to be sophisticated and help I'll move to the second question which is like really you know deliver the map or you know deliver the gpus and so forth so I don't know ran do you want to go first yeah sure I think um you're talking about speculation right but speculating look at from other side is basically like Capital they gave you like a initial

like Capital injection when the projects is really small think about you know you have like a three friends working together you have great idea to conquer the world right where the capital is coming from you don't want to go to VC on day one you know trying to do a hard sale hard pitch but you might TR try something by yourself maybe having a token maybe have a small community maybe have people believe in you I think that's kind of like a magic about about crypto uh about deeping especially you know iot networks has been known for you know have a huge upfront cost by Design manufactur distribution of the devices then you will see something will work or not right but deeping I think actually have this magic of like crypto on day one they gave you something you can try it out right maybe with some small

initial Capital you can also prove to your investor to the to the world something already works so I definitely think a really good thing for th and crypto good so it gets the capital so for us it starts by abstraction first having AR never have to deal with crypto onboarding them onto web 3 from a Fiat perspective where they they use the network to pay Fiat for rendering jobs and then the the great thing is it's kind of an onion they can peel back from there where they can contribute their node become a node operator and as they do there start earning and learning um what we do on on the crypto side is it really comes down to choosing the right model or or tokenomics and then being able to put the incentives to work uh

for various players in the ecosystem sy so a great example for us was we um upgraded our chain recently and and in that process we were actually able to put upgrade incentives to the crypto side so they could help us move over in a way that that really harnesses it excellent H well I I mean look if if I have a data center that's like a100 million and not too big um of h100 and if I just like basically buy that GPU power and I put it in the that Center and I'm waiting for my sales team to bring in long-term contracts to sell the gpus here um generally I'm like this time I'm losing where I'm not the my my

device is not hired my gpus are not hired for any workload whether AI workload or whatsoever workload um that is like this time there's a Time loss like opportunity loss here but when and data centers like Amazon or Google cloud or whatsoever when these Cloud providers when you want to deploy a cluster takes like 15 minutes to just boot up a node because it's offline they want to reduce electricity and whatsoever for us it takes like 30 second and you're ready up and running so imagine the speed difference but the reason why we have that and reason which gives an incentive for that Center to run on ION rather on on AWS if it if it has the chance is because on AWS there is this time that's gone right I'm not gifting anything

until someone comes here but with a token system it's so interesting that you can and emit him some kind of rewards coming in from the io coin rewards small enough like to keep him stay you know keep Cal the electricity bills are being paid you don't have to worry about your capex and so on until a job comes to you which basically if I'm just like that like if I if I'm not hired for the whole like two years if just my GPU is there I basically bring back the value of the GPU just because I'm on the network I'm providing that Capital compute Capital power to the network that itself give that someone like hype like I want to be get that coin uh and be part of such such thing and then you know reduce my my expenses

that's why I like the token tokenomics and deepend projects I love it I think I think that's super powerful you can see how that can deliver you know cheaper but also more adaptable than the centralized model using token economics Ariel yeah I'll give you uh two examples so the first one is this guy Brad so Brad basically bought I think about a hundred of these devices pretty early on actually when we were just launching and then he figured out who were the right kinds of drivers for the map right so obviously Uber drivers lift drivers and like that and so he bought these with his own Capital right but then he distributed them to all these Uber drivers lift drivers Amazon Flex drivers and then he split the tokens with them so that really helped the network grow

because he was identifying the right kind of driver he had the capital right a lot of Uber drivers lift drivers they may not have the capital so that was one really clever example and he's now taught other of these entrepreneurs who come into the network saying if you want to replicate what I'm doing here's the Playbook so that's one example the other example is now more on the commercial Fleet side so this guy who operates I think about a thousand car rentals uh in his Fleet he bought a thousand right and so he's putting them in all his different car rentals all over various countries and so that's a little bit on the commercial side so both of those are really amazing because they help the map grow in a very effective and cost effective way with while also rewarding that person who's putting a little bit

of Capital At Risk in terms of buying these devices that's great that's super interesting because like you know different takes right the way across across the group so so the next sort of question area and take it you know with examples wherever you want to go really is is this question of matching supply and demand um so you know for example you know we don't want to map the hell out of Central London and have a you know 100,000 versions of the same map and obviously there's many kind of layers to kind of what you are exactly mapping and but yet have different areas that are not mapped at all and I guess another one would be there could be a lot of uh demand currently for mapping the puddles around Dubai like how do you react like that and the same a different

