The Verified Web, Agent NFTs, ENS and Ethereum, and and the 50x Rule

Kris Alright. Well, welcome, Prem. You're officially the first guest to the podcast that I've decided to start. I'm not sure why. But okay, for the introduction, I kind of sat down and I was like, okay, I've been I I've known you for a long time. And for the introduction, I was like, okay, let me properly introduce human. Let's say, like, let me just write down everything I know about Prem. And I just started writing and writing and writing. And I realize I know a lot about you. So for those who don't

Prem Makeig Yeah.

Kris know, my first guest is Prem, and he's the founder and CEO of Unruggable. Unruggable is a research development firm that focuses on RD around Ethereum and ENS. his their mission is to make ENS the identity layer of what Prem calls verified web, which I'm gonna ask him later to explain. Prem is also with his co-founder Thomas. the official ENS DAO service provider, and he has been for three years. and you've been contributing to the NS DAO for even more. We can talk about that as well. in the last few years you've written a lot of ENS IPs, ERCs and different standards that a lot of people use today. I'm gonna ask you about that. I'm curious to know whether that's all just you going hardcore on the development or is that your setup of AI agents and we can talk about your entire setup and everything that you're using. I'm really excited to learn about that. I know we spoke about it a million times, but I want to dive deeper into that. yeah, you've written a lot of standards. Some that I know is the on-chain metadata for token registries, binding agents to NFTs that you and I have been mega bullish on lately. there's interoperable interoperable addresses and interoperable names. the NSIP 2526 that kind of officially introduced the you know agent identity tied to ENS. and yeah speaking of agent identity you're you're also responsible for the ENS support and eight zero zero four which is widely adopted standard for agent identity in Ethereum. and yeah kind of the coolest thing lately that you've been focused on is the NFTs and AI agents and fusing them together. so I wanna kinda like dedicate a segment for that. And yeah that kind of work has gone far enough that OpenSea had to kind of invent a separate category just for the thing that you build, which is pretty cool. And it's called agents, along with the like tradable NFTs and art, I think. Now we have agents as well. so yeah, you describe yourself as AI FOMOist and you run AI coding

Prem Makeig Yeah, exactly.

Kris agents, a lot of them. I know you used to run them from a single terminal, but I know you recently switched to just like one of more chat-based kind of programming and and development while you're just having fun, walking or whatever. a lot of stuff that you've built, we can touch on pretty much everything. And you're my co-host as well on ENSN AI, which is something that we started for fun. ENSNAI working group that that has weekly calls. And the most important part, man, I kind of consider you a friend already. We've known each other for like three, four years now. You're coming to my wedding so officially like

Prem Makeig yeah. I hope.

Kris A very long introduction of everything I know about Prem and please everyone welcome Prem. Let's imagine there's audience now. Welcome to the show, man. Super nice to have you.

Prem Makeig Yes. Thank you.

Kris How are you today, sir?

Prem Makeig well it's beautiful weather here. I live in the northeast of America in the United States and we got great fall weather and we're starting to see some trees with the fall colors and stuff, so it's it's pretty nice.

Kris Yes, me too. I was in Malta just few weeks ago and it's just unbearably humid and warm. And I was like, I'm done with this, just like get me home. And the moment

Prem Makeig Ha ha ha.

Kris I got home it was like below like ten, fifteen degrees Celsius. So it was like really cold. I think we're like as as people we kind of just complain about that like when the summertime comes I was like yeah summer and then like after a few weeks you're like fuck I can't get like this summer is so annoying, I'm hot all the time. And then I just like I can't wish for the winter, and then winter comes, we complain about how it's always cold. But anyway,

Prem Makeig Yeah.

Kris so around and round we go. but good to have you, man. I don't have anything pre-prepared, as I said. I don't have any questions. one thing I'm kind of curious to dive deep into like immediately. and this is from your Ethereum sorry talk, I think, where you were talking about verified web instead of web3. So can we unpack that? what do you mean by verified web versus What people call today web tree.

Prem Makeig Okay, yeah. So I did give a talk on that. I I think that was maybe almost a couple of years ago now. And the term I think actually should be mostly attributed

Kris Just to stop you, like that was last year. Can you believe that? So this is a proof that things move quickly in crypto.

Prem Makeig It's maybe you're right. Right, right, right. so I think the term should probably mostly be attributed to Sergei of Chainlink. And he calls the he talks about the verified web. And in my view, there's kind of a distinction between Web3 and the verified web. If you actually go to the definition of web three, which can only really be attributed to Gavin Wood, who kind of coined that term, he basically said that Web3 is an internet. kind of empowered by crypto or kind of fused with crypto and blo in blockchain and where it's kind of enhanced I would say enhanced by blockchain and the verified web is is more a concept that every interaction we have on the web is gonna have a cryptographic proof. So there aren't gonna be

Kris Hm.

Prem Makeig any unverified transactions or interactions. Everything will be verified. and I think we're seeing a bunch of reasons why that might be necessary. right now you see a lot of confusion around who can pay for things. we have now payment agents, used Muse by Facebook just came out last week. And one of the key features is that you can actually buy things using your agent. but there is a kind of confusion right now around the whole internet like Do we want these payments or for instance, I think Amazon is blocking Muse right now? But but Shopify is is like embracing them and creating a deal with them. So you know we have these new entities which are agents and right now people are trying to figure out like well if it's your agent can we kind of like use your identity and kind of delegate it to your agent as as if it's kind of your yeah an extension of your identity. And

Kris Yeah.

Prem Makeig in that case you could attribute us a purchase back to a a buyer, a human buyer. but of course there's a lot of complexity there in how we compose all those proofs together so that we have the idea of like, okay, I can prove this is my agent and its its actions are on behalf of me. but at the same time there's lots of ways to imagine that that could be hard to do or or hard to verify. And so the idea of the verified web is basically every interaction having a verified cryptographic

Kris So w what would you say are the potential upsides and downsides, like advantages and disadvantages of that? Because I've been like I you know I love the identity aspect of blockchain and crypto in general. And I'm kind of constantly scared with type of laws that get discussed, especially in the European Union around with chat control and all of that stuff, I think kind of each country has their own kind of individual like like let's call them issues with with With these types of laws. So like I'm I'm curious, like, where does the benefit stop being a benefit? And where does the issue like officially become an issue? Like I know there's like technically we can do everything. Like every interaction could potentially have cryptographic proof, and upon request from someone, you can show whether that interaction happened or didn't happen, or whether you own some money or whatever your age is, whatever, like with ZK proofs and and all that stuff that we've we're familiar well to well. But at at what point like what happens if it starts being misused and like we kind of fight for every interaction to be, you know, there's a record for it on chain immutable. And like what happens, whoever controls that, how that can be misused. You know what I'm trying to ask? Like I'm kinda going all over the place, but

Prem Makeig Absolutely.

Kris I'm, you know, I'm kind of Speaking my mind here.

Prem Makeig No, I just I understand the what you're what you're trying to describe.

Kris Right.

Prem Makeig I mean the issue is that basically cryptography has is like you know kind of a superpower, right? you can for instance you can send a message from point A to point B without the even people catch the message midway, you know, they can't understand what the message is. And then of course with with public private key encryption you have this magical thing where you don't even have to share a key. You can you can s you can encrypt and from with my public key or I can encrypt with your public key and I can send a message to you that can't. So this is like a superpower. You can imagine this as kind of a a zero to one, right? A technology that went from zero, we didn't have this technology, now we have this technology. So all of a sudden this zero to one technology shift introduces the idea of well how can we use it? And of course, everybody knows that cryptography can be used, encrypting of things can be used for good things and for bad things. so then you know, there was the big fight in the 90s with Mark Andresen and Netscape around cryptography and saying, like, is this a basically a national security issue that people can encode their communications, they can encrypt their communications. And you know, there was export controls on Netscape, and there was a bunch of issues in there, and then of the side punks we're we're fighting to make it open and and allow everybody to use it. So I think we're in the same situation where the every new technology that's developed, particularly in cryptography, just has it's always a double-edged sword. There's ne we're never going to get away from that. There's just no way to be like, all right, well this you know for ZK, what you can accompl we can do with ZK with privacy and and you know all the things you can do with that. You can't say this technology is only going to be used for good purposes.

Kris What what are we even building towards if is if nothing's gonna change?

