Appchains Vs. L1s: What is Crypto's Endgame? With Illia Polosukhin

Transcript

I have a feeling you guys are going to go at it and I'm going to try to I'm going to try to mediate but Ilia is going to tear it apart and rip me a new one as well as Rob so that'll be fun so the answer there is actually users right at the end it's all about users welcome back to the rollup for today's episode we had Ilia on from near after he saw one of my tweets saying that every app will eventually be their own chain he pushed back on that idea throughout the first half of this podcast and really illuminated why he believes this is a flawed model it was really interesting to be able to go back and forth with him for about half an hour and kind of hear his perspective throughout the second half of the podcast we talked about why

AI matters in this conversation and where the impact of AI comes in we talked about the future of user owned Ai and why Ilie is so passionate about this sector of blockchain we really enjoyed this conversation hope you do as well let's Jump Right In after a quick word from some of our awesome partners [Music] cartey is your home for app specific rollups at the Forefront of developing new tools and platforms that enhance the capabilities of blockchain Technology carty enables applications that are flexible and built to scale build on a full Linux environment and use

traditional software libraries but in blockchain development go to cars.io that's cy. to learn more about building your own app chain today welcome to the ignition Mantle's newest campaign to liberate the growth power of Bitcoin fbtc ignition has officially begun with a total token supply of over 2,000 fbtc with a tvl of over 150 million right now you can go on the fbtc website to see all of the featured campaigns staking opportunities and liquidity pools where you can utilize fbtc on chain today for some of the best yields available on l2s the

ignition Sparkle campaign can be found at fbtc ongoing Dash campaign going on guys welcome back to the rollup uh today we are talking about why every app will eventually become a chain and why sharding is not going to work for that or at least that is my perspective but Ilia is going to tear it apart and rip me a new one as well Rob so that'll be fun um and yeah I second time back on the rollup so happy to have you back and um yeah man how you doing yeah thanks for having me uh I'm doing pretty well we've had a pretty amazing August so

far uh with launching chain signatures and night shade V2 and so now you know I'm ready to talk how the future will look like and we have all the tech actually there yeah we had Kendall on to talk about chain signatures and he gave the the approach of uh chain abstraction in that perspective um and so maybe we'll touch on that a little bit but yeah I have a feeling you guys are going to go at it and I'm gon to try to I'm gonna try to mediate so uh yeah I mean I had this angry this angry man in in in the comments here um basically saying that like sharding is going to be better

scale than any single chain and I kind of responded here but I think there's even like a a bigger tweet um but I just want to explain my perspective first IL if that's okay and I I know this tweet is like pretty polarizing uh because it's like very like black and white which is never the actual case but is great for engagement bait I think but like my general idea here is like okay so if you're an application on ethereum L1 if you're an application on on a L2 if you're an application on even like like like a very highly performing L1 you're giving away in exchange for composability and network effects um from other apps on that you know kind of like Universal chain you're giving away

like a ton of sovereignty by means of um sequencer revenue for l2s by means of uh controlling transaction ordering by means of having to worry about whether or not there's going to be certain uh events on these chains that are are going to cause the experience on your app to be kind of um problematic in in in in the sense of like nonp parallelized chains um and I mean there there's kind of a whole host of reasons why uh application Builders I think should be launching their own chains uh from from that perspective I just think that people that are building chains or apps on universal chains are thinking like a tenant who is renting a property

from somebody rather than like a landlord who is acre Equity over time it's like you're you're paying to be there rather than owning and and kind of earning with the kind of uh value of your block space so I I'll just pause there and that's like my general take now I think there are instances where this may not make sense and I think this tweet is pretty kind of black and white but I'd like for for the purpose of this to kind of stand pretty strong on this and not just give in so yeah IIA definitely curious uh as to explain your Tweet a bit and yeah for sure yeah I mean my my response was like very specific but yeah I mean I think there's like a whole spectrum of things to

consider and the the specific thing that I wanted to you know start with is definitely the reality is every single like single chain Tech right now is actually non-scalable right and I think like every time one of the chains actually starts to get to a level where they actually have users they have ton of trouble right the Bas is good example recently um and like there are a few examples before as well uh and so the the challenge is like none of that you know op stack arbitrum like or even Cosmos kind of uh SDK is actually able to handle like

million plus or like I mean even for example 10 million users right and if we are talking about like truly build kind of you know next Generation applications that's a scale we need right so how do you do that well you do need like you just cannot fit all the user data and all the transactions on a single machine like that's what all this kind of approaches do and you can buy a bigger machine for sequencer for your L2 you know and kind of uh potentially scale that way you know plug in more ssds there's actually like actual limitations how much ssds you can put on one computer um but that that's kind of the the the story there or you can actually go the

