Episode 82

Sharding Scalability with Bowen Wang

Watch on YouTube

About this episode

In this week's episode of Untangling Web3, we're thrilled to host Bowen Wang, head of protocol at NEAR, a leading figure in blockchain technology.

This episode delves into the cutting-edge topic of blockchain scalability through sharding, a pivotal aspect of NEAR's development strategy aimed at overcoming existing limitations and propelling widespread adoption.

Key highlights

  • Blockchain Scalability Challenges. Bowen elaborates on how scalability is a bottleneck for blockchain adoption, describing how traditional blockchain architectures struggle to handle large-scale usage without compromising speed or incurring high fees.
  • Sharding Explained. The episode explains the concept of sharding, where blockchain divides data across multiple nodes to enhance transaction throughput and efficiency, likened to how vast social media platforms manage vast user data.
  • Nightshade 2.0 Upgrade. Bowen discusses NEAR's latest upgrade, Nightshade 2.0, which employs a sharding mechanism to dramatically increase transaction capacity and reduce latency, aiming for true mass adoption scalability.

Today's discussion with Bowen Wang sheds light on how sharding and other innovative approaches can resolve longstanding blockchain challenges. The insights provided highlight NEAR's commitment to enhancing blockchain usability and accessibility, ensuring the technology's readiness for a global user base.

This leap in blockchain technology marks a significant milestone in fulfilling the promise of decentralized technology for everyday applications.

This episode is sponsored by the VeChain foundation. Learn more about VeBetterDAO here: https://vebetterdao.org/

Transcript

Auto-generated from the episode audio.

welcome to the untangling web3 podcast joining us on today's show is Bowen Wang head of protocol at near one the R&D Hub driving forward the near blockchain and ecosystem we've invited Bowen to explore the challenge of scalability for major blockchains this is something that stands as a banana skin to mass adoption we want to discuss with Bowen some of the exciting developments with n's latest upgrade Nightshade 2.0 in the context of scalability so let's get started Bowen welcome to the show it's great to be here nice to meet you Jack and Alex no thanks for your time B I'm really to speak to a proper professional you know someone who's properly deep into the weeds of this topic of scalability you know your your head of protocol at near one so I can imagine you're uber technical which is great and we'll get into all the weeds of that first and foremost why don't we just start off by you know grounding ourselves in this idea of scalability so how would you contextualize scalability for

blockchains and that challenge because it's something a lot of people talk about it but what does it actually mean to you yeah so it's basically the problem that we need to solve in order to get to mass adoption of of blockchain of crypto uh this is because you know one of the fundamental Promises of blockchain is is to build this um Global Financial system or at least um be a foundation for that and then you can build various things applications on top of that um but this foundational layer needs to be scalable enough uh in order for everyone in the world or U you know of people to use it together um without having to either wait for a very long time or um pay an extraordinary amount of fees uh to use the system it's like a good summary I think what most people think of blockchain they think of crypto when they think of crypto they think of you know digital money or electronic cash and these kind of things and when we

look at you know the classic examples of Bitcoin ethereum and their TPS and he transactions they can support it's like in the single digits and double digits right it's up to 20 30 these kind of things and when you look at like The Current financial systems the Visas of the world they're processing what like 50,000 transactions per second and it's a real kind of fundamental like problem there where we can't actually get to that mass adoption because you know we don't have the capacity so why is this such a problem when we talk about like decentralized Ledger technology why is it such a problem has it been such a problem so far to get this like throughput this this scalability yeah so um in traditional blockchain design how it roughly works is that um there is some node that does the computation right let's say you have some transactions or let's not worry for a second about how you get the transactions and so on but let's just say there's some transactions right so what

