What Happens When the Internet Can't Tell Humans From AI? (ft. Alex Blania and Ben Horowitz)

Transcript

How do you prove somebody is human? It is a surprisingly hard problem. I think that people are going to start getting accused of being bots. What we currently see is less than 1% of what it will look like in probably [music] a year or two. >> The idea that AGI will lead to some very fundamental shift seems obvious. [music] Like >> AI's are really good at programming humans. Much better than humans are at programming AI's. Absolutely. AI [snorts] will be able to have a GitHub account and will be able to post and also attest to five other AI's that these are in fact humans and even though they're not. Also, if you don't take it seriously now >> [music] >> Alex, welcome to the podcast. Great to have you. Thanks for having me.

So, proof of human is having a moment right now. Won't you first give a background for people who are unfamiliar? What is the moment that's happening and how did we get here? Yeah, and what is proof proof of human? Proof of human, as the name suggests is, you know, do you know if you interact with a human or like something else on the internet? And I actually think the the kinds of questions that we are asking is are you interacting interacting with a human? Uh, an agent on behalf of a human or just an agent. Like I think these are like roughly the three the three areas that we want to split apart.

Well, and and describe a little bit the difference between just an agent and an agent acting on behalf of a human. How do you see that distinction? Yeah, so um, quickly explaining just the term proof of human and I think what is hard about it and then it will will explain how it fits into into an agent on behalf of a human. So um what proof of human really means is that uh, you know, every individual that interacts on a platform has only one ideally one account or, you know, a limited number of accounts and stays the owner of that account.

Like that that's that's kind of the property that you're looking for. So, like you're looking for a uh, initial verification uh, that ideally should be, you know, something like anonymous or very extremely privacy preserving and then on ongoing authentication that the same person remains in control of that account. Um, and then there is like some secondary properties that I think are good to have. But that actually tells you that the really hard thing is is uniqueness. Like like what what is happening on a platform like Twitter right now is that there's all these accounts you know, all these all these bots that are in the replies um, that, you know, there's probably one human sitting somewhere and and sending out 10,000 or like 100,000s of AI's and there's this catch-up game where like

uh, you know, Twitter and X are trying to just find them and block probably millions a day of these. Which is what, like a a 100th of the of the bots? That that's right. That's how it feels like. Um, and then agent on behalf of human, I think like how it will look like is uh, you know, I as a like I think all of us will have agents. Maybe it's unclear how it will look like. Is this going to be one or there multiple ones maybe with different tasks and different even even types of characters.

Um, and I think it will then come down to I, you know, I approve a certain action of my agent. I give him certain rights. So, like act on my behalf. >> Okay. Post post to my X account. Post to my Instagram. For example. >> But it's my Instagram and I'm a unique human that owns that. That's right. You know, that X or Instagram could could decide that if that's actually something they want as a platform. >> Right. Uh, but that's how you could do it. That makes sense. Um, and so >> [gasps] >> how do you How do you prove somebody is human? It is it is a surprisingly hard problem.

Yeah. So, you know, it's And those agents are very very clever. >> [laughter] >> It's uh, you know, it's funny. We started this company now a couple of years ago be before ChatGPT and before all of that. But we we kind of took that as an assumption that eventually we will have AI's that you know, both pass the Turing test, so they can just claim to be a human you will not be able to tell that anymore on the internet. And also that they would be, you know, highly agentic and just like run around doing their own thing. And so that makes it really really hard because uh, back then when we started the company there were like roughly three big ideas that people were interested in.

One was this idea of uh, web of trust or like related idea. So, this idea that you you look how someone behaves on the internet or did behave in the past. So, like usually a combination of you have these certain number of accounts uh, that you you know, you own since a couple of years and then you post regularly or you comment regularly to GitHub. Like these were the kind the kinds of things that people are using. And then let's say all three of us have them and then I attest also that, you know, I know you in the real world and I attest to you that I know you in the real world and that's how you would build a certain graph.

