Technology, Connected - Sam Altman’s World ID: Can Iris Scans Prove You’re Human Online?

Episode Date: March 19, 2025

Ajay Patel, Head of World ID at Tools for Humanity, joins Mark and Jeremy to discuss digital identity in the age of artificial intelligence.As AI-generated content spreads across the internet and bots become harder to distinguish from people, existing verification systems are starting to fail. CAPTCHAs are increasingly ineffective, while conventional identity checks often require users to hand over sensitive personal information.World ID proposes a different model: verifying that someone is a unique human without revealing their identity. The system combines iris biometrics, zero-knowledge proofs and decentralised infrastructure to create what Tools for Humanity describes as a privacy-preserving proof of personhood.In this episode, we discuss:How Sam Altman’s World ID system worksWhy World ID uses iris scans to verify unique humansWhat zero-knowledge proofs mean for online privacyWhether biometric data can ever be collected safelyHow AI agents and bots are changing digital identityWhat happens when platforms can no longer tell humans from machinesThe conversation examines the technical design, privacy concerns and wider implications of World ID. It also asks a more fundamental question: what kind of internet remains when proving you’re human becomes difficult?Please enjoy the episode and share it with someone interested in AI, digital identity and online privacy.–Timestamps(00:00) Intro: Future of Digital Identity & AI Challenges(01:15) Meet Ajay Patel: Head of World ID at Tools for Humanity(01:44) WorldID: Why Digital Identity Matters in the Age of AI(05:40) Global Digital Identity for 8 Billion Humans(07:38) AI in 2025: How Advanced is AI Really?(08:17) AI Bypassing CAPTCHA: The Security Crisis(12:54) ZK Proofs Explained: Privacy-Preserving Tech Behind World ID(15:20) How World ID’s Iris Scan Works (No Data Stored!)(18:54) World ID Data Security: Breaking Down Multi-Party Computation(22:56) Hot Button Q&A: Apple vs Android, AGI Threats, & More(24:19) World ID Use Cases: Gaming, Voting, & Anti-Bot Systems(25:13) Exclusive: World ID x Razer to End Gaming Cheats(30:56) AI Hype vs Reality: Are Bots Just “Yes Men on Servers”?(35:35) Decentralizing World ID: Open Protocol, No Corporate Control(38:15) Question for Next Guest: Hollywood’s David Bianchi on AI & Art–Thank you for Thinking On Paper. Your future self will thank you.   

Disruptors and Curious Minds, welcome to another amazing episode of Thinking on Paper, where we bring you behind the scenes of the latest technology trends, and we together figure out what it all means for future, for life, for work, for family. My name is Jeremy. This is Mark. Mark, what are we learning about today? And let's tee this thing up. We are tackling a question that could define every listener's future, every single one of you.

How do you prove you're a human in a world of AI? humanoids, deep fakes and bots. And why does it matter? And there was a time not very long ago when I would have laughed at that question. And maybe so much listeners are laughing now, but chat, GBT, Claude, right now it's making everyone sound like a faxie-similar emoji-wielding robot.

But very soon, AI-generated content and human-generated content will be indistinguishable. And it's all about trust, isn't it, Jeremy? And unless you want to be Fox-Modeled, Mulder, that's an X-Files reference and trust no one, humanity needs a verifiable way to prove you're a human. And today's guest is going to give us the tools for humanity. Great T-Up. So let's introduce. Who do we have today, Mark?

We have R.J. Patel, his head of World ID at Tools for Humanity. Yes. Welcome to Thinking on Paper. Thank you for thinking on paper with us, RJ. Thank you for having me. Thank you for being here. And Jeremy, I will let you paint us a picture. So one of the questions we always like to start with, Ajay, because tech can be pretty complicated. So imagine you've got a group of relatives sitting around the table. And they ask, so, Aj, what do you do? And what is this all about? And why should I care?

And how can I figure out what's going on? That's a great question. And how do you do that without using the jargon that makes everyone's brains break, right? Exactly. Yeah, I mean, I think AI took everyone by surprise a couple years ago when chat chip D came out and it's advancing faster and faster. And, you know, I think every subsequent release of these new things, that surprised the world. And a lot of them are pretty amazing and eye-opening. And so there's a tremendous opportunity for us to be more productive and, you know, outsource things to AI.

And people are talking about robots and a lot of cool things. And things seem to be accelerating. Things are happening faster than we expect, although some things I think may not happen as fast as people expect, but then other things are happening very fast. And so when we look at the Internet and the systems, the tools, the process, how we access things. The entire infrastructure and actually our daily lives are not built contemplating AI, right?