way that actually scalable applications going which is sharding right and so so this means like okay well you need a sharded blockchain to achieve this and your the the response kind of here would be like sure let's just grab near for my application right let's just have a near instance per application problem solved right near and chains H so basically app chains so basically app I'm going to I'm going to get you there but I'm I'm just saying like that I'm I'm I'm I'm showing you right away your argument right which would be like let's just have a sharded blockchain per every application right and so the answer there is actually users right at

the end it's all about users and the kind of like you I I I like this argument of like hey you trading off sovereignty but the thing is like this is not your sovereignty to trade off as a developer you actually providing a service to the user this is their sovereignty to tradeoff and the fact that you have an app chain that now you're controlling your user sovereignty is actually the opposite of what users truly want users want to have all of their state as as kind of composable and as close to each other and under the same correctness and guarantees as much as possible Right I think that's that's to me that's

actually the fundamental trade-off right and so and as we again as we doing the sharting and doing all those things like you want to have that tra and that and that's why I also think the layer twos are kind of have this limitation as well because now you have all the different levels of sovereignty and security and you know who who is responsible for what this is actually back in 2019 when we were interviewing Zaki for for Cosmos for the Whiteboard that was exactly example like I have my crypto KY that I own but it was bridged from a different app chain who is responsible for this right like now I as a user lost my sovereignty like Lo lost part of the sovereignty of

my assets right and so so this is why like for us it's it's all about this kind of the core value we provide is users owning their assets and data right and so everything we do is kind of how do we maximize that and kind of minimize the leakage of that for the user and so in a way developers are just facilitating that by building more protocols and more software they themselves don't need like as a developer you don't need sofware into you need to have a best developer platform that is you know easy to run on and then you can give users the functionality and so so where this gets practically speaking right is Integrations right okay cool you launched your own uh let's say in

charted blockchain well now you need to integrate everything and again as a user now this doesn't work this doesn't work it's not you know it's not propagated there it's not in those walls right you are losing kind of all of the like user state that I cannot easily bring and we are actually solving some of this this is why chain signatures are being a big push is because we're trying to solve some of those problems kind of on a web3 scale like how do we actually make it such that user has their own account that can can actually transact on other chains but again like every time you launch a new thing like it still has different set of parameters it still has different uh kind of you know additional

latency for transaction it still has you know rpcs going down indexers Etc and so the other question is cost right because this cost at the end is cost that bued by user and right now I mean right now there VC is kind of funding everything so uh it's a little bit hard to like uh like properly evaluating this but you know you can go back to the math right like user paying for a transaction right that cost needs to cover validators running it you know whatever settlement whatever the You Know da whatever is the guarantees you have there all the Integrations all the software maintenance all the sres that are you know sitting there running validators

and all of the other costs that kind of and there you have a very much so uh benefit of having you know all of that on the same shared platform right like the more kind of people transacting the like the reality is like even now like even on very used platforms right like we still subsidizing kind of all of those costs so like every application just will not like will always be underwater and we and we see this right now right all of this applications they like using their token incentives but the reality is like there's not enough Revenue being made even with sequencer Revenue Etc uh to do that now sequencer Revenue me Etc all of that is addressable for per specific use

case right so you know for swaps we all going into intent based swaps the intent matching is happening off chain you already don't have onchain like validators cannot capture M uh like there's all kinds of ways I mean we discussed with tun on whiteboard how to do me protection for UNIS swap like amm pools as well where you you know you you satisfy in price and you take off difference to your stakeholders instead of to mbots so like all of this is like you know solvable uh for a specific use case and so and you can actually again so like the costs are kind of shared between all the applications the

sovereignty is given to the user and the kind of Revenue Etc can be you know generated in in a more systematic way with all the Integrations to exchanges and all the stuff right it's still super hard to get like new things going right even if it's in on dexas because somebody needs to list that and we just that Etc yeah okay so bunch of really really great points there um kind of the first being that this like this idea of sovereignty for like developers to build in their own playground their own sandbox so they can make their own money is like it's looking at it from the wrong perspective like we should be optimizing for what the users really are

like after and what the users want is this experience which is composable which is fluid which is easy to operate in and doesn't require kind of the annoyances of like what one could say as like I guess cross chain interoperability as we know it now um important user wants kind of the same level of security right you don't want like you as a user never want to reason about security like that's not something that like you know you sit there it's like so 17 validators is that bad or good like oh what's an aamod efficient is like you know that's not like do they have all the audience like yeah okay and the kind of on the more economic side the idea of running of running your own