use case but on the GPU side how do you you know deliver low latency for inference in one part and how do you deliver you know the high throughput um that you need for training in another so uh do you want to go first AR work back the other direction yeah so so really cool example so we're working with this autonomous vehicle company um and they were tracking across four different you know cities in America like Phoenix and Austin and Houston and whatever they were tracking a specific type of issue it has to do with construction work zones and stuff like that so let's just take Phoenix so they said to us okay here's about 3,000 locations all across Phoenix that we want to see that location or we want that location mapped

three at least three to four times per week okay so the the the network in Phoenix Metro area for those of you who don't aren't familiar with Phoenix Phoenix is massive it just like never ends right and so the normal Network hit about 80% of them right three to four times a week which is pretty good but then in order to get to what the customer wanted which was about 98% coverage we introduced this concept called burst which is if you take a five or 10 minute detour off your normal course then you will earn a little bit more because you hit that specific intersection where we were interested in and that worked amazingly well we took it from like 80% coverage for the three or 4,000 locations to like 98% coverage

but again three to four times a week right so that's like you know a lot of coverage per week for you know 3 to 4,000 locations so being able to identify a very very very very specific location and then say hey if you're in this are area take a 5 minute detour and you'll earn a little bit more you can go down to that micro micro level and there's also things that we do more at the macro level in terms of like hey we want more of this region less of that region for various reasons fantastic examples and we'll come back to it maybe but I I'm I'm intrigued how can that be ultimately be automated so like the network itself works out where it needs to change the incentives to fill fill in the gaps I mean you can go now or um back I'll just

answer it really quick in terms of how we're doing it so like as customers are hitting the map and we're not returning good results we're then basically Auto creating those bursts if we think there's not enough demand there based upon what we're seeing on the demand side yeah so I'll pass it over to you I think having Supply figured first was it's more like the chicken and egg problem which we are like living in live um we first have few clients were like hey we want 3 A1 100s we want four 400 sorry 300 A1 100s or 400 A1 100s um and we're like we only have 300 and but they want 1,300 or, more so we didn't have

that Supply but then once the token incentive came in uh we managed to grow the supply what actually happened is really interesting so a client came to us reached out to Twitter some startup and he said hey guys we have gpus in Lambda cloud and ws and but our use case wouldn't fit hosting the gpus there and they don't have GPU Supply in Korea South Korea or Japan because we have most of our users there but there's no gpus we can lease there why because their app and and we're noticing this trend now like we're moving from waiting 5 seconds for the image to be generated to real time I want AI real time imagine if it's like Sora like I'm playing the

game and it's real time changing the environment then meaning that the industry is moving to more realtime model inference so I need the generation so fast and that can only be solved if you have gpus in every country where there are users of that product so this company that came to us said we have users in Japan and Korea and we want if we will do gpus in Lambda Cloud which they and it will take 300 millisecond for the trip geographically it's going to take time through the internet to reach there but our requirement is 150 millisecond for the image to reach generate and come back so we think you guys have gpus in the countries where there are no gpus so we

would like to deploy our model geographically closer to the end user so then suddenly we found out wait did we just build the CDN for ML models meaning that we basically distribute the client model on on 139 countries and then if a user came in from Saudi Arabia or Dubai the trip doesn't have to the the the request the computer request doesn't have to go to openai Microsoft back in SF or east coast and so on it could just go to the nearest GPU provider here right and that basically gives a 20 millisecond latency which was really amazing and and they moved all their infrastructure out from such providers because just availability but we couldn't reach that availability before

we kind of crowdsource this from the crypto Community like all of us work together to build this massive you network of gpus in in a lot of location around the world 140 countries and then from there with that massive Supply suddenly we found out we actually built something no one actually can build without having 6 700 800,000 GPU it's like no one have reached that level of having 800,000 gpus all across the world like maybe yes Z zberg has it in SF and his their data centers in Facebook but it's in one location here we have it you know across the globe which is something you new like we have been pushing the Frontiers like we have going where nowh