And so when, you know, when tools for humanity was created, it was really with that premise, like, hey, like if we rethink how the internet and how our daily lives that were going to work, we really, there's an opportunity here to really think about the tools for humanity and rethink them and then figure out what we should build in order to help humans usher in AI and then maximize benefit of it. So AI coming into the world is an interesting introducing a ton of challenges. And we actually see glimpses of that already, or the way I would explain

it to my parents, for example, is like, hey, when you go on a website, you're, you had to type in these fuzzy numbers and you can't find your glasses and it's super annoying. Why are they tween? Because they want to know you're a human and not a bot. Or you're going through some experience and suddenly this platform asks you for some information that you feel is kind of weird. Why are they asking for me for my driver's license to verify my age? It just seems like complete overkill. And these, these sort of things are indications that there's this primitive on the internet that's kind of missing, right? Or, and it's really about humanity. We don't really have a good way with certainty that can cover all of humanity in a way that's kind of a real baseline. So we think about a primitive as a baseline where

everyone can get equal access to a primitive. Most of everything we do is based on something financial or there's government legal identity, but then 800 million adults don't have a government legal identity. And so when we're thinking about how you prove your humanity, it needs to be something that when I prove it versus when somebody in another country proves it, that proof should be trusted the same. So you mentioned trust. I think that's kind of our mission. When we assess the entire landscape of what we could do, we said, hey, you know what? The first thing that the cornerstone that needs to be there is actually, hey, I need to feel to prove with certainty that I'm a human and not a bot. And if I can do that, and full control and also not expose who I am

on the internet, because there's almost this requirement that you have to tell everyone who you are in order to do anything online these days and without data minimize the internet, but then also could provide a lot of certainty in that interaction and trust starts with certainty. If you think about money, there's counterparty risk and there's a lot of things and you see a lot of inefficiency. And from a user perspective online, it's a good users end up paying with friction. Those are all the things that we think about when we're thinking about proof of humanity and building a tool for users to prove that they're human and not something else. Good users pay with friction, Jeremy. Yeah, fist pump.

I love the humanity. I'll rephrase that. I am on a different path to 100 million people in the world who don't even have access to a bank account. Somebody living on a farmstead in Somalia has a very different need for proving their humanity than I do and I will in the next decade. So it goes from creating a bank account

to verifying website access to voting in the future, to getting a mortgage in the future, getting married, having a relationship. It's very, I don't know what the word is to describe what I'm trying to explain, but that difference in humanity for different people and how do you make the trust equal for everybody? So when you think about like what makes us human, it's something that's actually, it's physical or something about us. Like when we, when we were looking at all the different ways that we could assess humanity, we realized that, hey, it has to be something that's unique about that person. And then you think about, well, what are all biological things that make you unique? And if you want to really build a system that can scale to 10 billion people, it has to be something that has a lot of data.

And there's things that are practical and impractical when you really think about all those things. But if you kind of in a scientific way go through all of them, there's certain what we call modalities. And but I won't say the jargon, but there's certain ways that you can, you know, like fingerprints or palms or face or eyes or whatever. There's certain things that, you know, if you think about face, right, there's doppelgangers. And so actually technically after about 20 million people, things start colliding. Right. And so you might be able to, you know, the system would mistake, any technical system would mistake you for someone else. For example, Jeremy is a doppelganger of James Hetfield for Metallica. Exactly.

And so, you know, if he got James' phone, he'd be able to totally get in. So, Adjie, real quick, I wanted to, as we continue to kind of set the tone for this conversation, why is it so important to distinguish AI specifically from humans? What's the overarching need for that? And then let's dive into what you guys are doing with it. Yeah. So one of the things to think about is if you're providing a service online, why do services online put CAPTCHA in? It's because they can't tell the difference sometimes between two clicks.

So very basic today, right? Today, probably. We're not talking about agents or anything like that. And so if a service can't tell the difference between two clicks, then a lot of things can go wrong. Right. So number one, accounts are being created. Hey, someone's doing something on our system that, you know, we don't like. And there's lots of examples from like social media or video or whatever.

And so you have that you can you're you're open for abuse. That's one thing. And then and that abuse can be automated. That's basically a bot. And then if you take that bot and make it smarter, it's AI. So bots are just, you know, bots are just kind of dumb AI. But then today's the dumbest AI is ever going to be. They're only going to get smarter. And so now you can imagine a world where, hey, like Jeremy, Jeremy clicks and does access to the internet. Well, something can train on how you act online. And then it's indistinguishable. I'm an online platform and the real Jeremy's clicking. But then. an AI mimicking him is clicking, how do I know? There's a lot of issues around trust where that certainty is important. And over time, as some of this automation and bots and AI become more sophisticated, and you can't distinguish between humans and bots online, guess what? You just have to assume everyone's a bot. And so then guess what? Everyone has to go through things like CAPTCHA.

But the funny thing about CAPTC, or not funny, but the interesting thing about CAPTC today is that there's already videos on X of AI bypassing CAPTCHA. So then what? well you know what now we need to kyc everyone the only way you can access is by giving me your government ID or your passport right and so now you're so now imagine if every website did that your personal data is just everywhere and by the way like documents are easy to fake too that's that's not robust either but from a from an individual if all your data is out there that's also even more data for AI to train on you know bad people to take and train and so there has to be a better

solution so can you so the problem is that hey can I identify myself as a unique human online without exposing anything about myself, providing that trusted signal that, hey, this is a unique human. And so therefore, it's nothing else. It's a unique human. You don't know anything else about them. Yes, you can collect more information. But at least as a primitive, you can differentiate the two.