what it what a transaction is is basically a kind of like instruction to do some computation if you think about it from a technical point of view and then there's just like whole bunch of these things right and then you you need to draw computation okay that that that's fine but like you know let's say you have um many nodes let's say 1,000 nodes or there's no or just machines right in this network and you you know for for like someone who's like completely um unfamiliar with broad chain they might think okay you have 1,000 machines then they should be able to do a lot more work than what one machine could do right they should be able to process a lot more transactions and what one machine could process but that's not how blockchain Works how it really works is that those one machine all do the exact same computation because they need to reach consensus on what the result of computation is um so it it it doesn't really scale that way and

because of that in this traditional like design it it is limited by what one machine could do and it's not I guess it's even worse right because you're you it's not just you you're doing like one machine doing the computation but you also need to spend quite some time um or like quite some resources reaching communicating with other nodes reaching consensus through some complex um algorithms and do things to prevent malicious attacks and all of that right so it's really worse than what what one machine could do so that's I think um big part of reason why scalability is very hard for broad CH that's a really good way of explaining it and I wonder if you could talk a little bit about the trima problem it's something that we've touched on very briefly but it seems extremely relevant when you're talking about blockchain technology and why it's like so far struggled with the idea of scalability so just the three pillars of scalability security and decentralization and why they're kind of opposed in some

ways yeah so um I think maybe the good way to to think about it still to to come come down to to to some concrete example right so let's say um you know imagine you have a system that is what one machine U right doing doing everything then you can potentially make it more scalable by I don't know throw more Hardware building the speaker machine uh and so on you know you can you can have security measures to to make sure this one machine is really um you know secure and and so on so but but then like it is not so different from any of the centralized system you have today right it's like Visa or whatever that that you you have talked about before um now if if you want to to to make things decentralized you introduce um a lot more machines let's say now you have 1,000 machines now the difficulty becomes like how how do you make sure like those 1,000 machines can actually agree on what the result um

of the computations are and and what the actual state of the blockchain is right so like you know fundamentally what there no what those machines trying to do collectively together is to agree on what the state of the blockchain is and continuously updating the state and and and always in agreement on what the state of the broing is um and that's where the the difficulty comes in right because you by adding those 1,000 notes you're you're not in increasing the throughput or increasing the the performance scalability and you are actually making things worse right you're you're because they they need to find some way to agree with each other and this this process is actually very complex and error prone and that's where the security concern comes in right so you in blockchain there's this thing called BFD assumption which which means that you tolerate up to onethird of like potentially malicious node and this is a concept that doesn't exist in traditional um centralized system right so in centralized system they like you know

machines could be faulty but but you know the worst you can do is to not not be online and and not doing the the computation but here you need to think about okay what if there's some guy who wants to actively sabotage the system right it's not like he's he know this offline but he actually you know is online but he's trying to do something that to to um to gain advantage of the of the system to maybe even like steal like user phones or or like or trying to Min money out of sing air or like things like this right because it's um it is not just a piece of infrastructure um but it also has this monetary primitive build into the system and that is really powerful but also it means that you know you really need to think very hard about how to make it secure and guard against offici PS yeah that's a such a a comprehensive way of explaining I think you know nine times out of 10 people would

struggle to give a better answer than that um and I think it's worth us just noting you know these are problems that are in some way specific to blockchains but there are analogous problems in you know traditional computer science right so if you take something like Facebook or Twitter or x what now it's called they have similar issues because they're operating this massive system they have you know huge number of servers huge amounts of data they have to manage and they are operating distributed systems themselves actually right so they they have to deal with something called the cap theorem you know consistency availability par and tolerance but you know all of these things that you have in the the kind of traditional web 2 world don't they all operate within a single trust boundary right it's still operated by Facebook still operated by X so now in blockchain yeah you have the additional complication of of how do we reach consensus among untrusted parties which is I think why the scalability problem in blockchain has

been so much harder to solve than what we've seen where so looking to solving the problem can you just describe some of at high level some of the the kind of major flavors of scalability solutions that we've seen out there because I know there are lots of competing ones but could you give us a kind of brief overview yeah um I I think maybe I would like summarize it in three categor right one one category is basically two I would call it like s Hardware the problem right so you know you still it's um traditional blockchain design um but the idea is is basically saying that there there's more slow and and you know Hardware is is growing and and and becoming more performance year year by year um you know the the the problem to make the the solution to make blockchains more scalable is to not not only utilize very expensive Hardware or like more expensive Hardware but but actually you know take advantage of or like trying to squeeze everything out of