And that was like a very hot idea back then for this. Um, but we disregarded it basically because we assumed that, you know, eventually everything that is just digital and AI will be able to do as well. Like Yeah, no. We're we're there. Yeah, exactly. So, [clears throat] then AI will be able to have a GitHub account and will be able to post and own an account and like also attest to five other AI's that these are in fact humans and even though they're not. So, uh, so you know, there was there was area number one. Area number two was to just, you know, uh, use government IDs for everything which uh, we just also immediately disregarded for a couple of reasons. One is you know uh, I think

you know, it's strictly better if the government would not control such an infrastructure in terms of free speech and actually breaking that apart. But then also Right, you lose anonymity. Instantly right? >> You could hypothetically set up a system that maybe preserves it, but it's very hard to do. And then a second thing is also um you know, the government ID identity system is just not built for that. And and and what is so hard about this problem and it's going to be a global problem. And so it doesn't really matter if, you know, one government maybe has the perfect infrastructure. For example, Singapore is like an example of a of a of a of a government that has, you know, perfect infrastructure all around. But that barely doesn't doesn't matter

because, you know for example, I don't know, Meta has a global product with 3 billion users and there's a lot of other countries. Yeah, Singapore is what, like 2 million people or a million people? Exactly. So, we want to lock everyone else out. So, uh, so yeah. And and then there's a long list of other things um, why we disregarded that basically immediately. And and then last one is biometrics which actually, you know, immediately gives us this big reaction. It's it's like um and it even went further because uh, what is so hard about this problem as I mentioned the beginning is uniqueness.

Mhm. And so just like in very simple words, how you can describe the problem is um Well, first of all, for example, what does Face ID do? Face ID checks that I'm the same person again using my phone. Mhm. And uh, so it's a one-to-one authentication. So, there's an embedding stored on my phone. It takes a picture of my face, creates a new picture, compares it to the previous one and if that is close enough, I can use my I can use my phone. But uh so that's a one-to-one, you know, one embedding to one new embedding.

To solve the proof of human problem, you will need to distinguish one new individual from all previous individuals. Mhm. Uh, you need to make sure that, you know, Ben is trying to sign up and Ben did not sign up before. Yeah. Um, and then suddenly it goes from one-to-one to one-to-n. And n is the n is the size of your network essentially that you're trying to prove that to. Right. And then you can just do the math and you can calculate how much mathematical entropy, like how much information just information theoretically do you need to to prove that? And it turns out that's pretty high number. Cuz it's it's an exponential problem.

>> Right. And so then you can just do the math and you find out that, you know, things like face or or, you know, even fingerprints or something doesn't work. Like that then you would basically hit a wall after tens of millions of users. >> Yeah. And so then you end up with uh, you know, something like iris which is the muscle of your of your eye. That actually has enough entropy that it's unique. That is unique. That is unique enough. And how do you also then solve the uh you know, one thing that biometrics have been subject to historically is just replay attacks.

Where okay, I may I may not have your eyeball, but I've got enough information that I can run a replay attack on you. Um so I there's there's now actually, you know I get it is important I think to split up the problem and verification which is essentially in, you know, old terms, it's like you're getting your passport. Right. And then authentication which is you showing your passport uh, constantly for certain kinds of things. And on the, you know, on the verification piece um that's, you know, we we we've went down if you know World, you know that we've built this thing called an orb. So you know, it's it's doing a lot of things to prevent these kinds of attacks. So, it's for example, it has

multiple sensors in the, you know, electromagnetic spectrum to just make sure that you cannot show a display to it and then and it would recognize that. Um, so I think on that side we've, you know, we we've got it handled. On the on the consumer side, like you know, that should then re-authenticate it turns out to be much harder because uh, you would need to trust the phone in some sense. Uh, because what we actually do in that moment is when you verify with an orb, we not only do we check uh, your uniqueness in a fully anonymous and privacy preserving way and we should talk about that. But also we send to your phone a signed face image that you then can later use to re-authenticate against it.

Right. Um and you know, with a new iPhone, you can have meaningful amount of trust against that, but with old Android phones, basically no. And so Oh, yeah, yeah, yeah. Yeah, you know, because like you can just uh you can just show a a deep fake essentially either through a display or just directly injected in the camera stream. So, um that's a problem. And so, it's going to be a mix of uh you know, if you have a new enough, let's say iPhone or a general phone, um then you can just re-authenticate against that uh picture you took on verification. Otherwise, you would probably have to even go back to an orb somewhat frequently. Um like let's say a couple times a year if you just I see.