And that's how I explain it. Is this about proving I am a unique human or proving that I am Mark Fielding? Or is it both? You think if it's building blocks, right? So imagine Legos. The first Lego has to be that you're a unique human because your mark is another Lego block. And not everyone needs to know that. If you think about it in like blocks, then this bottom block doesn't exist.

This is the only way to know that you're unique human is by you telling somebody your mark. And that's where the problem starts. Now being able to tell it. Now there's digital identity things that are happening, right? Like mobile drives access to the US or some of the stuff happening in the EU where they're focused also on this upper block on being able to tell. that your mark, you know, using a digital credential, but you still need that building law.

I'm going to try to not dive down too far in a rabbit hole. Mark, ring the bell if I get there. But you got me thinking on this Lego block thing. So Lego block number one, I'm a human. Lego block number two, I'm Mark Fielding. What could other Lego blocks look like? And could those Lego blocks automate the connectivity of community? So, for example, layer on top of Mark Fielding,

you know, I love a specific science fiction writer. I like to snowboard, like different qualifiers of someone's identity to a community. Yeah, totally. I think the way to think about it is that World ID, the Humanity and World Network, is solving that first block. And that's actually a really hard block to solve because you have to create something where 10 billion people have that unique block and you can guarantee that that is unique. And so then if you think about like your online, your online activity,

in the Western world, we buy things online all the time. But the average user, like my parents, what they do online is email, like go and watch some videos or whatever. Nobody needs to know who they are. They can go days without ever anyone needing to know who they are online. And so to the previous comment about being something ubiquitous that's global is really our focus. Because if you're somebody in Africa or you're somebody in a village that doesn't have a, is not known to any official government system, eventually right now online, you don't exist. It's like you're not even on the planet.

But if you have this double block, a digital credential that identifies you as a unique human, you're known. That's like the most basic access credential. There's obviously you need internet and all that other stuff, which are other higher layer problems, other blocks. But this basic block is needed and that's really what we're focused on and scaling. So now you can layer things on top. So what are we building? We're building a primitive that's actually a protocol that anyone can build on top of. So you can have, so there's a lot of interesting projects out there already.

where users can build their profile, but it's self-custodied, but then they can share things in a controlled way. The digital identity framework in general has the concepts of selective disclosure where you share only the bits of information that is required for the job you want to do, and the business on the other side or the product or the service on the other side will only ask for the things that they need, and that's actually cool because it's a part of GDPR. But underneath that, the service still needs to know that there's a unique human behind it. So that's the way we think about is we're kind of agnostic to those layers. Now, as we build this bottom Lego piece, you know, we can create the hooks to kind of those things.

And that's something that we're exploring. But right now, our focus is actually get this proof of unique humanity into everyone's hands. And then there's enough use cases where, you know, even if you just said, hey, you don't have to do fuzzy numbers anymore. It depends on your World ID and you go right through. You've just explained what most people get wrong about World ID. And I think that me and Jeremy probably misunderstood or didn't understand probably. before we started researching this as well as that you're really about building that bottom

Lego before things get built on top of it, very important. So let's explore or how you do that. And obviously we want to dear clear like the tech jargon and really figure out like demonstrate how you take a person in and you know, help them create this ID and

how does that work? But you can talk about ZK proofs because then we can link to it. Yeah. You love it. You love this. Yeah, ZK proof. It just seems like something which it's one of those, if we want talk about building blocks and Lego blocks and all of the shows that link our shows together. ZK proofs this and what they do in proving what you want or who you are without disclosing other information seems an integral building block in in the future of technology. I love it. I love

the idea of it. I don't understand it, but I love the idea of it. Yeah, it's actually really complicated. It's based on super cryptographic math and I think there's a point in math where things get really abstract. It's like physics, they get pretty abstract and it's like there's special brains that understand that. But the end result of a ZbK proof, CK proof is basically there's something that's real, there's something that sits in between, translates that and said, and then so you can ask, so basically you ask a question, is this person a human? And the only thing you get back is yes,

and something that identifies them as unique, but you don't know anything else about them. And that's basically how it works. So how do you get that proof of human? I talked a little bit about what modality, this face, his fingers, they're this. And so we're always exploring how to build a proof of unique humanity in a way, you know, using different modalities. But the one that we settled on and we're really scaling right now is based on your iris. So it's not your eyeball. It's not your retina. It's your iris.

And again, if I'm explaining this to my dad, I'll say, think of iris as like a really beautiful QR code. It just has a lot of data in it. And if you can take the picture of both eyes, there's actually a ton of data that you can extract and build and identify someone as unique. What we built is a piece of hardware called the orb. There was a lot of engineering that went to get to that point. What we realized was, hey, the fidelity cameras, the sensors that we need, all those things don't exist. There's a lot of things that we built to get the orb to where it is today. But it's really like a really high resolution camera.

It has a regular camera and it has an infrared camera. And you walk up to the orb. And it's actually, there's a lot of fraud risks that is built into, which we don't get anything to get into. But a user walks up, has an anonymous interaction with the orb. So nobody knows you went there. It takes pictures of your face. and your eyes and it pushes that data to your phone. You go up there, takes a picture of you, pushes it to your phone.