have gone before so it's really interesting absolutely exciting and I I was just actually um catching up with um zuck's announcement in the last couple of days llama 3 U you know being um uh open sourced which we're you know very supportive of and but also I haven't tried it yet but apparently you can be you know talking of inference and latency you can be typing your prompt to generate you know image imagery and as you type the image is is is changing so like I mean that's the expectation we're going to have when we shift to a generative model and this is one of the things I took away from actually the the keynote at um the Nvidia keynote from Jensen the other day like we the whole architecture of you know the physical part of the the the internet is going to

have to change to adapt to that generative model so I think that's exactly in tune with what you're you're talking about Trevor wow well scaling a a two-sided Marketplace is hard it in fact it's one of the hardest problems you have running one of these networks uh typically you start by getting enough supply of of compute out there so that people can onboard and use it but at any point if you break that you get into a situation like we had this Le this week where you can't find nuber and you can tell right away it's it's not in Balance um for for us it was it was really focused on first getting that initial batch incenting those users to stay those notes to stay and then going about it from an artist perspective and ramping that up and um it's been a phenomenal run while we've done that

we've had a weight list that we've grown it's got to about a million GPU so it's it's been an asset that's built up outside without breaking that balance and the the great thing for us was this year we were able to actually open that up again as we started moving into Ai and bringing on Partners like IET who who were then able to leverage it in a different way but it it's a really hard problem and honestly you're you're out of balance at any moment in the system so it needs constant Focus I just I want to shout out for render because that moment when that client came to us we were like no one knows about us ion was on Steals and I called Trevor like look we have gpus in Canada and Iceland from you know through

fcoin SPS but we need gpus all around because this client came in or like but we have Community everywhere let's let's plug them in quickly and so we plugged in their gpus and that was the beginning because they already have the massive Community which is all around the world and without them honestly that have wouldn't have happened so shout out to them good job yeah so that's kind of interesting question about like a DP usage right or use cases um yeah so deep is definitely like a design for people used by people so we're not saying like a deeping replace like a entire physical infrastructure on day one but it actually gets like a fill into some small niches here and there I can give you some examples so there's a project

who's doing like a a Tesla like a war charger sort of like a Airbnb for Tesla War charger on iotex right so idea is pretty pretty simple so they have validated this idea within a small community basically you can share your War charger with other kind of Tesla owners and they charge ma that actually works in a small environment then they're kind of like a scaling up using iotex as an underlying infrastructure so that's is like one use case is how ding can make this like a Tesla charging more efficient more convenient and cheaper and another example we sour so far is uh it's called Network 3 it basically use devices you know they have people to running their server uh R running their software actually on their computers during the idle times and they measure

like the internet or scraping information from the internet in a decentralized way right and all the information this script will get aggregated all together to produce like a image for internet so this is also another very interesting one that is not has a huge t or like a market right now but it's actually used by a lot of people you know it's kind of useful I feel like deeping will start you know here and there and eventually we're replacing the world we're seeing like a web through version of The Uber one day one day one day yeah hopefully soon so um so obiously Uber had its own challenges with regulations and different geographies in the in its early days since it kind of you know pushed the frontier um I'm intrigued and actually we we didn't really kind of do

any pre- discussion on this but um you know it would seem we've you know we're all enthralled by the power of the token economics here to get people to deliver this new physical infrastructure and that it's it's going to be you know very powerful um what are the challenges or questions that you wrestled with from a regulatory perspective um with you know unleashing that power you know perhaps commentary on different geographies I mean I think we know that Dubai and the UAE working very hard to make this a um a welcoming environment for from that perspective so I don't know want want to go first go yeah I mean we have it from both so we there's obviously the crypto regulation side which is what you're

referring to but you know uh we have also because we operate in the physical world and these devices are driving around in all of these different locations um we also have privacy regulations as well as naso security regulations right so South Korea for example has very specific um regulations related to their National Security and mapping and where all that data can reside and so forth right so so that's a a good example Europe as many of you know has very very very strong uh privacy regulations right so there's a whole set of things that we need to do in terms of processing all of the data on this device you know so we have to privacy mask effectively blur

all of the N uh names of specific objects uh people's faces bodies cars license plates all that kind of stuff and that's why we're actually doing a lot of the actual map AI processing in terms of detecting okay that's a 35 M Power speed Lim sign there's a construction zone there so that in some locals we're not even uploading the imagery right because we're not allowed to and so that is you know depends often times by country and even within that country by specific state or region of it so you have to be able to balance all of those different privacy regulations in addition to all the crypto regulations so we kind of have it from both sides do do take it from that that you you're not commenting on the crypto regulations I'm not commenting on what