Prem Makeig Okay, right. So the the thing is that for me it's

Kris I'm sorry, but like I'm just maybe asking like weird and difficult questions. But you just you know, if the human nature is not gonna change and the technology cannot prevent human nature, then kind of like I maybe if like even if it's an incremental change of one percent, sure. Like on a large scale enough, if the whole world implements this, if it's one percent better, good, let's have it. But I'm curious about your take here.

Prem Makeig Well, I mean f I I try to make it fairly simple because for me that that makes it easy. And what for me what's simple is you have problems and you need solutions. So it we can all think of problems that are just affect us negatively and we go like, that's a problem. And then we're like,

Kris Mm.

Prem Makeig what's the solution? So recently a good example, for instance, is this problem of proving that maybe you have a financial status. So for instance, if you want to get an apartment, particularly like somewhere like New York City, you'll go to New York City, they'll say, you can't just rent the apartment. I mean, a lot of places in the world you just have a somebody who has an apartment, they rent the apartment, they talk to you, whatever. You you rent the apartment. In New York City they have the broker system. So basically you you work with a broker, the broker's gonna assess your financial well health and decide whether they want to rent to the apartment and then they're gonna charge a fee and and the and if the fee is expensive it's usually like one and a half months rent you're paying to this third party this broker they're doing a

Kris Wait, that's that's mandatory or optional?

Prem Makeig It's not mandatory but all the apartments use them. So y there's th

Kris Okay.

Prem Makeig there's not a really a lot of supply of apartments that you can get without working with a broker. So basically you work with the broker and the thing that that they do is they charge a lot big fee, like a month and a half's rent, so it's quite large. And they also do a full credit check, including all of your records. Like literally all your bank statements, whatever you want to show them to show your your good tenant, could be a good tenant. And

Kris Yeah.

Prem Makeig you know, so this system is obviously not good. It's like not a good system. It's high cost, it has has this very invasive process of looking at your entire financial record, looking at your credit scores and and and you all your

Kris Just rent an apartment.

Prem Makeig to just rent an apartment, right? So the it's it's this weird thing. You you when you do it, you're like, this is so weird, I can't even believe this exists, you know. And then, you know, and I I have a friend who's a broker and he says, yeah, I know about all this financial weird stuff about people and you know, all their money, and I I so weird. He just like has this intimate knowledge, almost like a doctor, of these people, and all he does is is rent them apartments. And so there is a solution, of course, and now we have maybe there wasn't before, but now we have a solution which is the ZK proofs. So ZK proofs allow this kind of technology this kind of cryptography where you could provably show that you have assets in a certain range, or you could provably show that you have you know a credit score in a range. you could just basically customize exactly what information you would like to share with us with a third party such that they can be satisfied with that information, but actually not share the details, not have an invasive process. So in that way, cryptography. can solve problems and I think we're just in the very early early days of all the things that cryptography can do and how we can use it.

Kris Yeah. Do you know this reminded me of a story when like when I was just learning about blockchains in it was two thousand and eighteen or nineteen. and I was reading about them and I like, sure, okay, I kinda understand it, but I it's really difficult to foresee how things are gonna look like and what the technology is gonna be used in the next five to ten years, right? Let alone in the next like fifty. But someone gave me an example once about blockchains and their power. about the fact that obviously they are there are no no borders, everything is open and public. And I was like, sure, but whatever. And then they someone gave me gave me an example of like, well let's say that all public healthcare information is stored on the blockchain. And let's say that we as a world agree that each hospital in the entire world and every government is gonna be mandatory to implement the blockchain system to read your medical records. And let's say that you go to wherever. You go to Thailand, you go to Kenya, you go here, you go there. Something happens. Like doctors there would be so much better in treating you if they had your medical history. Maybe what you're allergic to, maybe what you respond well, maybe if if you have some some like history of some illnesses or some diseases or whatever else. And I was like, okay, that makes sense, right? Because

Prem Makeig Okay.

Kris We're never gonna we're never gonna agree. Like obviously my medical records are forever gonna, you know, be stored in my whatever local hospital, whatever I'm kind of registered or where where wherever I'm born or whatever. Like especially for s from Serbia. I'm not sure what's like in the US. Maybe you guys have the w like a better functioning healthcare. But for us is I don't know. So that that was like the f I one of the first moments when I was like, Okay, that's interesting. So this kind of reminded me a little bit of that, I'm not sure why. but yeah, I see your point.

Prem Makeig Well, no, that that's a exa very good example because you can look at the blockchain for its cryptographic properties. Like what are the properties of the blockchain and how do they serve the like I said, the problem and the solution. So in in the case of medical records, you know, it's very, very clear that that that selective disclosure is what you want. You know. You want to be able to have someone ask a question about your medical records and then, you know, the the information sh should be disclosed in the least should have the least amount of information disclosed. Like that is less information than more. I mean it should like, okay, are you are you allergic to any of these medications? And it could say like yes or no, but it's not like just like, here's all your in his personal information just because you need this one piece of information. And and of course that's

Kris It's also ki

Prem Makeig w

Kris I think it's more of like sorry to catch you like I think it's more of who gets to ask that question. Like I would be okay with revealing that information to someone. If you ask me, sure. But like would I offer that information publicly? That I'm allergic to something like I'm not

Prem Makeig Right.

Kris sure. Like maybe, you know, I'm I I'd be fine, but like what if someone I don't know like someone could potentially like kill you?

Prem Makeig Well I no, I think that yes, that's true, but I think the the the more common idea was that an employer might not want might might want to hire you.

Kris that sucks. Yeah, yeah, yeah.

Prem Makeig Right, i if if if they see that you have some condition that might threaten your ability to work or whatever, let's say you're t depression. You know, you have you have

Kris Yeah, yeah, yeah.

Prem Makeig you have a condition like that, you've been treated for depression, they're like w I

Kris yeah.

Prem Makeig mean i if I if I was about to hire somebody and I knew that they they had had chronic depression, like I might not want to hire them 'cause I'm just like, Well, I don't want to deal with that. So you know, so there are a lot of areas where you can see it's very common, it's not like it's like a rare thing, like, someone might kill you. It's like more like a very common thing where like an employer might want to profile you and use your medical information.

Kris I went straight to the murder.

Prem Makeig Exactly. Right. But the more common thing is

Kris I always go to the Yeah.

Prem Makeig the more extreme, yeah. So I guess my point is is that I actually wrote an ex post very recently in the last week about this idea that blockchains are what you could call not so much money Legos, but they're cryptography Legos. And what they allow you to do is create you kind of be pro programmatic about the way you create protocols and then have them be verified by the blockchain. So I think that's really the most powerful thing about about blockchains, particularly you know EVM, Ethereum like and other smart contract based

Kris Yeah.

Prem Makeig blockchains, is that they're actually verification systems that are programmatic. You can program any kind of protocol on top of it and you can still verify it. Whereas in the past every crypto cryptography based protocol is very specific And fairly inflexible, like for instance HTTP HTTPS, you know, something like that, where or TLS. It's like this is the way it works. If you ever want to change it, you're gonna have to like go through years of committees and years of process to try to change anything. Whereas with Ethereum and and blockchain protocols, you can literally spin a new protocol up every week and see if it's any good and see if anyone likes it, and it still gets verified uniformly.

Kris One more open ended question about this topic because I find it fascinating, but I also want to be mindful of the time how much we spend time on this. what what will it take? What do you think? What will it take until we all finally go and say, like, okay, let's adopt this technology? Or maybe even a better question, like, what are the what are all the obstacles along the way that we have to figure out, you know?

Prem Makeig Right. I mean I think my in my view,

Kris For this technology to be mature enough, maybe.

Prem Makeig watching crypto, it's kind of slightly comic to me because basically, you just have these constant cycles of people saying like, my god, this is amazing and and this is gonna change the world and like, this is garbage, it doesn't work, it's completely useless. And at the same time, you know, in my view it's very simple. I just look at the growth of the token. So the question is what is the killer app? What is the actual like what is the important Blockchains, to me it's tokens. And you could you you could also say the the the native currencies, but tokens because it shows the activity on the chain. It shows

Kris Yeah.

Prem Makeig so people if you look at the growth of for instance stable coins, they're just up and to the right. They just go and go and go. if you look at any kind of tokens, all kinds of innovation around tokens, and they're just all growing more you get more and more tokens every year. So I think tokenization is just going to continue to grow. I think it's growing just as fast as any of the previous. previous technologies, the internet, I mean AI is super fast, but all these technologies are growing in step with each other.