It doesn't go anywhere else. From your phone, you enroll, you get, you register as a easy human. If the network's never seen you before, you're, you're given a new world ID and if it's seen you before, then there's like, it's kind of your authentication. Now you have this thing that on your phone that's not tied to the pictures that were taken to the order, but it's this number, let's say. And from that, a zero knowledge proof can be, can be shared with, you know, anyone that you're a unique human, anyone online or anything online that you're unique

human. It all cryptographically ties all together based on how that proof was generated originally. All of the pictures that you have are totally isolated from the proof of humanity and then again isolated by the proof that's generated for you online. So even if you wanted to, you'd have to take your own picture and send it to somebody. That's kind of the way the system works. And so that's the most basic building block. And then, yes, like there's going to be scenarios where you have to share real information, they're signing up for a bank account or whatever, but at that point, that's a different use case. And so you would then follow normal paths. But the beauty of this is that whoever you're sharing, whoever you're connecting to has a Sony human signal that they can then

build their account structure on. And on their, on their side, when they authenticate you or when you're onboarding, they can, they don't need to do CAPTCHA. They don't need to do a lot of different things and it simplifies things and data minimizes things. Right. And that's kind of the vision of where we're trying to go. I love data minimizing. You've used that a couple of times, and there's a lot of data out there, and useful information is a little bucket within that data. So the more we can minimize, the noise, the better the signal is.

I want to nerd out just a touch. Mark, keep me honest here. No, nerd out. I think I know what you're going to say. Wow. We'll see. We'll see if we're super positioned here. No.

So that's a really good point, Ajay, because when people are like, oh, man, I got to walk into this building and I know they're going to scan my eyeballs and where's that information going to go. And I love that you clarified that that information immediately goes to your phone and then also cryptographically stored. So can we talk a little bit about secure multi-party computation just from a high level and can I run my understanding by it by you and you keep me, keep me in level? SMPC, this is a way that iris data is stored as I understand it. So you scan it. It's in there. It gets uploaded, but then it's broken apart into tiny little bits and pieces, right? Two little bits are two little pieces that are then encoded 12,800 bits. Each bit is then encoded into a random

integer, and it's just, there's a way that it keeps this separated. And the only way for it to come together is these multi-parties when it's needed and requested. So how does the multi-party thing work? It might get my Irish scanned, and it's put in there and it's segmented and protected. Who has the little pieces when I'm looking to push the button to use it. Yes. So the secure multi-party compute implement. So I'm not an expert in it. So I'll explain kind of how I explain it.

So secure multi-party compute, our implementation, we call it anonymous multi-party compute. And the key there is anonymous. So what's stored is not, you know, a picture of your higher risk. Like you said, like there's some data that comes out and it's distributed across multi-parties. On our roadmap, what we did was our goal was to have multi-parties, hold those different pieces. And so none of them are tools for humanity or world. It's actually four independent nonprofit universities, those types, and we're going to expand that number. And so the

idea there is that in order to compromise a system like that, you actually need to compromise all those parties, which then becomes more and more difficult. That's number one. But then it's also cryptographically science. It's also cryptographically stored. So even if you get some of the data, it's going to be impossible for you to pull it back together. And then even if you have like that code that came off your phone, let's to say that's what it is, right? Then you can't reverse it back to an image. So the interest, the, the, the,

the, the, the, the, the, the, the, question before I joined. I was like, hey, like, and they're like, well, actually, if you reversed the code, you get this, like, blob. It doesn't even look like anything. And so, and so the good thing is, so the thing is that you can never tie back.

And so the user is kind of in, in full control of the data, right? That's kind of the key. How that kind of storage and joining works is really through this, this super complex cryptographic system. that's doing this math. Right. And so the way that I think about is that you ask it a question

and it gives you an answer. So it's really like you're just talking to this crystal ball. It happens inside and how it gets all split up. It's kind of opaque to who's asking the question. What that's called is privacy enhancing technology. So you'll hear about PET and there's a whole like research and innovation

that happens in that space. And then you have these like trusted execution environments which are like where you can put data in, you can process it in there and it will give you answers out. But you don't know what data it used. or you don't know or none of that data ever leaves that environment is sort of like the technologies. And so by having to taking on World ID, I was at Google. And what's interesting is like P-E-P-E-T and T-E is mostly happening in like Google Labs, Google Research.

But in production, like there's not a lot of that. At the scale, that implementation at scale, actually like World ID has it as far as I know. We're doing all those things at scale. We have millions of users interacting with these systems. So it's pretty exciting. I'm playing my silly question card. I've got three quick questions before we hit the hot buttons, which is more short, quick questions. First question, my silly question, when I grew up, I believe that my fingerprints were unique.

And it turns out, they're not. Are Iris is unique? That's the first question. Yes or no, you can feed to expand on it, if you like. They're the most unique. And so, you know, it's like any sort of like modeling, and there's always innovation in that modeling. So we invest a lot in innovating a modeling. and so the modeling is actually what makes things more scalable.