kind of uh system whether it be uh consensus and your your own nodes and uh your you know your oracles and your rpcs and your block explorers and your all different things for individual chains it it it basically it's part of your point is that in a world right now where things are very kind of like uh uh FL like pushed up by uh early stage funding that this kind of is just being absorbed by teams and isn't really a consideration but like in time when we're when we need to reach real revenues um and kind of product Market fit the idea of everybody having their own having to pay for all this own infrastructure is kind of problematic from a perspective of why not just do

shared costs across an entire spectrum of apps and users and developers and team okay to to okay so like both of those points so okay so on the second point I I I have this thesis I think a lot of people share it that like the cost of infrastructure in software is basically going to Trend very close to zero in crypto where the cost of all these different like operations um I is kind of akin to like what software is in like on your phone or like a mobile app or like even just like Twitter as an example is free all these different kind of like software applications are free and so why not the infrastructure for

crypto to be kind of uh Trend towards zero uh as well and just as different teams are kind of commoditizing each other um that the cost of running a chain what what is today was much less than what it was yesterday yesterday and will be far less tomorrow and so I think that's kind of like what I would retort with with regards to the cost idea and then with regards to like the composability and and security on the composability side I think what we're addressing right now is like the state of rollup interoperability and like chain abstraction as it is today which again I think is like is is what it is but I think it's it's like we are very

close to like at least getting some sort of stand standard in the ethereum space and I think in the broader crypto Community having these kind of like true chain abstracted ux so it's hard to make that argument now today um but I think that like in the world of user experience perspective while having your own a ton of app chains may never uh equate to just having like one chain the the the experience is going to be far better than what it is today and we'll be solving kind of the first broader issue that you brought up with regards like the single state machine uh not being able to to kind of scale now the security is kind of like another question right like is a centralized sequencer in today's format better or

worse than a L1 with a small validator count I would argue it's probably worse um in the future in with decentralized sequencers is it better I'd probably argue that some sort of decentralized coordinated block building is probably better than an L1 with low validator count but that that's kind of a tough one to get over but it does seem like the points that I'm rooting on like my my my current thesis on are kind of uh pushed out into the a future view of the world rather than what it is today whereas it seems like where where you're coming from with regards to kind of near sharing design um is very much like this is here this is here and now and it's happening so uh perhaps there's a difference of time frames in our views

but also definitely curious to understand uh your perspective on those kind of uh on those two push backs yeah yeah um and I appreciate those so the the cost one indeed like the cost right now is immense right like running a full fully like fully rolled out you know your own chain is like anywhere from probably 20 million a year to 100 million a year right depending on how much Integrations and kind of how much like auxiliary stuff you need to have and like if you have a separate VM or kind of more uh like you you sharing the tooling right and I think the the the interesting thing was like the you know

um like roll up as a service and similar things which are kind of like okay well we just have like one set of tooling that we're going to run on every chain um and so you know you can imagine that like some of this will get unified and it's like you know you still get some economy of scale even though um at least like developers developing the same stuff for all the UPS the thing is like then you're kind of losing all of the sovereignty because like if you try if you build if you modify your evm now everything broke right and so now you need to like Fork everything and hire Engineers to do it so I think that the important part here is is really it's about the this tradeoff between how much it

costs to to do to change things and innovate right because yeah if you're using exactly the same stack you like just clicked Fork then the cost is probably way lower because indeed the cost is now if we're talking about the future we have you know the the second part about AI we'll talk about that how a lot of this things will indeed change but yeah I am talking about kind of now like you know because kind of the question is as you're an application developer you kind of need to decide what to do now and as especially as you're planning to scale what do you do now I think is really again now and like for next year is really important um I agree that like things will change but in a very

different way because of AI and we'll talk about that so so and that's why like the costs are like pretty tremendous even for Layer Two right it's like million a year at least you know in the infrastructure in auxiliary infrastructure I'm not talking about da costs or anything right which you know you can use n labs and get very cheap da uh not to plug a n um but yeah I think the the cost of some of the things will go down but if you want to innovate if you want to change something if you want to have that sovereignty that you like want to pay for then it's going to you know you will need to have that cost added up the kind of the design with Nightshade we bringing in also allows you to have your

own Shard right so you can have like as much compute as as sh uh as any other chain would have uh so you don't and you don't have the same kind of problems of uh you know some some other app got popular and now everybody's struggling with that uh and at the same time you can also flip it and have sharted smart contracts right which are available in every Shard on the user account and have state together with other sharded smart contracts and user information right so you kind of like what we're trying to do is actually flip the script where it's users's information and stored in users's account and so when you're building applications you're actually storing them on users uh in user State and I think that is also