Right, to re-authenticate. Yeah, that's right. Interesting. And then, you know, one of the things to one of the kind of incorrect criticisms of the approach early was, "Oh my god, they've got my eyeball. Um you know, now they're you know, they they somehow have uh access to my privacy and they're going to you know, do all these things to me and and that's my access and then they can they uh Worldcoin can um impersonate me and all these kinds of things, but that's not the case and um so that was also like a non-trivial engineering problem.

>> it was very much non-trivial. Um it's actually I think one point on iris that I think people don't appreciate enough and that's a bet we took back then, but it was essentially that iris will turn out to be supernormal as a as a modality just because I think we'll all wear um AR and VR systems that do that. You know, Apple already does it. Uh Yep. Iris ID in in the Vision Pro. So, I think it's So, maybe that's a general point. I think it's going it's going to be become something that people use across many different devices and uh will normalize in that sense. But I think on the on the privacy piece, um that took us a lot of time because like when when we when we decided back then that you know, with our assumptions,

you know, which was 6 years ago, that we will need a custom hardware device for biometrics, it was actually quite scary um you know, to to come to that conclusion because like Yeah, that's an expensive conclusion. >> It's it's like it's like very expensive and then just having this idea that you would need to distribute them all over the world. Like that that just assumes that you would be able to like somehow bring up billions of dollars and to like a massive effort to to just resolve the world. Um but then also the privacy challenge of like how could you build such a system that has all the all the requirements that we care about. And the the two main high-level, you know, ideas on how to solve it were uh multi-party computation and zero-knowledge proofs. Mhm. And so, um

to again, what is different to Face ID because Face ID actually is it you know, can be very private just because the embedding is stored on the phone. It doesn't have to leave the phone ever just because it's just you against you in the past. But to check uniqueness, uh you need to check against all previous people. So, it Right. So, so something needs to leave. Yeah. Uh you know, something needs to leave something and be compared to someone else uh and that's and that's a much harder uh challenge. And >> [snorts] >> um how we approach that is through multi-party computation. And so, that essentially means that so, in our case, when you uh verify with an orb, you know, we we take all these

pictures, uh they get computed on the device and uh then they actually get split up in multiple pieces. So, we for example, we take a picture of your iris, we calculate an iris code, um then we break that iris code in multiple pieces and send send it to multiple computers such that um there is no central database in some sort. So, no one actually has the information about you. Right. And then you do some clever tricks of how these different parties need to come together to do a computation that still leaves the pieces apart Right, right, right. in such a way that >> Right, nobody has the whole thing.

>> Yeah, so no one has the whole thing and also during the computation, no one has the whole thing. >> Yeah. But they do some, you know, some clever interactions to come to the conclusion >> like a zero-knowledge proof kind of technique. >> Uh it's I mean, it it's very different, but I think in terms of the properties it achieves, it's somewhat similar. Where like you no one knows anything about you, but you can actually together make a statement about you. Right. And so, you know, you send it to this multi-party computation and what comes back is, yes, that individual is unique.

And then the second thing we do is we we separate all of this um from you with a zero-knowledge proof. So, meaning you have the secret on your phone, but no one else has it, no server has it, we don't have it. Um and then you can later go back to this multi-party uh computation and say like, "Hey, I have a secret that is part of that computation and I am in fact unique." Uh and then you can prove that to a platform. You could go to the social network and prove that you're a unique user to the social platform without us knowing anything about you or the social network knowing anything about you. And so, it's just like very counterintuitive property that you there is like even though it uses

biometrics, you you know, you preserve anonymity and and uh extreme levels of privacy, which I think is super cool. You know, social media is one kind of vector of uh you know, things that were annoying and are now becoming overwhelming in terms of uh just bots and particularly with psyops, propaganda, all these kinds of things. What are some of the other um you know, uses of bots that are going to be kind of impossible to live with if we don't get to proof of human in the future?