But from a physics perspective, DNA iris is where the most data is. The more data you have, the further it goes. Now, the problem with fingerprints is also that people that work with their hands, I didn't notice, but people that work with their hands, you can actually rub off your fingerprints. You don't even, you know, and so. Guitar players. Yeah, exactly.

Like, you've got guitar players, like farmers. And then also, interestingly enough, like, there's, if you think about, like, hey, we want to get your fingerprints. Like, nowadays, when you cross a border, they'll do all these pictures of your fingers. But if you have a self-served setup, you don't know what finger someone's putting on that thing. So it's really even harder. It's harder to assess uniqueness because the same human can come in and just put fingers in different orders and the system wouldn't know, right? And so I think that this challenge is when it comes to usability that also compound the uniqueness aspect of things.

So just things to think about when you're thinking about different methods of proving someone's human. Okay. The second one was you mentioned that there was different universities controlling the nodes. You said three or four, I think. People listen. to thinking on paper, some of them decentralization maxis, they will pick up on that. Is that enough? What's on the roadmap, how many do you want to have? How does that affect the security of your data on World ID? Yeah. So basically, we're already at a multi-party compute system, so the data is very secure. There's no magic number to the number of nodes. There's a lot of considerations there when you think about the data centers. If you just think about the data centers

as an analogy, there's a lot of things to consider there around speeds and redundancy and all that stuff. But yeah, so we have three nodes live today. You know, we're planning on on having more, but there isn't a magic number. But our multi-party compute system is, you know, already has like million, is already being used across the entire network. And from what I know, it's the largest implementation lies in production today. Awesome. Guess what timing is, Mark?

Is it? Hotburn's time by chance. Believe it is. Aj, are you ready for this? Feel free to take a quick stretch. Mark's got five questions for you. We're going to put a timer. 30 seconds.

Let's see if we can hit it. 30 seconds, five, one-word answers. Thinking fast, thinking slow. This is Daniel Kahneman's thinking fast. There are only right answers. Awesome goes. Apple or Android. Up to Android.

AGI, humanity's greatest threat or greatest opportunity? Opportunity. 1984 or Blade Runner. What is one book everyone should read? Read right on. Jazz or classical? Classical. And you've won a bonus question.

What would you ask Elon Musk if you met him? How do you do it? When you read? You and you won. hot buttons, RJ, thank you for participating. Amazing. Those were great. Let's dive into use case and then we want to run into our news segment for new listeners. We have a new segment. We look at what's happening in the news and tie it to the discussion. So, Aj, talk to us about how it's being used today, where you see it going, what you're most excited about from a use case perspective.

Building a network of humans. So where do they use, where do they use their unique humanity, right? What's really interesting is that there's actually a lot of segments and And problems that we have today that were this, we're proof of humanity is really, really resonating. We talk to certain creators or organizers of events and they're like, hey, you know, oh, cool, maybe we can use this to prevent scalping, right? You can actually say, hey, a unique human can only buy two tickets. Or are you the human that actually bought the ticket that's showing up the venue? I don't know if you've been to a concert where they asked you to show the credit card you used to buy the ticket. It's like really weird.

That is such a great use case. In England, I mean, EASIS recently, they launched their new tour and the tickets for Allscout by Box. Glastonbury every year, the same thing happens. Every concert you want to go to, you have three seconds. Like there's a problem, you know, you exclusive shoe drop. One person buys all 500 shoes and then sells them at a markup, right? So there's those types of examples. There's also some of these interesting free trial examples where like,

hey, I just get another email address and I can get another free coupon or whatever. And so that really is with like, there's ability to go with market and also incentivize users to convert and et cetera. So that's these business use cases that are interesting. Like I said, right? Good users pay with friction. So there's a lot of interest in, you can reuse World ID for authenticating or verifying accounts.

And that kind of gives us a certainty of a signal. The one that I think that's really exciting is actually gaming. My son's always complaining that a bot beat him. And so initially, I'd be like, are you sure a bot beat you? Or are you just not good at the game? But then the more you look into it, I haven't played a multiplayer game in a long time, at least not as much as I did when I was younger. but like it's a real problem where you want to have a game tournament or you want to you want to play poker sorry not poker but chess there's all these examples of AI or bots in chess and so here like gaming platforms can use the world idea to say hey this account is actually owned by unique human

they can ensure that they unique humanly has one account so you can't really gain that part of it and then they can have a real human gaming and so one of the things that we're announcing a GDC the game developers conference in San Francisco is a really exciting partnership with Razor Razor is one of the largest gaming companies in the world

they have a 300 million users and they're super super I mean this value prop definitely resonates with them and we learned a lot about this problem from them what you'll be able to do is verify your Razor account using World ID and then log in to the games that that support their login, which is, there's a bunch of new titles and a bunch of titles already. And then that game then knows that there's unique humans in that game. So, um, so if you think about the sports, online esports, you pretty, I think we're going to start seeing like human only tournaments would be something