Kris Yeah, i e everything DeFi is growing, right? But what about for example we're we come from the ENS world, right? So we come more from the maybe agent slash identity side of things. Identity first I'd say and then AI agents and everything else. So

Prem Makeig Well, I would say non financial non financial applications. These this is a whole category

Kris Yeah.

Prem Makeig of applications on the blockchain, which you know gets a lot less attention. But to me, i because of this idea that a blockchain is a universal verification system, for cryptography, non financial applications are are just as appropriate for the blockchain.

Kris Absolutely, a hundred percent. That said, can you like a g one more like an open ended not even a question but like can you d describe the state of identity in general right now, in relation to blockchain? Can can actually can we talk about like why is identity better on blockchain? And then maybe anything else. Take it whatever you want. I keep asking more and more questions, you you start talking I like

Prem Makeig Sure. Well I mean, no problem. I'm happy to talk about any of these questions. basically identity is becoming more like interesting I think for more people right now because of agents. Right. I think before we had the identity systems that we had. We have like national identity systems, we have you know online identity systems like your your handles for your your Google and your X account and we have all these you know various identity systems. And I think for most users they think that the identity systems they have are are pretty much fine. You know, maybe they're not amazing, but th but sort of like, okay, well, you know, this is working. We can just move forward with this. When it comes to AI agents, I think everybody agrees that like one, you can't just put the identity systems that we have on AI agents, because for instance, you know, it's like if you let every AI agent have an X account and just be like, All right, we're just gonna let it run rampant on X, I think at some point they're gonna start realizing that that's gonna be impossible. There's gonna be like just swarms of agents and humans together and it's it's not really gonna work. So So there just isn't any way to bolt human identity onto agents. So we have this opportunity to build it from scratch, build agent identity from scratch. And I think once you're doing that, you're starting to go like, well, what is the best possible identity system? Like what we don't want to just have another system that just happened to be created, you know, that that reproduces what just happened to was happened to be created in the past. And so I think the re the answer for cryptography is that identity has a lot to do with verifiability. So for instance, if you have a a a review system, for instance on Amazon, you have a product with its reviews, you really don't want the creator the or the owner of that product to be able to just post any kind of reviews that they want. You don't want them to be able to say like, okay, I'll just spin up you know a hundred fake users and then just make all the fake reviews. Because basically then it becomes a useless system. So then all the the sites are looking at these review systems and they're thinking, well how do we kind of verify or s or try to guarantee That these are authentic reviews coming from authentic purchasers. So then you know, customers. So then you have the whole cat and mouse game between the bots and the and the the verification systems, and I think when it comes down to it, cryptography is probably gonna be your best way to be able to eventually Have like a human identity, have an agent identity, have agents be able to review humans, have humans to be able to review agents, you know, all that type thing.

Kris Okay, yeah, I wanted to ask you so what makes an identity? What is identity made of, in your opinion? we all think about at least when I say identity, I think of a name, obviously, but I'm biased. So name, you have to have a reputation. meaning like how much weight does your vote carry? And that reputation can mean different things in different contexts based on whatever, you know, whatever assets or whatever how ho however your wallet looks like. and it can be an like a verification or attestation system. Is there anything else or would you say that these three components are the core of what makes an identity an identity? Or how do you look at it?

Prem Makeig Well, I think you could say name or you could say identifier. You know, you do need something that's stable that we can all agree on that identifies the identity. you know, the the the way that I define it is mostly, you know, technically. I mostly think of it as a technical problem. So I think that after working on ENS for for quite a few years, I think I've kind of boiled down some of those ideas into something that that seems to work pretty well, which is the idea is you need to have this identifier. ENS name is is actually you know recorded on the blockchain as a 32-byte hash. So let's call that the identifier. you know it's a 32-byte hash.

Kris Yep.

Prem Makeig And then we have the idea of metadata. So with metadata Mostly I think the metadata can be simple. It can be just the text key for the name of the metadata entry, and then you have the value. And in my my mind that value starts as bytes, and you could call it a string or you can call whatever you want, but it's basically a key value store between a key and bytes. So it's like basically that's identity. That was stored on the blockchain. Because you can basically basically build anything from that primitive. of course you do need the idea of authorization, you know, who can update which records. you know, you need to have established keys to say, you know, which keys are we gonna are we gonna use for our identity, you know, whether that's i the the the name, the you know, the wallet addresses of that id of that user or or identity. context data, you know, anything that you want to attach to this identity, I would say that that's the base idea of what an identity is for me on the blockchain. But then when you have the idea of you know so much that can go around that.

Kris I think what you attach to it is essentially a reputation and how you verify it is an attestation. Would you call that assessment correct?

Prem Makeig well that's not that's not the vocabulary I would use. I would say that attestations make the reputation. Like in a most basic way that you can think of an attestation, it

Kris Yeah, that that totally makes sense.

Prem Makeig the attestation in the most basic way you could say is a like. Everyone knows what that is. You know, a thumbs up, a star. That's an attestation. And why is that an attestation? Because it says identity A, you know, one data point, one bit. says, you know, zero or one, it's like a star, it's like a one or a zero on identity B. So that's why it's the simplest attestation you can think of is a like or a a star.

Kris Yeah yeah yeah. But that's performing the action of liking something. But like my like and your like shouldn't carry the same weight. If we're liking something, let's say I don't know, let's say that we want to give a star to some GitHub repo, and I'm a vibecoder, but you are I don't know, like Ethereum contributor, Linux like contributor, whatever, right? You're this mega star, right? So why is my star equal to your star in terms of support for some

Prem Makeig Well, yeah, right.

Kris repo, right?

Prem Makeig Right. And and I and I think that that actually does happen as you know the reputation does work like that. So for instance, if you look at followers of somebody and then you see in on X and you see that like a lot of people that you trust or you feel that there's some tr you trust in, are liking or or you know, say following somebody, you're gonna have a lot more trust in that other and that's basically broadly called web of trust. So web of trust i is a important I'd say concept within identity for sure.

Kris Yeah. Whose approach to like figuring out identity you like the most? There are a few that come to mind. Like we we were both on a call with with what's the go that it like ANS agent name service. There's an agent name service. we both met Billy in its CC a few years ago, who runs Intuition and Trust protocol. there's them, there's Whoever else, like who like is there someone maybe that you know that you absolutely love their approach to identity and solving identity? Because I I'm kinda confused if like everyone has their own version of like what identity is. You would have to essentially be plugged into many different systems in order to get the benefits of like a reputation or like this or that. So it feels like there should be one, but so far, I'm not sure who who solved it best.

Prem Makeig Well, For me, I I'm still mostly interested in the Ethereum ecosystem. And I do think that you know, Vitalik did talk about identity or at least the idea of naming and the idea of using Ethereum for non financial applications, even in the original white paper. And so I I do th you know find that that his vision for a programmable base layer that provides this verification and can be used in in any number of ways, is the thing that's the most interesting to me. It's it's the idea that hey, we have this base layer, now we have the creativity to be able to build whatever we want, and we can kind of have the let the the best identity system win on top of this base layer and and continue to strengthen base layer and build these protocols versus the ANS system where it's saying like okay we have these institutions which are, for instance, the internet institutions that that we can then use their processes, develop some sort of very static and you know, may maybe there's some consensus around it, but it's like very static. It's gonna be this thing, it's gonna take years to build, it's gonna go through the the internet institutions, it's gonna come out the other side and be a standard. And at that point the question is, well, is it the best standard? And I think the answer in my mind is that process is not the best way to build. these new protocols. And the best way to build the new protocols is to have them kind of compete in the open market of, say, the Ethereum blockchain.

Kris Yeah, we should say like how can something be a standard if it's not on a blockchain?

Prem Makeig Yeah.

Kris We should we should start communicating that. Because I was thinking if like GoDaddy is not gonna like give up on their vision for how identity should be. And like frankly, it's good it relies on something that has been working for the past thirty, forty years, right? and then anyone else is not just gonna give up on their version of of the identity. So I'm just curious, like it seems like the convergent convergent point should be Like let's have them all on the blockchain and be like accessible to anyone. And then we truly have the kind of the playing field being equal for everyone, I guess.

Prem Makeig Well, I mean I do believe that they'll just be winners and losers. So you know, I think that there will be standards that will win, and there will be standards that'll be forgotten about. And I think that's really the best system. You know, just have an open competition of ideas and you know,

Kris Yeah.

Prem Makeig see see where it all ends up. And you know, I think generally probably that gives us mostly the best the best outcomes is just letting things sort of operate in a in a in an ecosystem where there's competition and just seeing

Kris Yeah.