cryp the crypto regulations yeah crypto regulations I leave I'm an American I got I got to leave that to uh to somebody else okay but it is interesting on the on the Privacy side I think obviously you in particular you're not just coordinating physical um infrastructure and activity you bridge across into the question of of data um and and I think uh all of this Innovation taking place at this layer is going to be pivotal to try and make sure we can do a good job at the data layer as well where one of the the challenges is that it's it's it's possible to kind of gain the system like um when you're trying to Define what the truth is I guess in your case you know the image is the image and can it's deterministic and in Trevor in your case as well it also

is um but you know this question of privacy and the truth is going to be a big one for the for the kind of layer above the physical yeah I mean um there was an image that was circulating um that the Pentagon was attacked right and they they it looked pretty real right um and the markets reacted to it it turns out that that was image was obviously fake that the Pentagon was not attacked and it wasn't blowing up and so forth but that is a major concern is if somebody were to hack into these systems and say hey Hive mappers detected that whatever a certain part in London blew up or was destroyed that would have severe ramification so yeah definitely something that we think a lot about is like how do we verify that data that's coming off this divide is actually

accurate you become a news agency or you become a source of alpha for hedge funds maybe we already are maybe you already are exactly um good so any Trevor do you want to comment on that that uh truth verifiability question at all yeah yeah I mean so much in that on on the regulation side I think decentralized gpus really help this problem number one to start with because um you can then do it on a country bycount basis you know we're we're in the middle of a a real battle for gpus and so being able to actually you know go only choose some in certain locations I think is a real benefit here um but when it comes down to the the data side that you were referring to um for us and and our

artists it's really critical they are angry because they have centralized companies training on their data on their artwork um and and that's something that we're really looking to try and help change we're we're actually moving towards uh generating synthetic data on the render Network to be a base model for people to be able to train without it having stolen data in that model and then layering on top of it and and using the blockchain in a way that is verifiable and and and really goes to the core problem we're about to have here when you talk about images we're we're about to move into a state where um generating an image or a video is Trivial and there's going to be an abundance of these everywhere and and in

that case it's really really hard to distinguish between what is generated and what is created and and so using blockchain for me is is fundamental to this to show proof of creation to show proof of render um and and is you know one of the reasons that really drives the render Network on on these problems yeah I guess it's a good point that that you know delivering a truly decentralized physical infrastructure helps with the argument around crypto regulations that this is um not a security but it also helps us deliver the infrastructure that we can um perhaps trust the trustless infrastructure we can trust to be outputting um um data that around which we can get consensus on the truth um so

you know it's it's a a foundation that we definitely need and on your side on again I'm trying to tempt people to get into the crypto regulation conversation def cryto regulation is super important I think th is uh least likely to be considered as a security right because everything we talk about is about utility right like for transportation for your data like for Ubi even those things are all utility but you mention about I think I think one point which is very important is D verification so this is a layer which doesn't really belong to the blockchain right now so that's kind of like one thing we're working on in the entire like a deep in protocol stack is you know you have to have an ability or another Network which is offchain that

collects all the data or information from GPU from sensors like this right so you have to have a like a some some some people to put the rule here into this layer says okay so here is a rule if this data points shows that can passes the rules it's considered as like a valid data points this can be using like a z snark this can be used like a te there's bunch of Technologies we can use to provide this validated proof this proof essentially goes to the blockchain says oh so this device has done its job or this GPU has done this computation or rendering that's why you should gave this guy like a 10 tokens right so this layer using any blockchain you cannot do it because it's computational intensive right now that's

why like we are kind of working with few teams you know here and there they're doing ZK they're doing like a GPU we're trying to packageing all the thing together to provide this de uh de verification capability I think deepen has a very challenging approach to the regulation because you have to fight at like it's a battle of two fronts the first first front is you're fighting on the crypto normal regulation if this is security or not security and so on and normal crypto regulation for custody and whatsoever then the other front that you're fighting at is the new executive house order and the infrastructure as a service regulation so once the you know once the US regulation in terms of the uh Commerce or or um the NSA or so on like once they notice that there are