that would be really awesome and just bring certainty and trust into that entire ecosystem. And as you guys know, the gaming industry is this massive. And so there's a huge opportunity there. Other, other examples that we're things that we're working on or that that's already live. So in World App, which is a consumer app that Tools for Humanity builds. So if you think about Tuesday humanity as a first developer on the World Network, we have a consumer app and there's a mini app platform in the app. And we have, we already have 100 developers there. So a million users a day are going in and doing things with mini apps. And you can do, you can play games. But then there's interesting things that where users are, you can buy ESIN, so you can donate through that app. There's a polling app. So

to your earlier point about surveys or votes, simple voting mechanisms where I can ask a simple question and I know every answer is from a unique human, so it's not being shifted in any way by bots, right? So there's a bunch of these sort of like, they're real use cases for real humans, but as a product person, they're kind of toothbrush tests, right, where you're sort of putting stuff out there and seeing what people use and that kind of gives you an idea of where, what other segments or markets that World ID could be relevant to today. And what's interesting is actually the more when you talk for a lot of business people, a lot of them don't know that,

that, that, that, however are you thinking about AI as a something, as a problem, as a, as a challenge that need to overcome, especially when it comes to how users are interacting with their platform. And so we also spend a lot of time educating the market on the value of proof of humanity based on use cases today, but then also why it's important to be able to distinguish humans versus other things online because it's going to become a more,

more important. And it's not a problem, it's an opportunity. But if we can think about that and build systems and tools based on that, then we can actually really maximize, put the human in the middle and maximize the experience because ultimately that's how commerce is going to be done in the future. Brilliant. And one quick thing I want to highlight too, you guys have a very cool ecosystem approach where you're not just focused on this one. I mean, you're hyper focused on this one Lego block to get people, you know, proof of human. But you're showing where this thing's going

by little experiments and tests, like you said. And tests will become new products, will become other products and it's really fascinating how you guys have it set up. Could we double click on Razor just for a minute because my kids aren't big gamers. Jamie, are you familiar with Razor? What kind of games do they have? And is the World ID news mean that it's going to be incorporated one game, lots of games, all the games, how's it going to work? If you go to Razor.com, they sell like a lot of different things to enable gamers.

It's actually really immersive. They sell keyboards of mice and headphones and it's like very cool brand and LEDs and you can control everything about your gaining experience like their entire environment. And then they also have these interesting like haptic products where like the games that integrate the Razor ID can then when you're when you're playing a game, you'll feel things about the game, right? So it's their goal is to how to make this entire immersive experience. And so the games that integrate Razor ID, right?

So think about when you log in with Razor underneath that account is verified by World Eddie. That's the way to think about it. It's that Lego Block. Lego block, one is I'm a unique human and I own a Razor account. And so their Razor account is then used to log into games. And so whatever game support Razor ID today and there's a lot of them would, Razor would be able to share that this account is a verified game in. That's a huge experiment then. This isn't a small experiment. This is across the whole Razor ecosystem. If you're using your Razor account to log into a game. Yeah, it'll, I mean, it'll take time. But that's that that's what we're starting with, right? We've just announced it. We're just getting started. There's like an initial, you know, user, the user, you know, user bootstrapping and then we'll kind of scale from there, right? We're excited to partner and learn together and make this, you know, verified human being.

Okay, Jeremy, are you happy with your kids having their iris scanned to have a razor account? Does that is, I'll show you confirm that. So to have it, you would have to have your iris scanned. You would, but to participate in the World Network, you have been in a adult. That's right. Okay. Sorry kids. Good questions. That's awesome. And what an amazing, tangible, applicable, pertinent use case of which you guys are building. You know, a lot of people are in idea phase. A lot of people are in test phase. A lot of people are in building something cool. People are jumping in, not sure what to do with it yet. You're down the road, man. This is, this is very exciting, very exciting. It's time for the news, guys. We talk a lot about AI. We said AGI a couple times today. In the news, according to TechCrunch, Thomas Wolfe, the chief science. officer of hugging faces. I had an interesting quote about AI and kind of disappointed with the

progress and basically saying, hey, AI right now is just yes men on servers, quote, as opposed to when are we going to create the next Einstein that's able to synthesize these amazing disparate things together to create something novel? What do you think about that when you hear that as AI today is just yes men on servers? I think that's true probably in a lot of, if your expectation is that you're going to have this sort of sentient thing that's going to, you know, be Einstein. But AI's already doing a lot of amazing things from productivity to, you know, I think I was reading about like how scientists are using it to, to analyze a ton of data, like, I don't know what the right sci-fi thing is, but it's like augmenting humans, which I think is really awesome. And maybe eventually it'll be smart enough or

or sentient or whatever, whatever the right word is. But yeah, it's not that today. But I think event we're on that path, which is what everyone's kind of curious about and probably afraid of if you've watched enough dystopian movies. As much as I as progressed, it's still very early. It's only been two, two and a half years or something. You made a great statement earlier where you actually answered the question in the episode earlier, today is the dumbest AI will be. That's a really good way to think about it. I'm going to push back against the Wolfman as well because I think that, A, if he's right, I don't think that matters when AI is being used so much in the West

that even if it is just a yes man on the internet, people are instilling a massive amount of trust and time in it, and it is percolating around the internet. So if it's true, we still need systems to protect ourselves and prove if we are human or not. And if he's true, and in 10 years, 20 years, 3 years, it changes from a yes man to Einstein, what are we waiting for?