Yeah, actually I think the the simple model I have for it is every moment on the internet uh that is primarily about humans interacting with each other. You know, or or even indirectly interacting with each other. So, uh you know, you can you can start with simple ones like dating. You know, it really matters. >> [laughter] >> It is what if the other side is in fact a person. Yeah well they I've got bad news for listeners. >> Well, I I I >> [laughter] >> and the person who you expected to be. Yeah, yeah. Yeah, exactly. Yeah, we had problems even before the catfish thing.

Yeah exactly. Yeah, yeah. So, that that that's that's an obvious one um and and so, for example, Tinder is already using it for that reason. Um I I think >> And what what's the uh the Tinder use case? So, So, we started we we started in Japan uh and >> [clears throat] >> like as as a test as a test market and it's essentially exactly what we just discussed that is um if you verified with an orb, you get a little badge that, you know, signals to other people that you are in fact a human. So, it has a high level of verification. Um and then also um I I don't think that's live yet, but what will come next is that you're actually the person you claim to be. So, meaning you have a World ID that is associated to the kind

of profile pictures that you use. Um so, you just run a quick check that uh this is all correct. Um and so, you know, you then know you're not interacting with bot, but also you you know, you interact with a fully authentic profile. Yeah. Another fun one because I think it's somewhat counterintuitive, but I think it will be video conferencing. Mhm. Uh because you know, you already have deep fakes. Yeah, just to just I don't feel like going to this video conference. Just put my deep fake up. >> [laughter] >> Yeah, and actually you um you raised it to me first and that's why we started building a product for it because, you know, it will actually start with very high-value users. You know, like for example, people, you know, like yourself that maybe manage a fund and you know, sometimes calls actually could be very

high value if it's about wiring money or Oh, yeah, yeah. Well, so so so so so somebody can uh be me and say, "Eric, can you please wire this Nigerian prince $400 million." Right, exactly. Um >> be good to know, yeah. Yeah. Yeah, like, you know, that that's still slightly hypothetical because these these things are not fully real time and you can somehow >> They're very close. They're very close. And so, I think you know, in a year from now, it's just going to be a full commodity and it's going to be super photorealistic and absolutely real time and you will just not know anything anymore on these video calls. And and so, I think that's another one. Uh I think another one then will be just I think it's fun, but it's it's going to be

gaming. Uh you know, because Oh, yeah, yeah. because like gamers really care. Oh, yeah, that I'm playing a AI. >> [laughter] >> Holy cow, that's frustrating. Yeah. Especially if we bet money. Yeah, exactly. And you you you lose money, you train multiple hours a day to get like really good at this thing and then suddenly you get, you know, Yeah. you you get destroyed by an AI that is just superhuman in every dimension. Um funny enough, I was like I don't know what you think about this, but because I don't have a good mental model about it, but even the the whole model for video platforms, I think, is about to break because there's a couple dimensions to the

problem, but one if if if the if the creation of content is is becoming super scalable. Like for example, I I heard about this one guy that uh created, I think like it was like on the order of a hundred videos a day on YouTube and made tens of thousands of dollars a month. All of them were fully AI generated. Yeah. Um and people just fell for it. So, now the question is is that actually something that YouTube wants to monetize that way? Yeah, like is that Yeah, well, it's it's interesting, right? They fell for it. Um But maybe they liked it. Yeah, yeah, yeah.

That could be, but it it would sure be nice to know like, okay, this is a human video or this is an AI video. Um Actually, my thesis about this is like something something along the lines of I think there's categories of content that are clearly just fictional. Yeah. Like movies are that, you know? It's like you you don't care that there's any connection to reality. It's just a fully fictional story. But now if you think about something like TikTok or you know, all all these kind of things like people actually really care about them mostly because there is some connection to reality.

Yeah. >> Yeah. Well, there's reality and there's connection to human, right? So it you can create a pretty good like you can take a scientific paper and give it to Gemini and say make this into a podcast and you know, it'll be like a pretty entertaining podcast and it will be reality in that it came from, you know, some real thing. But you would like to know that. You would like to know that. Yeah. [laughter and clears throat] I would like to know that. >> it continues as an advertiser you would like to know that a human watch it. Yeah. Or did an AI watch it?