some people building some layer that basically takes out monitoring and censoring and control out of the government hand in a way and making it more private it becames really really sensitive and especially with the geopolitical tension between China and Taiwan and all the other things regarding the GPU um bans and on exporting so all these embargos that's happening on China it's causing um any infrastructure as a service labeled provider to to soon to ABY with new regulation that's coming out that's not only you have to do the normal crypto uh stack but you have to add to that the infastructure service tax so for example

like this new regulation it's an executive house order and if you provoke it if you if you you know if you do it you provoke the patriot act like this how serious this is so uh the for example like what happened something happened to us so when we launched we had like 40 to 80,000 G gpus like not we launched like two months ago in March February we have we reached that number um then a workshop happened between NSA uh you know all render Solana Foundation Fon they summoned everyone in the deepen ecosystem immediately everyone in the deepen

ecosystem was summoned like guys in a way where like this is could go up out of hand another way around and like honestly you don't want to give this compute power to China period so and that was when 880,000 gpus now it's 800,000 gpus it's capable of 51 attack on every GPU Network in the crypto Market that's how dangerous it is so imagine how much attention from hackers they want to access this compute power and do a 51 attack on ton coin or do a 51 attack on uh other GPU based mining algorithms that's one thing the other thing is with that massive compute power

you could you know train a machine learning model that is hacking that it's malicious that does malicious things so how could regulation now that's the new regulation how can we as a regularity be careful that you are not going to give the service for D5 sanction countries embargoed countries from accessing massive compute power without kic and long processes how can you basically the compute power it's going to you're going to come to to our platform one day and you will go in it's like level one level two level three kyc and each level will give you what amount of GP can put Power you can access it's not I want that it's if I don't do that they're going to shut me down it's there's no joke here like it's it's they will really shut me down and if they are so much serious about

this more than they are serious about money lering like defi normal someone doing transactions without knowing know who is okay but someone accessing 100,000 gpus without knowing who is oh no no this is Patriot Act invocation so now we have to like battle at the front like we want privacy to our people we want what is the right way for the internet of gpus to be owned for the world like it's owned by the people for the people in a way where it has like verifiable credentials and whatso on that keeps the Privacy at the same time but also abides with the regulation that could you know make big trouble for us so regulation for deepen is coming and it's coming really hard and it's coming

so fast it's faster than the SEC it's faster than this these guys are like really serious like no joke well okay so we thought we had a hard time with polit politicized regulation good Round of Applause wo yeah um and we had we thought we had a tough time with regulation sometimes being co-opted and politicized we we you know around Defi and um you know Securities and so forth um I think we've now ticked off kind of like technology sovereignty the intelligence Services um you know Kym we got all data privacy we got trouble in on every on

every quadrant we're surround we're surrounded and it sounds like it's going to get more complex at least rather than than less complex so that's good we're in this we're in this for the cause for the excitement as a as a community together we will you know do it slowly the right way owning our data having the ability to for everyone to have access just like the internet we all should have access to compute power that's it like anyone should have access to compute power no one should limit me from how much compute power should I access so how could we do it in a way where it's the network verify that the user is not doing anything malicious or trying to D be dangerous to others and but still giving that freedom to people

that's I think the future of deepend 100% certainly if I if I build on the um examples given by my two daughters access to compute power in form of their iPad is a human right and they need to go get it but it's applies to everybody else too no Ju Just on the crypto regulation probably was didn't say this too loudly but it's uh you know six months ago a year ago in DC it was like crypto has no use cases right like it's just meme coins as this as this as this as this right and so all of a sudden it's like hi mapper shows up ion shows up render shows up these are real use cases across massive markets and all of a sudden at least in DC you know that part of the argument seems to have gone quiet right and so like you got to show

up with real use cases we still need regulation obviously that is clear and all that kind of stuff but I think we're doing a better job educating those folks of like no there's real utility going after really massive markets against incumbents and so if you just want Google Maps to own maps for the rest of the world you know for all time then sure that's what you're going to have if you want competition this is one way to get competition yeah no that's where I put it helps have a good use case to evidence utility would you see good beginning of the conversation Trevor do you want to comment yeah yeah so um I think we're we're kind of Cruising Into the last five minutes um what I'm intrigued about is that you know you know as has just been commented this has