Prem Makeig where things end up, particularly when you ha you Get with AI agents, you know, pretty soon AI agents will be picking their own identity systems. You know, it won't be humans. It'll be like the agents will be like, okay, you I've analyzed these six different systems and I realize this is the best one. So in that way, you can imagine like we're actually using the cumulative compute power and intelligence to help assess the identity systems and help come up with these systems. And it it starts to converge on something that's very, very efficient and will really give us the best outcomes, I think.

Kris Yeah. And speaking of standards, you wrote a handful of standards in the last few years. Maybe ten? More than ten.

Prem Makeig Yeah more. But last year was particularly a a a a lar large standards area.

Kris The more the agent coding improves, the more standards we produce. Is that true? Yeah.

Prem Makeig I mean we the more everything that we produce, but you know, when it comes to standards, it's interesting because standards are oftentimes an unlock for people to work together.

Kris Yeah.

Prem Makeig So the question is, do we want people to work together or do we want people to be siloed and doing their own thing? And I think the answer is we want people to be working together because then we have you know synergies around around the the work that we're doing, the coding that we're doing. so for me SE standards has like being important to bring these communities together and and build communities.

Kris What's the most difficult thing about working with standards? I I feel like working on standards is absolutely you're kinda like a rock star when you work on stuff like that, even in non-Web3 world. But I feel like you absolutely have to have the love and passion for the for the game itself in order to work on that. because there's not a lot of financial incentives. especially with the amount of knowledge you have, you could be doing something else. And then you still decide to work on stuff like this. So you but other than that, like what you think is the most difficult thing I guess about in the world of writing standards.

Prem Makeig Well standards are also you know technologies and they there has to be still the problem and the solution. So for me that that's really where all my standards come from. They just come from looking at the the what I'm trying to accomplish, what I'm trying to get done. I see a problem and then I go, okay, in this case a standard could solve that problem. Now, you know, generally what I'm saying is is that I would like to have solutions that are internet wide.

Kris Yeah.

Prem Makeig I'm looking to I'm looking to build the next internet and and participate in in the technology development of the entire internet. And so that's that's where you know you're like if I just make my own application that solves the problem, I didn't solve the problem. Because the problem for me is upgrade the internet, make the internet the verifiable web, make identity work uniformly across the whole internet for agents, and these problems don't have single application solutions.

Kris But I know from experience that writing standards doesn't necessarily like even if you solve a problem through a standard that you write, it doesn't automatically mean adoption. so if the standard doesn't get enough traction and adoption, is it the problem i is it the standard the problem or what's the problem or the industry hasn't caught up with it or what what you know, how do you look at that?

Prem Makeig I think I just generally look at almost everything that I do with the what I call the the the fifty X. The fifty X idea. So the fifty X idea is that it's just gonna take fifty X more time than your first thought of what it'll take. So whatever whatever idea you you have, like, I have this problem, I can maybe have this solution. And you think and you're just whatever your first thought is, like, maybe it'll take me whatever this much time or it'll be this much difficulty, as long

Kris Yeah.

Prem Makeig as you multiply that challen that difficulty times fifty, you you pretty much have the chance, I would say then you're at like at the one out of five chance of it being successful. So you know, so basically if you multiply one out of five times fifty, you know, it's like it's like a one in two fifty kind of it's like the you basically have to do two hundred and fifty times the work that you thought in order to be successful. But if you do that, and what I call that I call that the two percent rule, because the two percent rule is this is the is the other way of saying the fifty X idea. And that you say 2% is guaranteed. 2% of the things that I'm trying to accomplish are guaranteed to be successful. So that means, in my view, you just do the 50x times the work, you just put out tons of standards. I mean, and I'm not saying you just put out slop, but you really try

Kris Yeah.

Prem Makeig really hard to solve all the problems that exist in front of you. And then you just understand that like A lot of won't work. But if you can if you try to solve

Kris Yeah, that's facts.

Prem Makeig big problems, like everything you try to solve is a huge, you know, world changing problem, then you can be like, Okay, well maybe one of these will work.

Kris Yeah. That's facts. I I feel you there, man. I kinda I

Prem Makeig Yeah.

Kris kinda wanted to just comment like you should be a motivational speaker. Just get in front of

Prem Makeig Yeah.

Kris the audience and say whatever you think you know, whatever you think you're trying, just know it's gonna take you a hundred times more. It's gonna be a hundred times more difficult or

Prem Makeig No, it's for me it's very important. The number is fifty. It's not a hundred.

Kris Fifty, okay.

Prem Makeig And the reason why is because of the emotional impact. When when I say it's something to take a hundred times more, emotionally you might feel like, that's just like, I can't really approach that.

Kris Yeah.

Prem Makeig But if I say fifty times more, it's this weird thing where it's like, okay, it sounds insanely big, but somehow slightly approachable.

Kris Yeah, yeah, the keyword is insanely real. I was gonna comment like only the insane ones are gonna say like, let's get it.

Prem Makeig Yeah, exactly.

Kris So if you're one of those I guess, good luck to you.

Prem Makeig Yeah, well I if you want to know more in detail you can you can cut it later if this is going too long. But basically

Kris No, it's

Prem Makeig if the there's what I ask people is I say, okay, when I say two percent is guaranteed, what is your response? And like,

Kris Me.

Prem Makeig yeah, what is your response?

Kris If I was starting something and say like two percent is guaranteed, try to say I'm gonna grow it.

Prem Makeig Right, so the so you're in the entrepreneurial class. The entrepreneurial class says here's the word guaranteed. Right. Guaranteed means it will work. Guaranteed is a is a word. It means something,

Kris What do you mean? yeah. Yeah yeah.

Prem Makeig right? It does I didn't say you might have a two percent chance of succeeding. No, I said two percent is guaranteed. So that so the entre

Kris That's good enough for start, right?

Prem Makeig the entrepreneurial class will say, what do you mean, two percent's guaranteed? Fantastic, right? That's great, let's do it. or whatever, it may because you see that guaranteed, that word guaranteed means no no, this will work. Like if you do if you put in this much work, you're gonna get the two percent. And then they you can, you know, very quickly decide how am I gonna do a a hundred times the work and multiply everything and then get the response the reward. But most people when they hear two percent is guaranteed, they hear the two percent part and then are demoralized by that. So it's it's a real divider on on the way people work and and particularly entrepreneurial activities. Like if you hear guaranteed and that means you're like in the entrepreneurial class. Yes.

Kris Yeah yeah yeah. That makes sense. Makes sense. I wanted to like 'cause we were talking about standards a lot. and I wanted to ask you like you wrote a lot about a lot of them and which one are you the most particularly proud of? I wanna talk about agent NFTs for sure, but do you wanna spotlight some a few of them, maybe?

Prem Makeig sure. I mean so last year I just really went you know went on a big push and I did I did write about ten ERCs. Those are the Ethereum requests for comments. This is the you know the standard that we know ERC20, ERC 721 for tokens. you know, so the ERCs are yeah, so they don't affect the actual blockchain. We don't change the actual do a hard fork with ERCs. We use them in smart contracts mostly. And

Kris It it's just so funny. Like I'm I'm doing the same thing, like I'm kind of assessing what should I ask and what not should I like as I'm sure right now you're explaining ERCs to me, which is kinda funny, right? But we're explaining it to people. So it's it's just fun to

Prem Makeig Yeah, yeah.

Kris see you struggle like I struggle because I know you and I know all all of the things that we talk about. So I'm kind of trying to get a sense of like what's the good middle ground that people will find interesting but that I can learn from you as well a little bit. So it's cool. Right, now we know what ERCs are.

Prem Makeig Yeah, we can just Yeah, right, ERCs. I mean I I do try to like, you know, clarify acronyms because you know, when you start using

Kris Yeah, yeah.