Why is he going on TechCrunch and saying, know it doesn't matter. It's just a yes band as you were almost. It seems very, I don't just say, naive from one of the companies that are at the forefront of AI, but it seems almost like clickbait. The way I think about it is like, I think there's going to be people with different opinions all the time. You think back to like the early days of the internet when I forget, you know, those people were out like, I assume you nothing. Or when the iPhone came out, like, I think so many people were like, ah, it's dumb. Like how you can't live without a keyboard. Nobody knows. But the thing is that the really interesting thing about AI today is that you don't need to be a super genius like Einstein to do amazing things because you have a super genius, a sidekick.

If you're Batman and AI's Robin, that's cool. I don't need AI to be Superman for me. If AI could make me Batman, I'm happy being Batman. And I also think that this idea of super intelligence is still an idea. We haven't proven it. And so the other, the counter questions, what if we never get there? Does that mean we don't even, we don't do anything? I think there's a lot of ways to think about it, but the fact that AI augments us, just like our smartphone augments us, I think is really awesome.

And I think it should be accessible to everyone because you don't know which kid out there that either has access or doesn't have access to internet and all these tools, which one's actually with AI maybe going to be more accomplished than someone like Einstein or whoever else. We don't know that because raw intelligence is not really going to be a thing because you have a psychic that has that. Great, great thought. And listeners, if you want to dig back into that more, we unpack Packing McCormack's most human wins paper, which talks about humans getting to a higher level of abstraction and using AI to create, to build with new Lego blocks. We use the Lego Block reference a lot. Hey, listeners, every show we like to tie and thread to the other show because knowledge and wisdom is from connecting the dots between very different and disparate things. So we pull together the shows through our carryover question mark as we wrap up this wonderful. conversation with Ajay. What was our carryover question from Diego from last week? Diego Borgo Web3 Marketing Big Brand Specialist. Are you ready for this one, Arja? How does World ID ensure that the orb and all of your technology remains decentralized,

voluntary and resistant to misuse? Two questions. If our aim is to build an IDA network that every human, that is basically represented by every human, our goal is to open source and decentralized that. So that's very much a part of our mission. Our mission statements accelerate every human, but decentralization is like a core pillar of that. And World ID as an open protocol, right, allows anyone to build on top of it. Right. So if we think about like opt in, you as a human have a choice to go get a World Eddie, I personally believe that it's a human right and you should have access to it. We don't have to, right? So nothing's forcing you to do that. Then once you have a World ID, you have complete choice

into which systems you access and no one knows anything about you unless you tell them, right? And so whether you're private or your private company or a nonprofit or a public government, you're building on top of these primitives, but World Eddie itself does not restrict anything. It's an open protocol that is enabling access for real humans and the ability for humans to identify themselves versus other things online. It can be used for benefits. It can be used for exclusive merch. It can be used for a lot of things where you want to provide a service to a unique human. also can include government benefits. And, you know, we've seen a lot of in the U.S., like fraud when it comes to, you know, unemployment during COVID. Imagine, like, if you had World ID, if every

American had a World ID, that that could be a part of the solution to make sure that, hey, that this unique human got a payment this month and not too. Those are the types of things that I think World ID can be used for and then, but not to restrict access or not to do anything like that, at least from the protocol. That's not something that is possible. Wonderful. Got it. Ajie, thanks so much for joining us today. This has been a fascinating conversation. I love what you're doing. Would love to stay in touch as you guys keep building because things are going to evolve out of this

that are going to turn into things we couldn't have even thought of. Where can people find more about you and your work and World ID? Yeah, so tools for humanity.com, ahold.org. There's Twitter handles. There's a YouTube channel, which I actually recommend to all of my friends. There's a lot of great content there,

especially if you really want to go deep on some of the technologies. But at world.org, you can see the stats of the networks. It's all there on the homepage. And then there's just a lot of interesting documents and develop API as if you're interested in playing with it. And then, yeah, you can find me on social media as well. Sounds great. Thanks for joining us. Listeners, stay tuned for backstage with Mark and Jeremy. Mark continue to, yeah, go ahead. Follow over question. Good catch. A question for our next guest, Aj. Who's our next guest, Jeremy? That's a great question. We're going to Hollywood next week.

Awesome. We are going to the foothills of Hollywood to speak to David Bianchi, NFT, poet, Web3, movie, producer, and actor. Yes, I've seen him. He's a very intense actor. I actually really like him. Question I have for him, which is actually something that I've been thinking about, because I'm a huge fan of sci-fi growing up.

When I think about what my mindset was when I, you know, biked over the theater and watched Terminator 2 or Saw Matrix or any of these things that really kind of blew my mind when, Most of my friends didn't understand what the hell's going on. When I watch his stuff, what I realize is, like, a lot of the fears or maybe, like, kind of, like, uncomfortableness that people feel about AI is really from, like, storytelling. Obviously, like, the more exciting you can make it, the more people want to watch it, right? So the dystopian stuff is always awesome interesting.