Hardware um that that means making the computation extremely parallel within within one machine and also um fully utilizing the the network bandwidth of connections and and trying to push as much as possible uh through that um so I think in some sense that's the approach that salana took and what what they believed in you can have more you can you can basically run more and more expensive hardware and um to keep up with the demand and also they did a lot of optimization on the hardware level to to fully take advantage of it's not just you have the expensive Hardware but you actually you know try everything you can to to take advantage of the the CPUs network band and and everything that and B on that point does that in your mind like say reduce the idea of decentralization in a way because you say increasing the requir of someone to participate in the network because all of a sudden instead of running on my Raspberry Pi I have to run it on this

you know incredibly powerful machine that only a few people can purchase um I I don't think it's as Extreme as as people put it in in either direction right it's it's it's neither extremely centralized nor is is very good for c for decentralization so it is definitely not at a state where um you know there only a few people can buy the machine like like I think people under appreciate um the development on on the hardware over the years and how cheap things have become right like this St let's say you know you need um a machine with 128 cores you know on the CPU side and you need like 512 gigabyte of RAM and then you need like one terab of SSD and and something like this right it's or like and then you need like 10 gigabit of network connection um I mean it sounds like pretty crazy right but but I think this thing actually is not super like not extremely expensive to to build right if you if you do it

um on the on the baram metal level then like it I think it cost like I don't know 500 months or or maybe like 700 months um which is um you know I wouldn't I wouldn't call it crazy but but at the same time it it's it's still like not cheap in the sense it's not something that that everyone can can afford and and it still like takes um money and and you know time to set up um and you need to get it on the on the you know param like data centration and and do it uh on the paramal servers instead of doing on the cloud um so it's I think it's still limiting to to some degree but it's not as bad as what some people um make like make it to be it kind of makes sense right like $500 $700 that's not that much money not everyone in the world wants to be a validator on salana or whatever right that's a very like small sect of people but I

cut you off so you talked about the three different buckets and you just covered the first one but throwing Hardware at the problem what are the other two in your mind yeah so other to all like fall into the category of like trying to scale on the kind of software design level right on on the on the design of the system um so one is this homogeneous sharting approach that near takes right so basically the idea is that um well the idea is not foreign right I think Jack already mentioned like you know how Facebook handles like billions of users every every day like they they they don't run on a single machine they have they have a very shorted system that runs on uh a cluster of of many machines the idea is that you you divide the the state of the network into into many shards so that um no single machine actually tracks all The Shard or the entire state of the network and then you like logically each Shard is its

own blockchain so basically instead of having one blockchain now you suddenly have like let's say 100 blockchains um and then your cut is basically 100x in in the ideal case and and and and coming back to the original analogy that talk about um when we started so now instead of like those 1,000 machines all do the same computation they're divided into let's say 10 different groups and all the machines within those groups do the same computation but between different groups they do different computation so you scale that way so you you essentially get the throughput of 10 groups rather than or like 10 machines rather than one machine today's episode is sponsored by the vchain foundation now we often talk about how emerging Technologies like blockchain have the potential to revolutionize sustainability and vchain is at the Forefront of this green movement with initiatives like V better Dow let's take a closer look imagine tracking small actions like recycling at low cost making them valuable assets these verified records build a massive Collective impact benefiting