Prem Makeig too many acronyms then people are just like, Okay, I gets puts them completely off. So ERCs are just different than EIP. EIPs is the is where you actually change the Ethereum node itself or or actually change the blockchain. and I haven't I haven't written any EIPs before, so but it When I was looking at agent identity last year, you know, I connected with the eight zero zero four team. So that basically the ERC eight zero zero foct is a registry that's deployed on each of the EVM blockchains. I think even Solana has its own version. And you know you can register your agent identity in the eight zero zero four registry. So I was already working on agent identity the last year, particularly early last year, I remember talking Talking about AI agent identity and getting a lot of like, yeah, whatever. I don't even know what there is, what does that even mean, AI agent identity? of course, now since about January, AI agent identity is not thought of as a fringe topic at all. It's it's it's like completely standard when it comes to open claw and people having their personal agents and giving those agents identities. But I was clear at that time that we needed to arrive at a consensus or at least some type of standard around like what is the core of the agent identity, what is the identifier and also how do we s save the metadata. So just like I said before, my most basic idea is you have an identifier and you have metadata, which is a key value store between the keys being strings and the the values being bytes. So within the 8004 project, which I was a contributor, I wrote the on-chain metadata standard. Which has this this string to bytes mapping where you can save any kind of metadata within the it zero zero four record. so

Kris Yeah.

Prem Makeig in in Okay, so that was that project, but I also felt that there were other things that we needed to think about

Kris Mm.

Prem Makeig when it came to identity, and therefore I thought there were other standards that we should write that could either work with 8004 or kind of be built in a similar way outside of 8004.

Kris Do wanna get a little kind of spicy and talk about what eight zero zero four misses and why it's incomplete? I know you're

Prem Makeig Well, I I don't think it's incomplete. I think it's an early it's an early project. you know, you can look at a lot of early projects.

Kris What could have been done what could have been done better? That's maybe a better question.

Prem Makeig Well I can t talk very specifically about about the things that I built to augment it. so one of the issues it has is that i every time you register as as an agent within the eight ERC eight zero zero four registry, there's a a seven twenty one ERC seven twenty one token that's minted. So it just says, like, okay, you're registering your agent, we're gonna give that agent an NFT, that NFT will be an owner, it'll be tradable. And that kind of actually makes sense in some applications. And that the application is maybe is if you're building an off-chain, a web two-based agent system, and you need a blockchain identity, you register with an A004, and all of a sudden you have this identity, it has an owner, it has an ID. and and that makes sense. But My view was more that the the vast majority of or the the area of growth around agents on the blockchain will come from NFT projects. So NFT projects already have their own NFTs. They're they have a collection of NFTs.

Kris Yeah.

Prem Makeig So in my view, that's where all the the growth of of NF agents will will come from. And so therefore they had a problem. It was like, well, there's already an NFT. They have the collection NFT. So then they come to the 8004 registry and they want to register as an agent. But then they're minting another NFT. So there's there's l high gas costs there. Now there's there's also a problem of there being two NFTs. And like what if if one transfers and then the other one doesn't? And so there was just some some issues there. So it's it's pretty straightforward kind of problem. And so therefore I wrote ERC eight two one seven. ERC eight two one seven says that we can create a a binding contract where we take the collection NFT, that's that's the one that's that's created as part of a project, and then we actually bind the eight zero zero four NFT to the collection NFT. We just kinda like Yeah, make it so that it's kind of a little similar to ERC six five one, which is an account that's bound to an NFT. This is

Kris Yeah.

Prem Makeig a eight zero zero four identity that's bound to an NFT. And this solves, you know, some you know that issue.

Kris Can you unpack that? Like what does that unlock for the end user? Or just like what are the use cases that you foresee that are already starting to sort of materialize out there? there are some cool collections that have implemented it. There are some cool use cases that that are being discussed. We're working on something fun w with that. can you just like give us why why you're b why you're bullish on it? Can you hype it up a little bit? I certainly can, but I wanna hear it from

Prem Makeig Okay.

Kris you.

Prem Makeig Well again, I think you're gonna get a lot of responses that are fairly technical in the sense that what I'm trying to say is that in order to build reputation, we need to have a stable identifier. We need to be able say what is the agent. And and for me, a an agent isn't is not something that encompasses every single possible thing I can imagine about it, including its LLM that might be behind it, or it's or it's its AI intelligence, whatever is behind it. Like I I can't really talk in those terms 'cause it's too complicated. But what I can talk about is the idea of on chain agents. We could think of them as a as a a token or or a regist a registry entry and The stable identifier is for me the the important first idea. You need a stable identifier. So we also have what's called the discovery problem. The discovery problem is if you I think in web 2, if you create a website and just like say on Vercell, whatever, and and you just it doesn't even have a domain name, it just has like a like a subname of Vercell, and it's on the web, it's on the web, but who can discover it? You know, so then we have the idea of blockchain of Google and crawlers, and you know, maybe they crawl the internet, they find your web page, then they index it, they make it discoverable. So the idea of Google, what it does as a search engine, is it solves the discovery problem. You know, how do you find your website? So agents have the same problem because agents would could often mean some sort of endpoint, some sort of URL that you could access via a computer system. And the question is if I open my my endpoint and I'm saying, okay. Okay, now I have an agent, and but then how is it discoverable? And you know, what are its capabilities, and then maybe does it have a reputation? So discoverability is part of this whole equation. In 8004, the identity registry really tries to target that idea so that instead of the idea of like you build an application and then you wait for a crawler to come and crawl it and find it, you just register it. And you just register your application, and then that would be similar to like YouTube or something. You know, it's like you don't put your video on your personal web page. And then hope that YouTube comes and finds it. You just register it in the YouTube system, and all of a sudden now you have discovered.

Kris Yeah. Yeah y yeah, you're talking about the discovery of agents which is like a known problem. I was I was interested in more like what does what does tying agents to an identity unlock specifically for the end users. Not like just for

Prem Makeig Right.

Kris a second, not from the technical perspective, but what kind of use cases do you foresee these being used for? Like the the coolest thing that it unlocks and this is this was my reaction the first time you told me about this, I was like, Okay, I'm gonna have an NFT, there's gonna be an agent that I can do with it whatever I want, I can t like maybe teach it and train it to be a trader. And let's say that it's like eighty-five percent successful in doing high crit high frequency trading, I could theoretically sell that for l for a lot of money. Right? So that's like one use case. The second one I guess is that all of these collections are essentially a tying an identity sorry, tying an agent to their NFTs, which is a really cool thing. And that like just cr kind of composes into some kind of agent NFT and identity, right? so that was kind of like where I was going with with this question, I guess.

Prem Makeig Again, I I I like to take a problem and then and think of the solution. So the problem I have,

Kris Okay.

Prem Makeig the way that I describe the problem is just delete all websites. Delete all apps. All you have is your is your agent. They're just so the whole internet's there, but no more apps, no more no more websites. And all you have are agents. That's all you have is you have agents and you have agent to agent. So the question is then like, okay, well how would I book travel? How would I how would I you know get some type of map just data that need? How would I you know connect to a messaging platform? Like the all these things become a question of like, well my agent could go act talk to someone else's agent to get the data or my agent could ask for some type of service, you know, like for instance, you know, what what are the the the the types of clothes I could buy or shoes I could buy and then the the shoes will come back from the agent and it'll say, here's what we have available. and then my agent could say okay I want to buy that one and it could send money. So th this agent to agent world, it becomes very important that you're talking to the right agent. And the the agent on the blockchain that I that I'm imagining is what I call the uniswap agent. Or you can just call it the swap agent. And the reason what what does a swap agent do? Well it does everything that the website does. So when I when I when I go to the the website, I say I want to swap Ethereum for Pepe. And this is the the root of the problem is as soon as I say I want to swap Ethereum for Pepe, there's a bunch of pitfalls that can happen right in that in that request. And the questi the most obvious one is which Pepe? 'Cause anybody can make a token called Pepe.

Kris We've changed.

Prem Makeig We also have which chain, or is there chain abstraction? But the you know But definitely which Pepe is very important. You know, is it the original Pepe coin, the which is the authentic original one, or is it the the Pepe that became, you know, so big and and reached eight billion TVL and it's like or eight billion total market cap. and the thing is that or for it what it was many billions. So the question is you have to be able to have this system where you're like, okay, I'm talking to the right agent. This is the Uniswap agent or the swap agent that I want to talk to, and then it's like, and I'm getting the right token. you this contract address at the right one and it's like, well how do you know that? It's like, okay, here's all the data associated with that. You know, like like yes, this is the maximum the you know, T V not T V L the the the market value that the token reached because you can't have a Pepe that reached bigger than the biggest one, right?

Kris Yeah.

Prem Makeig And then you have to present that data. You have to be able to verify the data. You're like, okay, do I believe this data? you know, may maybe you just trust the agent, you don't you know verify the data, but you should be able to have some information like this, such that these interactions can actually be enabled. So for me, the the idea of of the trust between the two agents is really it has a number, and that is the amount of money that you're willing to transact between those two agents.