But I guess, like, for him, my question would be, like, what do you see as the intersection between, you know, humanness, art, and blockchain, and how he thinks about storytelling around that, right? How do you balance the positives of like, you know, storytelling can be like very positive and help people understand exactly what's, you know, what the value of all these things are and what the role of humanness and art is in the future? Or for entertainment value, there's always going to be the action movie or the sci-fi movie aspect of things. And so I'd love to understand his view on the intersection between, you know, humanness

or humanity art and blockchain. But he's kind of across all that stuff. I'm sure he does. That'll be a very interesting question. Can't wait to ask him. Well, Adj, thanks so much for joining us today and stay in touch and keep up the good work over there. We're excited to see where it all goes. We'll do.

Thank you very much. Thank you, Aj. All right. Mr. Fielding, we are backstage after talking with Ajay from World ID. What are your immediate takeaways? What did you come out of here unexpectedly learning from him today? Other than a new Hotburn's question, which superhero are you? which I'll now be asking you.

Very happy because I think a lot of people who are in blockchain know about World ID. A lot of people have heard about the orb and it's very difficult to write a news article or to report on a company who are using an orb to scan people's irises. Because we're so programmed as a society to hear I scan orb, those kind of words there. They're trigger words, aren't they? And they're just going to have a real powerful impact. And to hear the truth was very refreshing. Agreed.

Agreed. Where I go with this is I start thinking, I start thinking, I just mentioned this carryover question was how does stories affect all of this stuff and generate fear? When you hear the orb and you hear eyeball scan, and it's not an eyeball scan, it's an iris scan. The thing that opened my mind up to it a little bit more is the fact that the scan goes directly to your phone. And the scan doesn't go to some crazy database and live there with your name on it that can be taken and manipulated and used for other purposes.

It comes back to you and you automate the permissions for how it goes. So after today, I'm actually more inclined to go get my ira scanned and try to figure out how to use this. And I also started thinking about how the adoption of this one Lego blocks. We've got this Lego block. You know, someone's raised their hand and they've said, hey, I want to fix a problem that's in the world. hopefully what I'm doing to fix the problem will be adopted by others because we're doing it authentically. We're doing it with trust. We're doing it with certainty.

I love the use of this word certainty in it as well. But it's going to be really interesting to see who adopts this as that layer of trust and identity across platforms, whether it's the old existing platforms that are out there right now, whether it's largely brand new platforms. What did you think about that? This is more a day two kind of conversation. You cemented the word I was looking for when you first came. Backstage, VIP access. Relieved was a word. I think that I felt like listening to what they're trying to build

and how they're trying to build it. I felt a sense of relief that what I'd heard about World ID didn't reflect what they're actually trying to do. The word was release. And I agree with you. But unfortunately, I live up a mountain and I don't think that they're coming here. But you live in Atlanta, so maybe you can have yours, Scandon.

You can let us know. what you do with that data. To your question, I want to take it to step 20, because I'm still vibrating from reading Nexus. Still shaking that off too. Holy. Yeah, book clubs. We've read Nexus. Still shaking off. Proof of humanity. At what point do we need proof of AI? When will you need to prove which AI is which, which of your AIs are yours? And if their AIs are interacting with AIs, how do the AIs prove that the AI that they're interacting with is the AI that it says it is? So this would be the turtles all the way down discussion. We talk about this a lot, Mark, and the idea of a chain of custody, right?

And this is more agentic AI. So if I spin up an agent to handle things on my behalf, if you spin up an agent to handle things on your behalf, there's got to be something in there that, you know, that ties. me to the creator of the agent, especially if that agent is being released into the world to operate on my behalf. There has to be proof of humanity there, I would think, unless we want to have some more Harari moments where we move towards what was the pendulum? It was the the doom utopia spectrum. I think it is a perfect application of having a little tag in those agents that can tell somebody, Not necessarily, hey, this agent is Jeremy's agent.

This agent was spun up by human. Here is some data that I'm releasing to you to understand who that human is or maybe up to me what aspects of my identity I reveal. You were listening, Jeremy. Wow. Thank you. All right. Well, there you have it. Backstage with Mark and Jeremy post amazing discussion with Ajay from World ID.

What do we have coming up? What are we excited about? Where can people go to see more about us, hear more about us? Thinking on paper, x, x, y, Z. we've had a very event for a few weeks. We've been speaking to IonCube, been speaking to StarCloud about data centres in space.

We have, as mentioned in the show, Hollywood actor David Bianchi. We have, of course, Kevin Kelly in a couple of weeks, the famous Kevin Kelly. And in Book Club, we have a new book, which we haven't started reading.

You have another two weeks. If you want to get a copy and come and read this book, Consciousness, Life, computers and human natures with us, then wherever you're listening to this, via comment, we'll send you all the info that you need. Tell you what, book club participants from the past,

if you thought Carlo Rivelli's, the order of time melted your face, this will be four or five steps deeper. Come join us, we'll unpack it together. Be curious. Stay disruptive. Keep thinking on paper. See you next week.