fundamentally kind of moving us in a direction like of that Vision where user own everything and then you just have kind of state synchronization smart contracts which are defining the rules between kind of how we peer-to-peer transact now not everything can be implemented like that there will be still shared smart contracts and they can live in dedicated charts if needed the and so going to interoperability I I mean I again we working on chain obstruction like we will we will make it work work between all the L twos and Chains No and Bitcoin and Dodge and Litecoin like this is this is all coming we have like ton of interesting applications on that but again they it always comes with latency

it always comes with additional um with additional risks as well right and so for example uh you know we're designing one of the kind of deck incentive kind of intent based decks and you know you need to consider the fact that some chains will go down and disappear right the layer twos will shut down you know maybe some layer ones will just stop right for a few hours or or longer um and so how do you like deal with those situations what do you do what if person deposited money on that chain and wants to withdraw in a different chain right so all of those situations need to be handled this is not something you need to deal with when you building kind of

in a single environment um so I think it's indeed like longer term we're going to change things and you can imagine like okay cool I want to start my own like server with my own app like that feels like a right model but because the your users are not there right the users want to contain state within within there that's the whole point we offering with you know user ownership um and the just like the fact of like Integrations being so much more complicated I I just don't think that's a right model and uh again we see like the inov like people who are actually starting new Chains It's very

infrequent that they end up building like truly Innovative products right there's only like few probably that are because they spend so much work on building the chain right you end up just like spending so much uh effort there and building the ecosystem Etc now economically right now you know obviously there's a premium for doing that and so people are rewarded but if you're truly building an app right you should be focusing on product on end user on figuring out and maximizing how much kind of um how much impact you can get to the end user and you know using shared infrastructure is just so much like easier cheaper and you know for the most part the tradeoffs

are there to you know to get you going uh in a successful way yeah I think um I think our ideas like in the sense that you see like the way that like n's model is isn't just like okay let's just max out a a single L1 and just kind of scale as much as we can with with kind of like one state like sharded State sharded user accounts allows you to kind of like basically what you're saying is like near is solving like near has solved what ethereum is trying to do right now in solving its interoperability problem just by charting at the L1 layer instead of having to launch all these new chains

so like the same idea of horizontal scaling is like applied it's just that in the in the kind of modular world in in the ethereum kind of rollup World putting them together has been way more difficult because well they're not touching the same state consistently like they are in near model but you guys have basically achieved what that kind of is the original yeah the original thought process was yeah 2018 each two circus 2018 is near yeah and so somehow along the way we got we got lost there in the ethereum land which is which is is what it is so yeah I mean I think you there's quite a few incentive misalignments that you alluded

to I think one being like the valuation game the investing game of launching your own chain uh as well as just kind of like yeah that idea of we can you know make money off the sequencers and things but again if the fees in crypto Trend towards zero on infra then you you got to think like fees on on rollups fees on l1s l2s are also going to Trend towards zero because that's just a state of kind of using anything again like I would you pay to send out a tweet every time you tweet I don't think so right and so like or maybe you would I don't know if you got enough economic value from that system and so I think like in that sense like these these applications have to create real ways to make money and not kind of rely on on on their kind

of like sequency Revenue um but yeah I'm I'm kind of once again seeing that okay the the vision is the same with regards to how we scale we can't scale single state machines we must kind of scale out in this like sharded in this rollup in this horizontal way um build interoperability first yeah with interoperability first well that's the big problem is that people in ethereum land are like okay guys it's been like what if ethereum was around 2016 so it's been almost nine years and we still haven't really figured this out and so that's kind of like the question here by the way one fundamental difference that near that we've done on near is that we have a single state space so on ethereum every single kind of Roll Up is a copy

of a state space and so you have a on everyone you have Unis swap on every one you have my account on every one of them instead we say hey it's single state space right we we sh we shed with split it into kind of cohorts and like they communicate with each other uh we can split it it differently every like whatever every 24 hours but because of this it means like you have one account right and just communicating with everybody else and so you can still mimic a lot of the same stuff like you can have multiple Unis swops or whatever if you want but like you now don't have this problem that like everybody's trying to copy the same thing everywhere yeah I mean I do think that's the

problem of not having interoperability baked in from the beginning like each ecosystem can have like its native decks its native lending Market but like you know if chain signatures work if if if Zion's formula works if a gor gets orchestration right and all these different interoperability things come online having a a DEX on one chain is gonna be no different than having a DEX on another chain it's gonna just be a DEX so it's like there's no Edge there so like that that's gonna be a whole cycle of kind of of getting rid of the Recycled products that we've seen as well so like you know as an investor there's no really value proposition to investing in the first decks on a chain theoretically because I me know the

chains are just going to be abstracted anyway so it's just going to be a deck you're just GNA end up using the biggest one well I mean you're going to use like overlay pretty much that can route through any yeah whatever underlying pools which are like both offchain and onchain through like that's kind of what intent intent based architecture for dexes allows you to Route liquidity through anything right oh you have better liquidity on binance right now cool we'll just like match you there it's better on whatever Seven Layer Two like in fifths pool cool right will do this so so it kind of unifies the space yeah but it but it removes all the like sequencer rollups and like all the