Yes. [laughter] Yes. Right, right. Well, right. That's the other thing is I created 100 AI videos. I had a million AIs watch it. >> [laughter] >> And then I made a lot of money off of YouTube. Exactly. And actually saw that video today of of a YouTube farm or like they had just like thousands of phones that just watch videos all day for a reason. Yeah, yeah. And then like that's got zero value to the YouTube advertisers. And so that's that's actually a real problem for them. >> Right. Well, the whole sort of the creator economy platforms of the last decade, you know, Substack, Spotify, and all the people who support artists or you know, Patreon etc. creators, YouTubers, they they have a personal relationship with with with these people. It's not just

they like the the art. And so if they all of a sudden found out that they were, you know, bots, that might, you know, they might not want to support them in the same way. Yeah, [laughter] you you might want to give them a a big YouTube tip or Yeah, I think there's a certain subset of people who support, um, you know, want to support actual people and feel like they're having a real relationship. And and the thing that I think like people don't really get is that you know, it should be obvious, but I don't think people really understand the the consequence of that. I think two things. One is that what we currently experience is like a super super tiny thing of what is about to happen. You know, just because like >> Yeah, right. It's a glimpse. It's a glimpse. Like you know, cost of

intelligent intelligence is dropping almost exponentially. Agentic capabilities are increasing, you know, in like some super linear form. So like yeah, what we currently see is less than 1% of what it will look like in probably a year or two. And then so and then second, these things will be actually they will be superhuman in many ways. They will be like perfectly able to understand you and like talk in the right way to you. Yeah. For example, there's this like one paper that I that I think you got to read after. But it was, um it was uh the change my mind subreddit. Mhm. [snorts] Um where the University of Zurich did this thing where they had AIs actually interact with change my mind.

>> Mhm. Yeah. And they were like superhuman in their ability to change it because they they were going back to the profile of the people posting it and were like understanding their political motivation, the way they talk and like and they're just interacting in perfect in the perfect way. Yeah. You know, I just like hit all the buttons and AIs are really good at programming humans. That's much better than humans are at programming AIs. Absolutely. There's no question. And so I think that's going to get quite scary also. But uh I think at least if you know you're being a victim of a psyop >> [laughter] >> then or or or it's a very advanced one done by an AI, that would be extremely useful to understand. Totally. Talk a little bit a bit more about the state of

the product and the business today. Like how many IDs are are out there? When do you give a little bit of an update? Maybe talk about the evolution as well. Well, first of all, it's a multi-sided problem. And I think there's like roughly three that you have to consider. One is uh well, you need platforms to use the technology. Uh then, you know, like things like Reddit or, you know, X or, you know, things like that. Um secondly, you need uh distribution of these devices. And I think the right mental model to to have for it is how many minutes does it take a person to reach such a device on average? And, you know, currently it's if if you would take the global average, it would be a

terrible number. It would be like you know, days or something because many people would need to fly. But uh but you know, how do we get that down to below 15 minutes across the US? And so that's probably roughly around 50,000 devices that you need to deploy. That's like it's not crazy, but it's also not nothing. It's it's you know, it's it's hard to do. And then the last one is how does all of that come together to something that a lot of people really want to use it? And that's a combination of you know, the utility of all the sub platforms essentially. But but all of that layers on top. Like maybe you can use a new Reddit account. Maybe you get like, you know, certain amount of ChatGPT subscription for free or like So I think it's going to be a combination of things, but you need to you need to

land all three at some point at the same time. Which is uh which is hard to do. We are now at 18 million users that are verified. For- 40 million in total in the app. Uh but the biggest thing is because of the past administration, because we use, you know, we use crypto, we we did not really invest in the US for a long time. And um that's not [clears throat] the main shift that we're going through. It's like Yeah. For all of this, the main thing that matters is the US. And ho- ho- hopefully uh we get the Clarity Act passed shortly. That will help us.