people through their own actions sound a bit complex but vchain made it fun explore our apps ATV better.com V Chain's goal is to power a new era of Sustainable Solutions so if you've got an app idea or just want to earn rewards for small sustainable actions every day go check out vb.com to find out more and now back to the show okay that makes sense so you know we're here to talk about sharding right because that's what you adopted near so should we dive into it a little bit the idea of breaking up the network into different locations and like having people rout to different parts of the blockchain when they need to I kind of get it on say the Facebook level if I'm based in I don't know say Europe then I would have a Europe server that covers all my friends that messaging in Europe but how does it work on I say a blockchain level where people are transacting across border all over the world like why don't I need to

have the entire history of the chain to actually support that kind of transaction yeah that's a great question um so basically how it works is that um um all the user accounts are in the in the blockchain state right but instead of just just having like one Global State it's actually divided into many sh States so so let's say you know um Alec want to send me some some here so your account may be on sh zero my account may be on sh one um but then there is a way for us to to communicate with each other right you can still send a transaction that goes from your account to my account and actually you know grounding that in in nearest design people don't actually even think about which Shard their account is on is like not not something they worry about it's what the system um handles behind the things and they just um you know send send this transaction and then somehow magically it gets processed even though accounts can be on

different sh and how exactly that happens is quite complex and it's basically all the um thinking and and design that that uh that went in in the past you know five six years that's a a very nice way to ground it in an example I think because you know when you start talking about people's accounts being in different shards that's when my me as a technical person my head starts to explode a little bit because I'm thinking okay what potentially are the problems with this right how do we get consistency between between them and things like that but I think that that makes it a little bit clearer um what what are the kind of technical challenges then to to making this work because I know Shard sharding has been around as an idea for a long time people have you know talked about the potential sharding of ethereum and things for years um so can you can you talk us through like what are the what are the technical tradeoffs and and things you

have to think about here and specifically is there anything that you're kind of sacrificing in order to get that benefit or is this kind of a net gain to to assist ecosystem like like blockchain like Nia yeah um I can definitely talk about some of the challenges or difficulty with shorted blockchain so right so um one I think the first thing is is security right what I think that's a big uh big reason why shorted blockchains are hard right so in the traditional blockchain or in all started blockchain you you know you go is this Assumption of um of up to one3 malicious um validators right and and know people come up with bft consensus algorithms and and so on so forth um and and and all is fine now in this shorted blockchain it becomes more complex right so let's say you still have up to one sort of malicious validators but then let's say you have 10 shards then you can imagine that um there is some mechanism to somehow you know assign

so to speak validator to Shard but you could easily end up with in a situation where A Shard has um a majority of Al being malicious right so let's say um not not just over like one thir but could even over two s and depending on how you do that you know it could be that malicious V can easily take Take Over The Shard and and and then you you kind of screwed right so this is because the reason why you screwed is because you know it's not that if they take over one Shard only that Shard become it's in trouble like you know because the the token or like they could basically just do something malicious on that shard and it actually propagates to the entire system right let's say they Min one billion tokens and then they sign it to accounts on other shards now even other even though other shards were never corrupted they like they you know they they didn't know that um that corrupted Shard was corrupted so they they

just like okay I process this incoming transaction that sends me some tokens okay things move on but but you know those 1 billion tokens now essentially crop the entire system You' got me thinking about loads of different kind of perspectives and points and like we're trying to keep it somewhat high level but you kind of already going into the realm with the more technical stuff it's like to keep it high level is the main benefit here of sharding that you can have this own self-regulated Shard that can achieve its own consensus in a way and be like its own separate blockchain and then say imagine if I separate it based on contents which is obviously doesn't make sense but I will and I have this EUR this European content that's doing all their transactions and if it's you know me sending to Jack and we're both in Europe that's fine because I don't have to worry about things that going on in America for example but then is that is that the main net gain

there that we can kind of self-manage ourselves until the point that I have to send something to an American Shard and then there's just that single transaction that goes between us which is still complicated yeah so I think there are different basically different designs even within sharded blockchain so certain designs um make The Shard more like you you explicitly have this basically they make each sh a blockchain um and um there are advantages and and disadvantage with doing that but but basically like if you if you still make the system kind of one system instead of just having like a 100 separate blockchains then the key difference is whether they share the security or not whe whether they share the entire like whether the same validator set is providing security to it right that's why like you know even though there are like hundreds of Cosmos chains out there they're not like the same system they're just like different blockchains um but here we actually want it to be within the S same system because