M Etc because now like that is happening kind of off chain you're running an auction pretty much saying like hey I'm willing to do this who who can fil me with the best price yeah that's what we want like we want we don't want rollups in the ethereum space we don't want like these different chains app chains general purpose to like be like headbutting each other infighting doing all this games with each other like we just want if if you know if $10 enters into crypto we want that $10 to be to be able to to go and be used where it is in the broader space not like fighting over you know that that user or that one like if we get a 100 users into crypto we should be it should be like that should be able to that liquidity should be able

to be distributed amongst whatever chain that those users want apps want to use applications on it shouldn't have to be like I come to a chain and I use that chain and then other chains have to fight for my liquidity with these games and things it should be if if I want to use this app on this other chain I just go there and it's just my liquidity is there as well so um I think the I think chain abstraction I think you've talked about this Ilia has bigger ideological impacts than just a ux impact for the way that people are going to be building in the space I think we're kind of touching on this a bit exactly yeah I mean I think it's it changes your like

instead of I mean instead of having this argument is like oh should I launch a separate chain or not it kind of changes this argument away from because like the the reason to have that argument it right now right I mean we talked about sovereignty and other things but is more around like oh if I launch my own chain I can get a better valuation right that's like that that's that's where like a lot of this right now like oh layer tw's got like premium layer ones were giving premium before right so like let's do this is that why near launched its own chain I mean so the honest thing was this is 2018 right we we looked at a lot of blockchains we obviously you know Bitcoin etherium we originally wanted to

do like payment for crowdsourcing for data right we actually wanted to build I mean like what would be a version like a gitcoin style you know few Generations before what gitcoin was like paying for work for people like um and kind of allocating that but you know we look through all the blockchains none of them actually would work right none of them would scale none of them would be cheap enough now would be easy enough to unboard new users and so we're like cool we we then went and talked to ethereum folks as well and we're like oh you know you guys are thinking about charting we can do charting uh we can build charting like my co-founder build Charter

database you know I worked on a lot of kind of uh scalable stuff as well we had like a bunch of folks we could bring you know what do you guys think should we like build it as part of ethereum how would that work and the answer was there was no like way to make that decision to kind of get external team to build something uh at least at the time and so yeah I mean we ended up kind of uh and we were already we had already VC funding from before from when we were doing AI stuff so we needed some some like VC back model so indeed like the incentives of you know VC back model led to that um if there was a way to just like contribute to ethereum and receive part of ethereum like allocation for example ethereum Supply if for doing

that I would actually prefer to do that so having such a mechanics I think is very important that's the thing about the lack of a kind of like a top like a or a bottomup I think is what they call it like a setup for a theorem rather than having like a a a single like you know kind of way of coordinating um it can be a little dysfunctional now um let's kind of shift gears into the AI Rob would' love for you to get involved here as well um IIA why does none of this conversation matter in a world of AI in the next two years why did we even just spend 30 minutes talking about this welcome aboard on our journey into the singularity this is the endgame road map

for frax and fractal the new L2 out of the frax ecosystem that is totally based on fractal scaling frax a modular L2 blockchain built by fra with modular and fractal scaling in mind fractal scaling is based on the mandal BR set of fractal images and and it explains how l2s become l3s and the entire app chain life cycle goes through its Evolution the frax ecosystem is made of some of the top builders in the space and has stemmed from its OG days of the frax stable coin it is now in the LST space

the LRT space and this fraxel L2 joins the existing Suite of fra products including fra Swap and fra lend you can see FRA faery the bridge to go from any chain arbitrum optimism or main net into the fraxel main net Bridge your tokens and bring your FRA assets into the fractal ecosystem join today to get the best out of fraxel you are still early join today take part in our Expedition into the singularity no so I think to be clear like a lot of the things actually do matter I think some of the accent

will be changing and and and some of those things is actually hard to predict even even as we're working through it it's kind of hard to see what things will happen at what timeline right and again this is like I you know worked on Transformers in 2017 in 2017 we thought is like okay this is happening you know is here and then it took like whatever four years for this to mature to scale and you know we're a couple years in and like we see we see the progress we see the momentum but we're still like you know do we have AGI tomorrow no like will we start ramping up into a lot of interesting things over next couple years yes so um yeah so I think maybe