>> That would that would be really great. So, um to get clarity on that. Uh-huh. Yeah. So, so the big focus that we that we now are going through right now is to kind of go all in on the US. So I think over the next year, 90% of the of the, you know, effort of the company is just going to go about the US. And how do you get, for example, device distribution up? How how do you eventually have this in every Starbucks? Um so it becomes just, you know, super normal and people just just use it every day. So that's kind of the And then on the platform side, actually we went through a It's um it was a very interesting experience to go through personally because I think um like a couple years ago, universally

people just made fun of us, you know, like just it was like the universal reaction. Uh well, minus and recent and then a couple other people who believed in it, but um Yeah, like in the press like like the amount of fun making of something that This just shows how short-sighted people are. That's right. >> It's like you don't think the bots are coming? What did you think when when we first pitched actually? Because even you must have thought this is crazy. Well, because you had the orb. Like the orb was so wild. Um you know, okay, we're going to scan people's retinas and that's how we're going to know they're human and so forth. And this was I mean, you pitched us Six years ago? Six years ago? Yeah, it was before COVID because you were

there with the orb. Right. Um and you know, AI just hadn't happened yet. >> [laughter] >> And and you know, you could kind of see, but there you know, there there's bots. Um but they were kind of very crude and, you know, compared to what there are now. Um but it uh it seemed inevitable. Um at least at the time, you know, the thing was it was so out of it was so from the future that uh you know, we always worry about, okay, like what's the timing of this and this and that and the other and and so forth.

Um But you know, you were impressive enough and it was going to happen eventually and it was an exciting enough idea that I think all those things kind of got us to go, okay, we're Um [laughter] but but it was not it it was one of the it wasn't obvious that like it was going to work in that time frame. It seemed very in obvious for a long time. And how different was that pitch from what it ended up being? Or talk a little bit It was actually pretty much exactly the same pitch. I think it's the same thing. The device changed. You know, you know, they've they've made it much more economical and convenient, but That's right. It's uh But the initial instinct was was there.

It was basically everybody's going to have to prove their you're either going to have to have some proof that you're human on in cyberspace or like it's going to be a very bad world. I mean, the robots are going to get us. We're done. Right. And then actually the second piece that was like this was the first thing is like it's going to be a that itself is going to be a big deal. But then second of all that, you know, when it's going to become a big deal, we will be able to build one of the most valuable networks as a result of that. Because in a world of AI, having a human network is going to be this incredibly important thing. And uh and so actually, yeah, two things. Like one, you will need to prove of human.

But then second, it will have very strong network effects. And even as the platforms, as you get into the platforms, even as the platforms' largest problem has been bots. I mean, you remember Elon and the you know, he backed out of buying Twitter cuz all the stats were based on bots. They still even knowing that it was hard for them to get all the way to the future in their thinking and go, yeah, we need proof of human. Yeah. Like it's kind of obvious. >> people were like, what does it even mean? You know, like what does proof of human even mean? We can just we can just, you know, And did you have the language When did you come up with the language proof of human?

We had actually we we had proof of personhood for the longest time. It's even here in this on this brief. >> Yeah. Yeah. But then uh at some point we were like, [ __ ] Well, at some point AIs will have personhood, too. So >> [laughter] >> Uh so like that's not going to fly. But they're not going to have retinas for a long time. That's actually Oh, that's coming eventually. It was It was actually really funny. It was like some of the some of the OpenAI people uh that I met were like, man, Alex, was going to This is going to be so dark. Like people will hate you for like not giving personal data to AI's and I was like "Jesus."

>> [laughter] >> Let's Let's call it Proof of Human then. Um >> That's funny. So, that that's how it changed. Um but then actually so then I would say like last year so post then there was like a big shift post ChatGPT. Like people were like that was like the AI suddenly got real to people. And then actually I think And so that's when people started talking to us but still we're not like, you know, like it's a future problem. It's probably a couple years out. Like we don't really care about it. Let's stay in touch. Like it was like the common response and then uh You know, and well but you also you had a couple CEOs that really believed that and were like willing to take the long-term bet um to to give them credit.