um it not only give give give rise to advantages like sh security or like being actually being one system but also things like composability and and that that is building build into the system natively right so so in nearest design it's not that every sh is its own separate blockchain per se um it's more like I'm trying to to kind of work on the analogy that you just talk about it's it's more like people you don't like you just use a system you don't actually worry about which content you're on and somehow mag being like the system figure out like you know um how how you know how it divide people on shards and and and and then figure out all this stuff behind the SC because um the the reason is that each like two shards on near are pretty much like equal to each other in the sense that they don't there's no like real difference other than the um the state that it holds like all the other aspects are are the

same right how how the transactions are executed and and how how the round time works and then so and so forth so um they that that's the homogeneous Shing that that I talk about I mean maybe in the analogy I was thinking I was going to say it's like the internet and you don't know where everything's coming from but you know we have domain names so if you do know what you're doing then you do you do know but maybe it's more like social media right when you're on Facebook or Twitter you have your timeline of post you don't know where they're coming from you don't know the country of origin of each person but it's still coming and you're not thinking about where it is you're not thinking about which partition or Shard of the kind of big set of things so yeah I think this is really interesting that distinction you made you know you're saying that it's sharded but is still one system because when I hear sharding sometimes I think of

it as okay is this a layer two solution you have like uh second layer blockchains but this sounds to me like it's a layer one blockchain so is that maybe if we can get into the specifics of near then and and also about night shakes I know there's a very exciting upgrade to talk about with with 2.0 so yeah how does it work this kind of layer one sharding model for you guys yeah so uh that's basically the the night shade design that uh we have been working on for for many years uh so initially we had this design uh of this sharted blockchain where um this you know uh this the state and transaction processing are all sharded um but the security relies on something called um fraud proofs and you're probably familiar that from like L2 word where it's basically the same concept except here it's much more complex because the consequences that you bear is is is much severe because whenever this thing happen the the state of the entire blockchain Bic

of all Shard simultaneously roll back together um and since 20 like early 2023 we have been working on this new design called stus validation that is something that also was originally proposed by vitalic I think back in 2017 for the ecosystem uh but for various reasons they never they never got to implement that um and we actually we actually realized that it's a really good uh design for shorted blockchain uh even though initially wasn't you know proposed to to be uh adopted in a shorted setting um and it it it helps us with many different factors including the the difficulty of um fraud proofs but also like just improving the performance overall and also adding um uh increasing the decentralization of the network uh because as the Nam so what it really means is that there are like a bunch of validators that don't really hold any state they just like but then they can validate the the blockchain uh validate the state transition without having to to hold any state so this means that

you know because they don't need to store the state locally they can run less expensive machines they don't really need much SSD um and it's really becomes really cheap for people to run um yeah so that's kind of the the the high level idea and then obviously there's like a lot of um complex Machinery that goes in behind the scenes like the to resolve the problems like well how do you communicate between different shards um and how how do you how to you make sure the state transition is valid and how they resolve the security problem that that I talked about before and so so so you you mentioned State a few times that the state transition without knowing what the state actually is is this can you explain that in a bit more detail like again trying to keep it somewhat high level I know these are deeply technical things but is it like a zero knowledge proof kind of style thing how does that work and what are the implications of that yeah

so zero knowledge proof could be used for this and then that's kind of the holy gril of proving State transition but unfortunately we're not there yet because the overhead of um producing xologic proof is really high this St um so basically what that means is that in order to produce a proof um the time it takes to produce a proof is something around like a few thousand Acts or even 10,000 acts compared to the native execution so so basically what that means is that if you if it takes you one second to run the program then it potentially take you like 5,000 seconds to generate the proof so the the overhead is really high and that's that's why it's not practical today um but okay that's a digression but coming back yeah basically it still is kind of a proof but it's based on um this this merco tree structure right so what what it what it does is that the block producer or the chunk producer which is a Shard block um the producer