just on the ey side there's probably two important Concepts so one is uh we've been talking a lot about um is user ownership right so we talk about it on block blockchain side but it's also as important on AI if not more important because right now all the AI models are trained in kind of for benefiting the company corporation that trained them right and all of the you know all those big labs are really big public companies they'll need to you know post more Revenue every year every quarter and so they need to actually you know increase how much revenue they're getting and remember most of them already maximized their target market

right it's not like they can add another billion users to their uh audience so they kind of need to actually get more money out of the same users right so so all the AI models one way or another will be used to actually extract more money from the same users right which are us and so so that's I call it like corporate Ai and so we need a version that's actually more focused on you right and kind again gives back sovereignty back to the user and so this is what we call user own Ai and so user own I Builds on top of user own infrastructure right the user own data user own assets and so you're kind of like linking all those pieces together now um the other concept is what we actually worked at neari back in 2017 is

AI development so the reality is kind of the current interface like as a software engineer you write some code you know you you're learning the machine learning uh you you're learning the machines interface but what kind of this AI models are showing us they can do they can teach machines to talk up in our way right they can teach them semantics of our own world and and kind of language and so what the transition is happening and again I was too early on this in 17 but I also for context I've tried to build my thesis around this back in University that was also too early uh but I've been trying to do this for for yeah many years now but the

reality is like we will be able to tell computers what to do and they will do it right and they will write code to do it for themsel to communicate with each other and so a lot of the coding process will be replaced from needing humans to be like you know anyone can actually um build software now that changes how we interact with Computing there also changes some of the cost structures right if we're talking about like you know how much it cost to maintain infrastructure if you can just like talk to a machine to tell it what to do and rewrite you know half of the software around you well the cost for that will be way lower so that that's an important you know note there that some of the cost and speed with which you can do

that will change but at the same time so those are two really important Concepts that's actually what we're working on at neari is like we we kind of building the infrastructure in stack for user own AI together with a lot of projects in the ecosystem right so we're not trying to do it ourselves we have you know incubator we have a lot of projects we collaborating with they plug it in into the platform we're building and we also offering it to for people to use so and we kind of have a alpha right now that's that's running and then on the other side is like we're using this stack to can make progress on the research of how to build a i developer right then uh you know from what we see now the front end part is getting easier and easier right

we already see like you can build front ends on a fly you can like click on button say this button should like send a transaction on chain boom you know it wrote the code for it now it works right so that part like similarly you know for smart contracts like some outlining logic Etc it's not there yet but there is a lot of uh interesting work there and there we also testing a lot of form modification and this is the other concept so most of the code right now has bugs uh like the software we're using right now has bugs right and some of the bugs are critical vulnerabilities that can be used to do something bad right AI is actually getting better and better at

detecting these variabilities because it's seen all the code and seen varability so it can actually kind of match really well uh I did actually run the test myself and found like vulnerability in some go uh compiler and what will happen is like if this continues and we don't fix it we actually going to have more and more hacks right it happening now in crypto we actually been working on this because we have a situation where a lot of money are at stake and we need to actually have very correct code and so we have been working on formal verification research for this specific reason and so the kind of one of the dimensions of work we're doing is how do use AI to do

form modification and smart contracts is the first use case kind of the because it's like straightforward and valuable and uh and this is but this is the way we believe kind of all the future software will be built is will be built by I with formal proofs and formal proofs have this like nice kind of similar to ZK uh proofs property that you know if you have some requirements and you're using some other code which has a form of proof if it doesn't match if there's something incorrect there it will just not like proof will not be correct right you cannot plug it in so so all of that like pieces you know piecing all this together kind of the vision is really to a have an open

source kind of user own AI EOS system uh it's going to be lots of different projects coming together competing you know providing different properties and you know at the end benefiting the user again sovereignty in the hands of the user uh and then one of those use cases will be an ability for user to build software without writing code and so that's kind of uh research we're working on but what this enables now is completely new experiences right where you know the chain abstraction will be it will be like next level from chain abstraction right now you can abstract user interfaces as well where you can on a fly generate a front end that you know interact with multiple chains with multiple you know apis Etc and gives you

the experience that you want at this moment and so you can imagine you know it even predicting that you want to interact with something and kind of generating that on a fly or you saying like hey actually I want to do this right now boom you have this custom front end so we have few projects that are working on different angles of that right like for defi specifically like hey can we like combine all the front ends can we have chain signatures together boom now you have wi can like navigate that and uh and you know initiate transactions like initiating transactions with initiating uh kind of interactions Etc so all of that is in progress like in different versions so we can like start to see how it will work um as it ties together again it's