But I think the second big shift was actually Claude Bots uh and Mobile Coin recently. Yeah. Just >> because of >> Yeah, that >> [clears throat] >> that kind of means like the the cow is way out of the barn. >> [laughter] >> Yeah, and and so like honestly if you don't take it serious now then I think you just you you should get a different job or something then. You're not >> are you doing? Yeah, they're just not not thinking about problems in the right way. Like it's And so that's that was like the moment when many many people started reaching out and now now it feels like much more of an execution problem. Not not anymore a market risk. Like a market risk or like a thesis problem or like like just a And which is still a big [ __ ] problem. Like how do you how do you how

do you get 50,000 devices out there? How how do you make it cheap enough? How do you make it economic? Like, you know, how do you how do you make all these three things at the same time? It's still a very hard problem. How do you normalize the behavior etc. So people aren't weirded out in a Starbucks or something. Although I I think that's now going to be go to use to I I'm just because I think people will hate the alternative so much. And I think people are going to by the way take a lot more pride in being human uh particularly online because I I think that people are going to start getting accused of being bots.

I mean Totally. it's it's going to get really weird um and without like clear delineation it's it's going to be a mess. Like I don't I don't understand how somebody can think they're going to have a social media platform that doesn't distinguish between humans and bots. Like that seems absurd to me. So it's absurd. I think we will My guess is over the next couple months we will see we will see things like these platforms trying to use things like face biometrics on the phone. Which, you know, I know it will break so it's fine but I think we'll go through that cycle now.

Uh and yeah so we just need to get to scale fast enough to to meet uh the market to what comes after which I think something like the Orb is the only solution. I think currently there's no real competition. I think we'll also see that. I have not seen a competitor yet. Because it's so Because it's so ridiculous. It's so ridiculous and it is so hard to get to in terms of building it and and then there's a massive network effect um Right. which like people are starting 6 years behind you on that but yeah I'm sure they'll come cuz it's it's just such an obvious problem now. What actually do you think about like AI continues

what in your mind are the economic policies that we will need to implement directionally? I think governments do have to figure out how to send citizens money. They're good at taking money from citizens but not the reverse. >> [laughter] >> I mean well just if you go back to COVID the stimulus program like I think 400 billion dollars was stolen. >> [laughter] >> And like that that that that's pretty good. You would have liked to know that you were sending the money to unique humans. I mean if even if not citizens as long [clears throat] as they were unique humans that would have [snorts] been good.

>> Yeah, I mean the social security system for example is a mess in the US. It's a total disaster. It's a total disaster yeah so We're going to have to get to some kind of way to cryptographically strong way to identify who's the citizen >> [laughter] >> of what country. Like like like that's going to be a really bad problem um I think. So otherwise there's no way to even have a democracy. I mean, you know, like the it's pretty crude what they're trying to do with the SAVE Act but it's not completely insane which is how do you even know like the people are voting are actual people or living people or anything?

And we really don't know now. Um we like we genuinely don't know and then if you go to I mean >> [laughter] >> the the the whole mail-in ballot thing like is built for whole very different world, right? That's right. Uh so like I don't think in an AI world where you can have like very high scale impersonation that and then with a broken social security system that like you're going to have the will of the people anymore. Like I I think that's going to be gone pretty fast. So I think we're going to need some kind of you know, cryptographically strong infrastructure on like who's who.

Um and then you know, similarly I think we're going to have to be able to get people money much more efficiently than through these uh this crazy apparatus of social programs that we have. Just cuz like how lossy is and fraudulent is social security or Medicare or any of these things. I mean like the Medicare is so frustrating for people that they shot the CEO of United Healthcare mail. Like and people are happy about that. Like really happy. So like think about how bad a system that is um when you know, and the government spends a lot of money sending you money for your healthcare but they do it in a like super inefficient way.

Um but we have the technology to do that now. So I think that AI is going to make that problem so bad uh cuz the ability to file fraudulent claims and create fake, you know, buy social I mean you can buy social security numbers on the black market. Like for those of you don't know that's an easy thing. That's a real thing. Like that is like everybody's social security number is for sale. Uh and so um you know, like AI is just a way of making that kind of loose black market underground fraud thing just massive and extremely scalable.