really started to take off fast just in recent you know times I think partly obviously driven by the maturity of the the web 3 Stacks the layer one blockchains uh we can get into a discussion about which one is which one is uh your Preferred Choice some moves to salana here um I um I'm I don't have salana bags um uh but you know we could definitely free feel free to comment about whether other um l1s you might want to go to I know on the meme coin side there's been a lot of activity on bass for example um and and I and I am a big supporter of near how but how do you define now as it's the activity is surging when you feel like you've got

some level of maturity in your core core model and like in these last few minutes where are you going to take it from there you know what what is the next level of innovation from from your perspective beyond the kind of the core utility that you're you're delivering yeah so I think one comments on the mean coin I was thinking about why you're asking is like mecoin is actually you know willcome by people right if you if you think about it's so simple and also carry some sort of like a emotion people want to express basically so that's why like uh it's very much understand by people and likeed by people by the community as well right so I feel like a deeping somehow carries the same spirit it's also like a use by people and should be

like understand by people because oh this is like a web version of uber this is web through version of ARA BNB so on so forth um but I think like a deeping problem is like a scale or scaling is really slow right so if you see uh you guys are doing great job B basically you know scaling like a high mapper down from scratch but we see a lot of projects especially you know on itex you know so they have like a this like a very steep curve to grow into a massive like a scale so people can understand and how can we educate the market so people can understand you know so this is some real right so we should trust this token I think this is like a biggest challenge in deeping and we should find a way to kind of come over kind of education yeah Trevor I I

think you're you're chatting to three of us are are on salana here so I I think that sort of shows a a definite impetus there on on the deep inside I I think for us there was a technology play that rust made a lot of sense um helium had just moved over and had a very similar model so that was good but a big part of it is also low cost and high throughput and when I look going forward I I think a lot of our original premise was putting proof of creation on chain and I think you'll see a lot more of that happening as we go forward more real time um you know leveraging that leveraging the data creation on chain um and and using the fact that we have a a high throughput lowcost system to do that our artists who create things create things

that are put on every chain and we'd love that to persist there are a lot of great use cases outside of that Beyond just the the creation thank you yeah um so today hi so we use salana we love salana you know the way I look at it is it's infrastructure and so the less resources from an engineering perspective that we're spending on salana the better right like if 25 30% of our team was on salana and working salana that would be bad right so today it's effectively one and a half people which is amazing right like that that's a good thing that it's one and a half people um because that means is the rest of the team is focused on product and value and utility that we can provide

our customers or our contributors so today all of the reward trans transactions and the consumption transactions in terms of people consuming the map that's all on salon on the blockchain we're starting the process of moving some of the map data itself to the blockchain right so people can say yes this is verified this location exists at this location there's construction and so forth right many of you probably had issues getting here I did yesterday and so when I got in the car um it said yeah it's going to be a little bit delayed usually it's a 30 minute drive from my h Hotel it was going to be a 40-minute it ended up taking me about 3 hours okay because a major road had been flooded now that

road hadn't just been flood in the last hour but I say to myself okay when Hive mappers in this fully developed State we will know about that flood right we'll provide Visual Evidence of that flood right because we have eyes all over the world and then we'll really understand that no no no Ariel this isn't going to take you 45 minutes this is going to take you 3 hours because boom here's 500 M of this road that is flooded with 2T of water and so how do we do that and get all of that data on chain that will take a a year or two uh but it will ultimately get there amazing it's going what happens is anyone got a burning question I'm probably not allowed to do this too many

questions one at the one on the back it's a very packed room how exciting is that it's like really you guys have really been um you know um you know to I think this is like real use cases practical physical relatable uh with real opportunity to to be involved and to earn and benefit from uh being part of these um these systems um and I think in a testimony to all of you for what you've uh managed to to deliver I can't believe this many people are that shy um but but you know I saw one hand at the back um but it any any final points I might we spit we um jumped over you at the last moment there

any final points you want to make before we wrap up me so so so for I think AI is going to be 30% of the coin market cap very very soon and I believe it should uh it's one of the most important things that's changing life it's like basically bringing internet to be 20 saying internet is 20% of crypto makes no sense it's everything but I I think AI coin market cap will be at least 30% of the market in the next two or three years look it's been super exciting you know being able to do this with you guys and get to know you very rapidly I said behind stage to people that we're actually not here for the money we're

here because we care about the things we're building I really feel that from all four of you and um I I think for that basis I want to have a selfie one H second you ready and I want a round of applause for everybody on the stage and thank you very much for your attention sh