kind of one level above even chain abstraction where you know you kind of abstracting the whole Computing with natural language but chain obstruction serves as a big function of that because we do need to move to a more trustless kind of world and that will be the big fundamental piece to power that I just went on like a 10 minute rant I think so yeah I mean it's yeah just real quick couple things um I feel like like some of these companies that are like companies that are Char charging for services that can use AI are gonna have to like kind of shift their business models or like figure out how to incorporate AI to and keep an edge you

mentioned audits um and some some of these other kind of areas so I think that's interesting I also think how you mentioned that like near in particular is set up for the expansion of AI based on how your how your chain is built it'll be very interesting to see how other kind of types of companies accelerate more than others with the use of AI and kind of which adopt it and and kind of instead of shying away really kind of like embrace it uh those are my two first main points but yeah Rob would love to get some esoteric entropy based takes here I I love the idea of user owned AI I think there's issues with the AI programs that we've seen out there now and they're owned by corporations

and they're operating based on the interests of these corporations that are profit seeking and they're they they've shown behavior of praying on their users and and extracting value from them so the idea of user owned AI I think makes a ton of sense and I'm I'm very curious to see how the incentive model will change um when we have these AI agents that are acting on our behalf um maybe they have I I I think I think these AI agents end up having these these uh the value that it can provide uh is going to end up for one okay so zooming

out I think one that there has to be an incentive for some platform because people aren't going to have the the technical knowledge to deploy their own AI so there's going to have to be some GPT model type platform that will allow them to to deploy their own AI agent and then it's a matter of how they want to optimize it and that that'll be the specialized value it provides to other AI agents um and then how those AI agents interact with these closed sorts corporate owned ones I think there there's this alignment problem or potential potential problem where um like for one how do I make sure I think it's pretty obvious that my

AI my interests are not totally aligned with the corporation and so if the if the corporation deploys the my interests aren't align with that agent if it's acting in the best interest of the corporation but what if we have like this this extrapolates to the end user but let's just start with two corporations right that each have their own AI agents and maybe they have contradictory interests what's to say that that one isn't able to convin my AI agent isn't able to convince your AI agent to like do something in my best interest rather than your best interest I I think those natural comp competitions are going to evolve and and how those things come out I think it's going to be really really cool to see

because we're going to end up with these AI agent competitions um and I think we're hearing a lot about uh so I think user owned AI to once I have my user competing against other users um there's going to be a very natural like healthy competition of like me specializing in one thing providing value to your AI agent paying AI agents paying one another to do that and there's going to be a whole economy that emerges as a result of that and I think it's natural for that to happen especially when the twos your own for that to happen on blockchains because it's already a natural place for economies to emerge especially for open- Source uh technology and end incentives well and in general like AI agents have trouble

like doing Finance which is not on blockchain so it's just like so much more straightforward to plug them in into an account and then have them transacting than doing that with stripe or with like any other infrastructure and like you can start them they can be like you know semi-autonomous they can raise their own money potentially right by launching a token and like pay their suppliers like they can do all this work and you know like you can go from this is my agent to like this is a fully autonomous agent that just raised its own money and now you know providing a service as like a you know and monetizing it and redistributing profits to the to the Token holders and it's not a security because it's not controlled by anyone do

you think it's like still aligned with the The Entity that deploys it or do you think need to yeah it can be like completely autonomous um and so like you have you know we have we're building enough infrastructure so that like the you can trigger l l from onchain right you pay for that onchain and so you can have a smart contract that literally does all that and kind of if if it have enough money it will be able to do it if it if it mismanages its money you know it's it's done uh and I mean you won't have some oversight so like token holders can have governance power they can provide feedback they can provide training data for this AI agent uh and I actually pitched this idea as like AI CEO right so you have like AI that you

have a board members or you know token holders who can like vote on you know what is doing well and and not but then it actually goes and executes dayto day you know does like day-to-day work yeah in the extreme case that looks like AI president but in the less extreme case that I think is is somewhat I've heard this idea where these AI these AI accelerationists are trying to basically come up with AI AGI and then if it if it's the AI for the company or the corporation then you just ask it because like you just ask it how do how do we optimize for Revenue right and you you generate the AGI and then you use it to to basically be the CEO of the company

do are you are you are you one of these accelerationists or or de accelerationist how do you philosophically view this verx Edge is now live on say Network say users can now trade 50 plus spot and per markets with the unified order book liquidity of vertex Edge across arbitrum blast mantle and say the next 6 weeks 850,000 say tokens will be up for grabs as Trading awards every single week exclusively on say join vertex onay Network now on vertex Edge that's vertex protocol.com yeah I mean