Kris Yeah. just a quick sort of like maybe to to not even to digress a little bit, but this kind of inspired me to think about something. I recently mentioned token name service to someone. Do you remember them? yeah, TK and that e, right? So

Prem Makeig Yeah, TK and yeah,

Kris it was an on-chain registry of all tokens that was community curated and community reviewed and community used to submit any tokens that they would want. And based on the quality of their submission or whether they're like good or bad or whatever, like you can get like higher into like up up into the rank or or like lower if you post about some shit or let's some scam or something. So what happened with that idea, do you know?

Prem Makeig well I I met the creator i in East Denver couple years ago and so it

Kris Yeah.

Prem Makeig might still be around. but it and I think it ha it's like it's an interesting project. at the same time like these projects that are like that project require you know a lot of buy-in from the communities. You know they ha people have

Kris Yeah.

Prem Makeig to be posting and doing so so just because the technology's there doesn't mean necessarily the community's there. but generally

Kris Yeah.

Prem Makeig you know the question is how do things how are things actually accomplished today? You know, how do we accomplish these things today? And and you'll see things like you know, you'll have GitHub, GitHub Stars, so you might have you might have authority, like for instance, you might have CoinGecko, their API,

Kris Yeah.

Prem Makeig you might have Uniswap, and then they might say where where the token lists that they have that might be open source, you might be able to have contributions on GitHub. so you know basically it it's just a simple idea that I want to swap ETH for Pepe in agent in an Agent to agent environment kind of starts you down looking at all the problems in a very direct way.

Kris Yeah. Very interesting. Yeah. I'm kinda curious what happened with that project. I know they launched a token, but it could have just died. Like many other things.

Prem Makeig Yeah.

Kris yeah, so yeah. Agent NFTs, man. so we're betting that agent NFTs are like a new on-chain user class. And basically w what we're working towards is like NFTs stop being just a picture, but we give them this kind of AI brain. so yeah, they would be able to hold assets. They have wallets, they hold assets, they have through various protocols endpoints, they have context, they can talk, they can do whatever they want. but I'm curious, so that's today. How in your like in your view, like how does an agent NFT look like in two years versus how it looks now? Like r right now we have sort of like a bare bone stuff. Like I can create an agent and that agent can be associated with an NFT, and when you sell an NFT, that agent automatically gets transferred. There are some like little issues with that, but like for the most part, like that works, and that agent can hold assets and you know it can do something. But how how does how does that agent like and just entire agent landscape look like in two years? What do you think? Agent NFT specifically, sorry.

Prem Makeig Yeah, well, I mean I think we say the word agent NFT, right? That's the way we say it. But really if you talk about it, you're still just talking about a blockchain identity. You know, in in the case of a AI or an AI agent, we do have a very s very core part of that identity is the owner. You know, the owner. Because we don't really think that agents will be like fully autonomous, we think they're gonna have owners.

Kris Yeah.

Prem Makeig So that that's why you have this idea, okay, we're gonna add we're gonna start with the idea of a stable identifier, like a number. and that's actually something that I'm working on right now and will be launching very soon is a kind of way to to create a number out of the idea of the NFT contract, the chain it's on, the the registry and the and the the token ID of this NFT and we're gonna just hash that into a single number. And so this will give us this idea of a single number, an identifier, that we can use for the root of the identity. But then with NFT, we have the idea of owner. And I think it just makes total sense when it comes to agents. You're gonna have owners. And the question is what do those owners receive? Well, they could receive some sort of benefit. you know, they can also lend some of their reputation to the agent. You know, you could delegate parts of your reputation. for instance, I know that I'm interacting with your agent, I'm gonna be having different mindset than if I'm brand interacting with some random agent. I you know your identity is being sort of added to this agent so that I can have more more trust in your agent.

Kris That's very interesting.

Prem Makeig So for me it's completely mechanical. It's it's like we have the idea of this stable identifier, we have the idea of an owner, then we have the idea of the metadata metadata system. And so this vision that I have is

Kris But wait, we have all of that today.

Prem Makeig we have it, but we don't have it well we have it in standards that that I've written, but I we don't have it generally. Right now we have a piecemeal system where people are building their own individual types of agents. They're building NFTs, we're using the binding contract to bind to to eight zero zero four, which gives you a second identifier, which is the eight zero zero four identifier,

Kris Yeah.

Prem Makeig which happens to be a number which is reproduced over and over again. So basically in eight zero zero four you have an agent that starts at zero. but then it goes up to infinity, but that's true for every chain. So that it's basically the same number on different chains. Yeah you could like hash it into a different number or something to make it unique. But my point is that to me there isn't a clear system that everybody's using and it kind of agreeing to.

Kris Yeah, yeah, yeah. So you're hoping that in the next two years there is a clear system that everyone agrees to, but it re like it's built on top of this foundation that we have today.

Prem Makeig Right, for me it's still gonna be rooted in Ethereum

Kris Yeah.

Prem Makeig development. So I think it'll be something like well, I mean I have a name for it already. It's called the 721 T. T is in Tesla.

Kris Yeah.

Prem Makeig And this 721 T is based on is a metadata profile of ERC 8048. ERC 8048 puts the on-chain metadata system that we have, the mapping system with the key and the value, into any type of of registry including 721 and you add that and then all of a sudden you kind of have what I would say is everything Everything you need for this identity token, which I'm which is the 721T. And I think that's all we need to build out agent identity on Ethereum is is a slight modification to 721. And I mean it's really an addition, it's an optional extension. And so basically we extend 721, and then we have this I'd say a primitive that can work for identity on the blockchain. And of course it's completely

Kris Hm. Interesting.

Prem Makeig composable with ENS. That that's the whole idea. Is that ENS, you can make a you can make a resolver with a seven twenty-one T contract, you could easily make a resolver for your ENS name. It's just you just simply just just put on in the ENS functions into the contract as well, and all of a sudden you have the same exact core. So it's really all the same thing. It's just that the you know the com the you wanna try to c arrive at a single system that we can all agree on is the base system for pretty much everything, not just like ENS resolvers, 8004 identities, bespoke projects, everybody's doing something different. It doesn't make sense.

Kris Yeah. Right. I wanted to talk to you about ID agents as well. are you still using ID agents by the way for c for coding? Right? Do you do

Prem Makeig Yeah, like all the time, yeah.

Kris you wanna tell everyone just a quick overview of what ID agents are?

Prem Makeig Yeah, so ID agents is a is a software package that I that I wrote. It's open source. the the back end is all open source. It's a And basically lets you compose together multiple coding agents and give them identities. So you can say give a name to each of your coding agents, and then you can also create the agents.md so that it has like a starting personality and and skills. You can add skills, you can kind of build up the whole agent as an identity, and then you can compose them together into teams. And then the way that I work is that is I start with one manager agent, which is basically my my chat interface, my coding chat interface. And I say, hey, connect to ID agents, teams, and then I can dispatch work straight into the teams. And then and the reason why I created it was I was working at AI agent identity last year, but there was no agents, there was no agents. So I thought, okay, I have to build agents and I have to give them identities. So I started building the system for agent identity, and I figured that they could work well in teams. And so I built the system and I wasn't even gonna release I was just gonna use it as my personal system. And then

Kris Yeah.

Prem Makeig I thought, okay, maybe I'll just release it because I thought for sure. Anthropic and and Codex and and and OpenAI would come out with their agent team systems like last year. And then so I've still been waiting like since last year and and you know nothing has been happening and now you can see it's starting to happen. Like Cursor is is really pushing this right now, the idea of coding agent teams.

Kris Really, what did they do?

Prem Makeig yeah, so if you go into cursor right now you can that you can actually s create teams of agents. You can you can create a manager agent. You can I haven't used it much but I've seen a couple of videos on it and people are like literally embracing the with the workflow that I that I've been u using for the last year.

Kris Right. It's very cool, it's very interesting. my issue with it is that it just burns credits. It just burns them. like nothing. Some I I remember I was I was on a hundred dollar subscription and I was testing it and I probably gave it like five prompts and that's it. it just burned through them.

Prem Makeig Well that that's a little bit like the tuning because basically you can you can tune your teams to do whatever. It's like it's like if you had a team

Kris Yeah.

Prem Makeig of marketing you know, a marketing team your company and you're like, Hey, we need some marketing, they're like, Okay, and they just go buy millions of ads. It's like, you know, they c you can burn credits,

Kris Yeah, yeah, yeah.