I agree, Tom. Yeah, so I I I think you know, Proof of Human is a piece of a very important puzzle where we have to upgrade the entire infrastructure or we're not going to be a democracy anymore. I mean that that's just be my guess. I agree, Tom. Say more You said okay, next year go to market. Let's focus on the on the US. Say more about how how you're thinking about that. Is the incentive for people to do it because they get to use a set of services? Is there some other economic incentive or how do you envision it? Basically a month ago we entered a very different phase as a project where I do believe many of the platforms that we're now integrating with will really, you know, bring a lot of users to our platform and that changes, you know, how you think about it

entirely. Like if you have a if you have a platform with a a billion users um sending users to you then it's really just all about like how do you meet that demand? It's like, you know, and that's that's that's what we're now entering and and so um yeah so I think the response is first. I think you will see and we're already working on it but you will see a lot of really large platforms that you know integrate in the in the near term future. >> [snorts] >> I think that will just set expectations.

I think they will be slow initially because it also should be just as you know, to to get understand the product. It will be focused on certain geographies. Like what we did with Tinder very soon in Japan just to, you know, to uh to test the product and also to just normalize the concept. Uh but that will happen and then secondly which is now becoming like one of the main priorities for me is just how do you get this Orb distribution up? Which is which is, you know, broadly speaking there's a couple different dimensions to that but one is first of all the product needs to work at scale uh you know, without supervision which is turns out to be much harder

than you would think. It you know, every engineering problem at scale turns out to be much more complicated than you would think because, you know, fighting for 1% of improvement in quality is this [ __ ] of you know, all these dependencies that come together. So that's the I think that's like one of the biggest engineering focuses right now. But then second um you need to find places to deploy them at and and the way to think about it is there are large-scale distribution partnerships that could be something like Walmart, you know, or if you you know, if you're very ambitious it could be something like Starbucks.

Um or it it can just be you go to one of, you know, hip coffee shops. And you just you just put it there or, you know, and then it you could go you could eventually even go to the DMV and just put it right there. So that's the problem we're currently trying to trying to puzzle together. Um And you know, it's going to be some some of all of that. I think there's going to be some large-scale distribution partnerships. Many one-off coffee shops. Oh actually one thing that we will uh we will launch soon and the team is going to hate that I'm saying this now but uh it's going to be Orb on demand.

So >> [laughter] >> Orb on demand? Yeah. So So in the Bay Area. Just because actually it's such a it's such a gnarly problem to you know, to get an Orb to truly everyone. You know, it's like to to get that the cap backs is insane. Yeah. So it's actually it's actually much cheaper and easier to just put an orb on a motorbike. >> [laughter] >> And drive it to you. It's It's It's crazy. As As crazy as it sounds. So like in in places like the Bay Area or New York, you will just be able to say like, "Yeah, I want to verify now." Wow. And 50 minutes later there's an orb comes to you to your work and you can you can verify. And uh Did you ever think about uh I don't know this is probably a terrible idea, but um having kind of

different levels like we know you're a unique human or like, "Hey, this guy may be a unique human cuz he's done it on his iPhone and it's not quite the the same, but Yeah, yeah, we have we have that. So actually we um you know, generally we just have to you know, we have the principle of you know, what whatever could be useful for this problem we just build it. And and uh And And so we we have something called face check that that does that. So it uses it uses face uh from the camera. It still uses multi-party computation but we've built the entire system so you're still anonymous. Mhm. Um and you know, it of course reaches way less accuracy.

So uh you know, as a system you will know something along the lines of "Well this is, you know, at least one person cannot create 100 accounts. Maybe it's just 10 or 20." So like it's like a at least it's some measure of rate limiting. Um and I do think just to set a disclaimer, I think with deep fakes and and you know, all this stuff, I think that will fundamentally break. So it's a Mhm. Temp. It's a It's a It's a temporary solution that I think can get us to scale. That's kind of how I think about it. Uh we also actually use government IDs uh similarly where like we we use a just the ones that have an NFC ID chip.

Mhm. Um and we use multi-party computation so you remain anonymous and platforms can choose to use that as well. Uh but no one really did. It's just tomorrow they have this like very negative stigma which I think makes sense. Yeah. Um but yeah, ba- basically whatever could do it Yeah. [laughter] by any means necessary. >> That's right. Yeah. Uh well, thanks so much for coming on the podcast. It's been great. Yeah, thank you. Thank you. That's great. >> Thanks for having me. >> [music] [music]