I'm definitely not a deceleration I think I'm like the the way I see this is how like kind of through this sovereignty lens right just a user and and Community like a I mean when I say user owned it also includes communities of people as well right so because I think it's it's important to balance the individualism and kind of more of a community feel as well and I think it's important that like right now the way people the acceleration is kind of movement is just like let's go right it doesn't matter how let's just like improve the technology and it doesn't matter how it's going to be applied right and so I like that part I don't agree with because I think like if we

mismanage that then we're going to get into like conc contion of power and kind of you know and we know that power corrupts and this is going to be absolute power so it will corrupt absolutely right so we'll have like one or maybe like two entities which are like controlling intelligence and through that are able to reprogram everyone so like I do think it's important to have this technology built in a way that is owned by everyone and everyone can kind of keep everyone else in check right like if you have an agent on your side and and like it's you know cryptographically on your side than if somebody else is trying to misuse this technology which people will like you have as much pretty much capabilities to

protect yourself right from this and and you'll have more people and more more kind of capacity to protect so like we we go back into the blockchain side you want more people on a good side than on a bad side but all have the same kind of level of of capabilities because if we only have like one or couple parties that are able to do that then we kind of screw yeah so what's to stop what's to stop people like I know for example right now in California there's like an AI law being discussed about kind of like control but what's to stop in the similar way of like the Space Race for going to the moon or for kind of getting uh nation states technology in space

same way with crypto where nation states kind of don't want to ban it but don't want to you know some will accumulate some won't what's like is there a is there a way in which we police AI that is fair and uh adhesive or like is is a part of its growth trajectory without kind of like stalling the growth or like how do you view the tragedy of the commons question which I'm sure you get a lot uh when it comes to AI yeah I mean I definitely think just like a pure political regulation is not enough and it's not going be effective right because it it's actually just stles Innovation I mean we we familiar with it

in crypto right because like every time by the time they figure out what to do with like this instrument right we already invented 500 new instruments because it's like oh you want a packaged yield that you know repurchase itself you know three times a day and you know shorts on the back end is like what what is this like like we just invented completely new instrument now somebody needs to go and like analyze if this is like a security a commodity or whatever an options or future no it just doesn't work like that you need you need a new way of kind of modeling it and it needs to be it needs to be product first it needs to be user first right and I think for many of this regulations they're like let's cover our ass and let's like protect the government versus like what

is actually benefits the user and so so I think like with actually Bitcoin what you mentioned right like some accumulate right and some you know maybe like not and kind of trying to to ban it but at least that's an incentive mechanism right like you know some people see it and so they'll benefit because like in the long term they were early on and so they were so I think like we need similar structure where we have incentives uh to do good things and detect when bad things are done and you know leverage the same technology and kind of in a way Leverage The also kind of crypto cryptography and other methodologies to protect it so to give you an example right

the you know uh we have deep fakes right you know anybody can generate a fake video fake uh fake image fake audio now and there's a very kind of straightforward although hard to adopt methodology which is well if you're a company you should sign all the content you publish if you're government you should sign the content you publish right White House actually at some point started signing videos they publish with cryptographic signature because now you know like it doesn't matter if they published a deep fake this is published from their perspective right uh or whatever like El musk published something like it's it's El musk signature it's not a signature of whatever uh whoever is in the video so

like we recording this we should like Auto auto sign at the end the content and now like if anybody wants to look at this you know recording they can see like oh actually Ilia signed this right and this is like kind of here's a Providence Etc so so I think like part of it is also thinking through okay we're going to be in a world where you know like the traditional way of like oh I'm giving you a piece of paper and you can trust me because it has a stamp on it it doesn't work anymore but we've been on this world for a while now it's not new world we just kind of been like lazy about like adopting technology to really deal with that and AI is just like kind of putting it in overdrive but like you know Photoshop fake documents

Etc all of this existed before we just now can do it at like way larger scale and so I think that's important to just understand we need to we need to build systems and kind of our society operating system into this post AI world now and cryptography and crypto economics is the way to do it it's like fundamental building blocks to build this you know kind of software our society I'm not super confident in the government to implement that standard I think you're right like I I polit Pol political regulation is not enough we need cryptographic regulation and that that CER that signature that we could put into any sort of content is an example of

cryptographic Regulation or verification that that can be embedded and then I think there's further verification that can be taken onchain by proving certain actions of AI agents and and I think that's probably the way to stop any kind of like Rogue AI from misaligning with its creator individual corporate or otherwise but that that gets really hay when smart contracts start deploying their own AI agents and we sort of it's sort of two or three steps removed from a human um yeah but then we start getting way too many app chains without the proper interoperability and then we have another podcast about it too many app

chains there's agent chain and then it's all like it's all downhill from there well it's been a pleasure man thank you for coming on and sharing some time I think I think it's a good place to wrap here uh really fun conversation and appreciate your insights and I hope you enjoy the time off thanks guys thanks for hear me cheers thanks man