not only produce a block but it it generates a a thing that we call State Witness what it means is that um it's basically a proof that the state associated with this block indeed belongs to the the state that everyone agrees on before right so so the how it works is that in addition to just providing you uh with the block which includes like transactions and receipt it also gives you the state that's needed to to execute that block so basically what that means is that it it includes all the pieces of state that's touch during the execution um but obviously you know we need to prevent people just um from providing random state right and then like to to that would lead to inval State transition so there is a proof that uh each piece of the state provided in this witness indeed belongs to the correct State um and this is a merco proof and then on the on the receiver side they can verify okay indeed this um this is this proof

is valid and the stating deed belongs to the state that everyone agrees s and then they can Val do that by executing against the St witness so that's kind of yeah how it works okay I don't want to say got it I want to say I think I've got it but it's because it's very complex but it's it's kind of nice to hear actually that obviously ZK stuff is is kind of the Holy Grail as you mentioned but it is nice to see that these problems can be solved via the let's say more traditional memes I love I love whenever a Merkel tree comes up I think they're super interesting facet of of crypto in general and obviously Bitcoin um so just to kind of summarize we're saying that we can do this kind of stateless validation because we are having this proof that show that that kind of is related to the subset of the state that is needed for the state transition right and and the Merkle proof is kind of an efficient

way to convey that state I think that's roughly what we're saying which is good I think you're nodding so I haven't completely misconstrued it that's and yeah I think that's that's that's awesome I think just to take that away from it because I that was something I didn't know and I'll definitely read up a bit more on this um I'm I'm C ious so how does this change this upgrade Nightshade 2.0 introducing stateless validation what does this change for the near blockchain you you've mentioned you know what what is well let's say what is the state transition for near blockchain but you've mentioned you know technically this is changing slightly how the shards can be validated and communicated but is there any other impact on the ecosystem more generally for users or is this all kind of just going to be if you're very very deep into the kind of protocol level stuff I'm interested to know oh so yeah basically this completes our original sharding design right so basically it makes the system fully

shed both state and transaction processing and no validating node actually needs to track more than one shard in the system um and yeah basically still stus validation is how we achieve that and and before we launched this we were in a state where basically this the state is shed but um but still the block producer needs to track All The Shard so it's not a fully sharted system and then and after the N 2.0 launch it's everything is fully shorted and essentially it completed our um kind of original vision for for the shorting even those through different meanss um and this is I think a huge milestone and and the um kind of attestment to the to to the AR design that you know this thing can actually work um and what what this means for us in the future is that we can now fully Leverage The Power of charting getting to to more shards and also improving single sh performance to we now have the the foundation in place to to build this

truly scalable system that can serve billions of users well that's exciting have you like have you started to Benchmark roughly what you think this will mean in terms of like usability in terms of like throughput is there like numbers to put to this and how much you think this will actually improve this you mentioned billions of users there which obviously a big number yeah yeah so I think I think it's actually there's two sides to the story right so surut uh is definitely one of them and and there we actually have a goal of getting to 1 million transaction per second next year um but it's also that to serve 1 billion users what what that means is that you need to be able to handle the state growth um meaning that you know it's not just about like you having to process a huge number of transaction per second but you also need to store the state of all those experience of users and that's I think a problem that's um often overlooked by

people when they talk about scalability but but it's actually really important because if if some point that your your your node is actually struggling to um store the the state then it well first thing will happen is that it will affect throughput because you um you know you for because you have trouble storing the the state so so you resort to means of like that that's are slower and and then that would affect the throughput and then like it also means that um you probably need more Hardware to just keep it up and and it becomes also more expensive uh and eventually this you know leads to it not being sustained okay that's exciting the idea of 1 million TPS by next year is incredibly exciting um I mean that's going to hopefully be somewhere we can actually start to replace you know the current Financial systems and when people complain that it's not there yet like it seems like you're actually going to be there which is really exciting and I'm sure anyone that's