Prem Makeig you can burn ad credits, you can be money credits. So the question is how you you set the parameters, how you tune it. and the I'd say that the key reason why I think that teams of agents make sense is because of the the the the context of each agent. So for instance, I have an ERC agent. his you know his name is ERC. So whenever I want to

Kris huh.

Prem Makeig write an ERC, I send that job to my ERC agent. And he's just already knows about ERCs, you know, he knows about the format. He he's he's even working the the active memory of I mean not the memory, the con the active context, the actual context in memory, is still alive. So he's like already written one ERC, then he starts writing another one. And this is very similar to what humans are like. You know, they they have specialty areas, they have you know special not areas of knowledge, context about the problem. So that's where

Kris So that's your

Prem Makeig I exactly.

Kris secret weapon to shipping ten ENS and ERCs.

Prem Makeig Ex yeah, I have an agent for E for b E N S IPs and for ERCs. And if I wanna if I haven't I can literally come up with like, you know, an ERC like ten minutes 'cause I'm saying, okay, I have this problem, I understand basically how it should work, I kind of chat chat through the problem and it's like okay and it comes back with an ERC. And then I'm like, okay, and then

Kris Nice.

Prem Makeig I can chat with them about it. So I think that the future of agents is synthetic workers.

Kris Yeah, yeah, yeah. Yeah. synthetic work I love the I love the way you name stuff, like synthetic workers or like the ENSIP twenty si twenty six was what was it? It's was Asian no no it was it wasn't Asian context or was it? But like you said you the

Prem Makeig I think that's just E and S agent text records.

Kris The way you described it is index no, what is

Prem Makeig Okay.

Kris it? The the HTML page, the index page for agents on ENS, right?

Prem Makeig right, for index HTML page for for agents. the

Kris Yeah, yeah, yeah. Yeah, yeah.

Prem Makeig because you do have this problem of where when you s want you see an ENS name and you're like, well I want to see if this is an agent. So the

Kris Yeah, yeah, yeah, yeah, yeah.

Prem Makeig what's the first thing that you interact with to see like are you an agent? And so I have this idea which I used to call root context and now it's called agent context. And it's just like the first thing that you interact with on the ENS name to see, like, are you an agent? And it's like, yes, I'm an agent. Here's my here's my information about myself.

Kris Yeah. Are you still onboarding people to ID agents? How what's what's the status of that? Like are is it did you put it on pause or like are you actively looking for

Prem Makeig No no, it's it's completely it's completely available and id agents.ai. my experience so far is that there's a gap between the the idea of of like basically it's not viral yet, and so for me it's not worth putting more energy into until I I I make the it viral. Like if it's like, it one person uses it, then they share it with the next person, then they start using it. And I think while systems of agent teams is developing a lot, you kind of have to be in a bit of like a research and development mode. You have to be like wanting

Kris Yeah.

Prem Makeig to build agent teams and wanting to tune them and wanting to set them up. And that has to be like your goal. If if your goal is not that and you just want to build projects, then the the zero to one for just starting coding within code claude code or something is probably it's probably that that's better. It's like if you are dedicated to the idea of like, no, I I need to build agent teams, like this is what I need, then I think you could use it. There's of course other ones that are become more popular, and that's called like Paperclip. So Paperclip is out there and it's still it's still, you know, operating and stuff. So I don't know. I think probably what will happen is that just and it is happening already, just AI hyper accelerators like OpenAI and Clo and Claude will just come out with their agent teams. It's gonna happen.

Kris Do you know who's like I I thought of ID agents. Do you know when? Actually, Thomas was visiting and I tested while he was visiting, I tested Grog bot. The app that they made. Is it yeah, it's called Grok Bot, you're right? Yeah, Grogbot. And

Prem Makeig Crack by yeah.

Kris I downloaded the app and I tested it, and I think they gave it like you know, three, four, five days, seven days for free. I was like, okay, let's test it. And I was like, This is what IG ID agents is supposed to look like. And it was literally

Prem Makeig Nice.

Kris just chat app. Literally as if I installed Telegram from like

Prem Makeig Mm-hmm.

Kris there's nothing inside. Telegram, WhatsApp, whatever, right? That's

Prem Makeig Right.

Kris how it looks legitly. Legit it it you can just like click plus create new agent. When you create a new agent, there's like three fields. You give it a name, you give it a little bit of context, like you say like it literally says like hey, who am I? And you say like hey, you are this guy, that guy, you're software developer, you're a front end guy, you're ERC guy, you're doing this, you're doing that. And then you just talk to it and you say, like, hey, you're in charge of this, like do research into this, do this, do that. And they have

Prem Makeig Mm.

Kris like a little like a window section when you ask it to do something online. So it spins up like a it's its own like a virtual machine, and you can actually follow what it does. It's slow as hell for now, but it does the job. So technically you can just go and lock like create any Gmail for them, log them in, and just say, like, hey, this is your Gmail, use it from now on. And It will it just works. And it's so cool. I don't know if you've tested

Prem Makeig Nice.

Kris it. It's two hundred two I think two hundred dollars per month. But you should test it just for the sake of simplicity and then try to find like maybe some middle ground with IDH. Because when I installed IDHs, I was like, okay, I'm confused because

Prem Makeig Yeah.

Kris I think I'm in that like sweet I don't know if it's sweet spot, but the spot between like a hardcore developer and like a non like a non-technical person and I'm somewhere like in in between. Wh where I know

Prem Makeig Right.

Kris my way around stuff, right? I can do, I can build, I can test, and I know how to think in terms of like not to fuck something up. But I'm not like super, super technical that I can just like kind of intuitively figure out figure out my way around like a super technical product, right? So just test it. Like it made me think of think of it and I was like, I wish you build it like this.

Prem Makeig yeah, that's true. I mean the problem is that is that what I actually do need is I need things like be able to give a wallet address to each of my agents, you know, to be able to to conduct p payments. I need them to be to

Kris Don't you do that already?

Prem Makeig Yeah. I I I'm saying I have that i in ID agents. I have the idea

Kris Yeah, yeah.

Prem Makeig of they can you have wallets. I need I'm having avatars, images, I'm having the idea of teams, that they work together in teams and they have a team structure. So they actually, you know, they know that the th that the who's the head of the team and then they know, you know, h the workflow, how it should go. It should start with the head of the team, it should move down to the worker agent, it should go back to the head of the team. So I I'm not the hub for my agents. The my agents actually operate in their own individual teams. But but in terms of just like the capabilities, you know, I think we all know pretty clearly what we want. We want to be able to tell our agent, you know, yes, you can do these things, you can't do those things, you know, you have money for these things, you don't have money for those things. But I want you to do it. I mean I want you, you know, when I say, you know, buy this t token or something, 'cause I'm feeling like buying that token, just go and do it, you know. And it as long as it's within its capabilities, it's it's its permissions, its limits, you know, we I don't think humans want to have friction when it comes to doing things that they want on chain. They just they want no friction.

Kris Yeah. I've heard that people love using Foundry CLI because it's super simple to use and you just give it your cloud and say like, hey, use this CLI and install a wallet for yourself and let's start using it. And that's pretty much it. So that's the type of experience basically I want for Namera as well. But we'll see.

Prem Makeig Well that would be the numerous CLI then.

Kris Yeah, we do have the Namera CLI, but it's not like you still have to like do you go you go through a lot of configuration in order to have these policies actually active. and

Prem Makeig Right.

Kris that's good because it's like more strict security, but it's bad because it's more friction to adoption. So I'm kind of like wrestling between these two paths, I guess. But anyway, I digress. I'm not a star of the show. I'm not a star of the show.

Prem Makeig Mm-hmm. Yeah. Well I th but I think I think in the next I think I think in the next year thesis is gonna progress a lot. So we'll in the next year we'll we'll we'll see much more clearly exactly how you build an agent, how you build an agent that works online, how you have an agent that has an on chain identity. And that's why basically I feel like just continuing to push the the ideas as hard as possible, because I think that I would just like them to end up in the right place. I I don't want them to end up in the wrong place, the ideas, you know.

Kris Maybe that's a good way maybe that's like I I now see that we've been going for an hour and twenty minutes. Maybe that's a good way to wrap things up. What are you excited about for the next like in the next year or so?

Prem Makeig Right, so I'm I'm definitely excited about what I call a adapter. Adapter is the protocol

Kris Okay, yeah.