like deep in the weeds of crypto and has tried to download like the entire State history and it's taken like three days or something like that know some of the problems that involved in this that's really exciting um I guess outside of just this night sh 2.0 upgrade what does the road map for near look like what are some of the other upgrades that are on your road map for say the next five years you mentioned maybe the idea of zkp and that being the Holy Grail is there other things that you are looking forward to or excited to build yeah so yeah integrating Zer LGE proof into the protocol is definitely something that we we plan to do once um thear proof becomes more mature in in the sense that um the proof generation overhead become much smaller um and and and then I think another uh another thing that we we wanted to do is to basically provide even better experience for users right even lower latency and so on so today uh

near blocks are pretty fast rate blocks produce roughly once once every second and then get finalized every uh two to three seconds um that's already pretty good but I think uh to truly provide um a seamless experience for web two users or for users who are not familiar with with crypto today we want something that's even faster right so we're thinking about reducing block time to like 200 300 milliseconds and um finality was in one second so that you know it become really a seamless experience for people building various types of applications because today I think still like I think depends on the type of applications you build I think for some maybe it's okay if people wait for a while when they need to confirm transaction but I think for some that people really wanted to um to be confirmed really fast and and and that's how you can build um new types of application that Li the power of blockchain and Leverage The Power of scalability awesome I mean that's such an exciting

Outlook and I'm I'm really excited to see you know you guys go and reach that that million TPS goal that's super exciting um you mentioned applications there and I just want to finish on this Bowen you're obviously super deep in the weed right head of protocol what is it that's getting you up in the morning right in terms of the application of web 3 crypto whatever you want to call it in general what is it that you're excited to see actually happen you know put in users hands in the future what is that for you yeah so we have this vision of building this user owned internet and we are also very heavily invested in Ai and and I think this comes from this Vision that in the future the uh the interface of uh user interaction with the internet uh will not be through browsers but will be through um AI agents or AI systems um and I think we already see this trend uh happening with people more and more going to chbt

or complexity when they need to search something uh versus going to Google and so on um and we envision this word where um you you do like everything including like all the online shopping all the all the trading that you do through like this AI assistant um and then we're that's why we're building this uh intense layer that um allow this settlement on chain this basically open permission system where it settles on chain but um you have the solers that give you codes of chain right so um we imagine this like cohesive uh experience in the future where let's say you want to buy a laptop then you go to you just ask your assistant like you know can can you give me some recommendations this is what I want and after they give you some recommendations you choose one of them and then this like solvers just plug into the system and just give you codes on like hey um do you want this is the best price I can give you and then

you choose one of them and then it it settles on chain and and and delivers to you some days later um so that's kind of the the future that that we envision um and I'm I'm super excited about this future of user owned AI user owned internet um that really transformed the the the day-to-day experience of um of every every single person out there yeah that's a really really interesting topic I think AI agents and yeah I think you you're hitting the nail on the head right that's something that you know people have started talking about a lot recently in the last kind of six months in particular but that missing piece of what are the payment rails going to be for these agents to go and perform certain tasks right to go and execute things with money with with transactions so I think that's a that's a really good Vision to be to be striving for listen b it's been such a pleasure having you on I think we've covered a lot of ground

and I can't believe we've managed to do it hopefully without losing the entire audience because it's such a deep technical topic but I think you know I've got a much better understanding just from having talked to you about this so yeah thank you so much Bowen for joining us I hope you'll uh you'll join us again on the show in the future maybe we'll we'll get you back when you hit that 1 million TPS and we can talk about how you got there um but yeah all that remains to say is to our audience listening thank you for listening and you can join us next time to entangle a little more web three

Topics: Blockchain & Crypto

Related episodes

All episodes