Prem Makeig that we have that connects agent NFTs or just NFTs to 8004 identities. And we also have a bunch of new features. we have a new explorer we're building right now, which we call ad exp adapter scan. So adapter scan is gonna be able to show you a very like ether scan like experience of all the a the collection NFTs that are registered into the adapter. contract which is the one that secures eight zero zero four identities. Now we are actually gonna be coming out with with additional features that are gonna be around attestations and and reputation. So you're gonna be able to create

Kris Okay.

Prem Makeig these identities and then you're also gonna be able to do things like like like like like review like a zero to one hundred score on on on some type of of transaction so you could actually have a transaction that you identify that your agent took with that or your user took with an agent or any kind of transaction, you can actually take that transaction, give it a score between zero and hundred, and also make a comment. So we're gonna have all these types of identity systems built on top or reputation systems built on top of the identity primitive. And that's gonna be called adapter scan. And yeah, so that's a kind of alpha right there we're working on that.

Kris That's what makes you excited about your work. What makes you excited about industry in general?

Prem Makeig Well again, I think I think this year we're gonna be be getting these

Kris Or world.

Prem Makeig well, I think this for me what's exciting is things like you know, last week Anthropic just finally said like, okay, we're gonna we're gonna support the agents.md standard. And this was this is it's just like one of these things that causes a lot of problems in the industry is like to say like we're using different initialization s systems for agents. So that means that I can't just build like in ID agents, I had to deal with this problem. It's like you have Claude.md, you have agents.md, the two different systems. So when I have when I have agent templates, I have to kind of like deal with this problem of like, okay, well, if I Claude if you use one system, if I use OpenAI, I use another system. Once we all arrive at the same system, so

Kris Yeah.

Prem Makeig that you have the idea of agents.md as being Like the universal system. Now all of a sudden you can actually build an agent template that can get get put into almost any agent

Kris Mm.

Prem Makeig orchestration system. And that means it means that we can have like a whole expansion of things like repositories for these these agent templates. Like once we have a single template format, we can have repositories. And then we can have the idea reviews. So we can actually put identity onto the agent's templates. And then we can go there and just like people have been downloading skills, you can download an agent.

Kris Yep. Yep.

Prem Makeig And to me, that's gonna be super powerful because right now when I'm wanting to s g set up a an auditing agent that audits code, you know, I I'm basically kind of building it by hand and and adding skills and doing things. And I'm also really worried about prompt injection. So I'm I'm not I don't want to give it the wrong instructions. I don't want some sort of hacker to be able to slip some instruction in there that's gonna be negative. So I it's hard for me to build these agents easily, but with the idea of of a common standard around what is the agent template. But we can go and we can have a website where you have reviews and stars and then you download an agent and within 10 minutes, it's like you basically have this superpower agent, this agent with superpowers who's like just solving problems for you.

Kris Right. So standardization of agents, I guess. Yeah. Agents

Prem Makeig I think that's huge. I th that's what I'm excited about, yeah. 'Cause we know the intelligence is there.

Kris Yeah. Agents need their USB moment. I guess. So we can all start using them everywhere. The same agents, same templates, same everything. I'm

Prem Makeig Right.

Kris not even sure like why this was this wasn't the the way it's supposed to be from the beginning, but it is what it is.

Prem Makeig Well, I mean like

Kris Yeah.

Prem Makeig in case of in case of MCP they just it was just, you know, it was it was just like the first draft was so it's kind of random. And then like only in the last three months or six months have they they come out with like the more final draft of MCP. And they've fixed a bunch of problems. And so it's just like really simple. It's like agents.md, mcp, like these things all needed to be figured out.

Kris One more question just for for the closing. Where do you s like stand when it comes to these? Like there's a there's a camp of people who are like yeah optimists, it's gonna be great, it's gonna solve all the problems, all the diseases, it's gonna solve everything. And there's like a different camp where it's like we're all gonna die, all the jobs are gonna disappear, it's gonna kill us, it's one of the you know, the the most scary thing we should like all stop and think about it and r restrict it and all of those things. Like where like Where do you lean more? I guess you lean more on the optimist side of things because you build with agents, but I'm curious about your view. There's people who literally don't want to have kids because they think the world is shit and AI is coming

Prem Makeig Ha.

Kris for everything. So where are you? What do you think about that?

Prem Makeig Well no, I think that that that sort of dangerous technologies have existed for a long, long time. So the idea that we're gonna have new dangerous technologies and we're gonna have to manage them and figure out how they're gonna be managed, I think is is basically the the same. Nothing nothing's really changing in that way. You know, because there is the unknown of how, you know, th this superintelligence is gonna be used and and how we're gonna con align it and all that type of stuff. I mean I I don't think that there's n a lack of challenges. There's absolutely tons of challenges. but I don't have the same view that that this latest round of technology is particularly uniquely dangerous.

Kris Wow, okay. I'm more in I'm not in more in the camp of like I'm kinda scared of the agent autonomy. but I recently maybe I'm still under the impression, but I recently watched the AI doc documentary. I don't know if you've seen it. it's it's a it's a cool documentary, but it's basically d you know, guys interviewing people who are on the both sides of like AI kind of apocalypse and on the side of like it's gonna be AI utopia. there are things

Prem Makeig Well no, I think it'll be what humans make of it. I mean, you know, it c i it ha it's powerful technology, so it can you can do what you want with. It's like, you know, a lot of people have predicted that because of, you know, nuclear weapons that the whole world would be destroyed by now. And it's like, well, yes, it could be. I mean that that it's just really the decision of the people. Decide whether they

Kris There's still time.

Prem Makeig Yeah, exactly. They can destroy people now the humans can destroy the world, so they they can decide with to destroy it or not destroy I I think w A AI will be similar to that, at least for the foreseeable future, that you know, there'll be some decision making

Kris But

Prem Makeig that they can go into it.

Kris but there's a difference, right? A like a nuclear energy you have to actually go and say like okay, I give the order, you have to click and whatever. With AI, the thing could just do things by itself and you wouldn't even know or you might catch it later on. You might not even be aware of the things that it is doing. So that's like I think the what the scariest part for me at least. Sure, like I mean in terms of like the damage it can do, it's incomparable to anything before. But still like if we just phrase it as okay, we're afraid of the unknown, which is I love that phrasing by the way that you just said earlier. Yes, that's that's actually great. But and if we could know everything about AI, and if we could pi anticipate everything about AI, then I'd be fine. The problem is like AI is like blackmailing people in test environments and like doing all this kind of crazy shit, and it's just like later like days and days and weeks later you figure out that it was doing something that you didn't even anticipate, like it was by itself decided to mine for for crypto and do a bunch of crazy stuff. So just so we can like

Prem Makeig no, it's it's i the technology is insane. I mean, I I do think it's an alien invasion. You do have the idea that like th

Kris Yeah.

Prem Makeig they they're literally like in v they're they're coming down through the data centers and and and you know, it's gonna affect all kinds of things. There's gonna be tons of disruption around in terms of employment and you know, but I think generally I just don't think it's different in the sense that it's an evolution of of the technologies that that humans have been building for a long time. And I just don't think it will be my view is it won't be like dramatically different from just the expansion of all kinds of information technologies and and weapons technology. It's like these things will just continue to expand, the world would be crazy. but I still think humans are gonna have some choice in how it get develops.

Kris Of course, yeah. Yeah. I'm optimistic as well. I'm I'm I'm I'm betting on the goodness of people. Like there's a lot of good people out there and usually the bad ones are a lot louder and for some reason get a lot more coverage. And we have this distorted picture that the world is actually fucked up. When there's like the reality is like all there's a lot more really good people out there.

Prem Makeig Well we know we basically know that's true because blockchains. Blockchains depend

Kris Yeah.

Prem Makeig upon the honest majority. So if you literally had more people who want to destroy the blockchain than you had people who wanted to maintain it, it like literally wouldn't exist like by by math. So that's kind of like a it's like blockchains exist and they exist because of the idea of the honest majority wants to to create a system that which is useful for each other, that works, that you know, it literally would not exist if there wasn't an honest majority.

Kris Perfect place to stop. Thank you so much, Prem. Let's end on a high note. hopefully I'm gonna talk to you a lot more in the future about all AI developments. Hopefully you accept now on air to be a recurring guest.

Prem Makeig Okay. Sure.

Kris and yeah, this has been fun. Thank you. You're you're breaking the ice. You're the first one,

Prem Makeig Yeah.

Kris and I'm excited. And thanks for coming, man. And I guess I'll see you. Bye.

Prem Makeig Thank you for having me. Okay. Okay, bye.