Episode 03

Dan Schmerin (00:07.152)

This is the Reimagine Podcast, where each week we're diving into how blockchain and Web3 technologies are transforming ownership. We will explore how decentralization is shifting power from big corporations to individuals, making ownership more transparent, more secure, and more accessible.

Join us as we uncover how these innovations are changing the game and paving the way for a fast, fairer, join us, scrap that, join us as we uncover how these innovations are changing the game and paving the way for a fairer, more equitable future.

Oliver Quie (00:40.152)

You

Dan Schmerin (00:50.81)

Welcome back to the podcast. We are thrilled today to have our friend, Oliver Kew, Ollie Kew, as he's known, joining us on the Reimagine Ownership Podcast. Ollie is the co -founder of Innerworks. Look them up at innerworks .me, a pioneering company specializing in intelligent bot detection.

Ali has become a leading figure in combating the misuse of automated systems across digital platforms, an issue which is very important to Yossi and myself. InnerWorks' innovative technology focuses on identifying and mitigating the impact of bots. How many of you have dealt with bots? It's only proliferating and they're coming after each of us.

And Ali's team ensures the authenticity and integrity of online interactions. Through his business, Ali is at the forefront of maintaining trust and authenticity in the digital world, striving to make online environments more equitable and secure. Ali is also a proud member of the Metaversal family. Hang on one sec, guys.

Oliver Quie (02:46.008)

You're still there, you see?

Yossi Hasson (02:46.437)

Yeah, I'm still here. There's just Zach's in our office and Dan can hear him. I don't think I can hear him though. I don't know if you could hear him talking in the background. Okay.

Oliver Quie (02:54.434)

I couldn't hear anything, but probably the mics picked them up too.

Yossi Hasson (03:00.33)

We couldn't hear him Dan, I think you could only hear him. We couldn't hear him when you were talking. Now that you have a nice noise cancelling mic, no one could hear him.

Dan Schmerin (03:00.828)

All right.

Okay, my apologies. I'm gonna pick it back up with...

is that right? Okay. Okay. That's good. I'm to pick it back up and then we'll dive in.

Through Innerworks, is at the forefront of maintaining trust and authenticity in the digital world, striving to make online environments more equitable and secure. We're very proud that Ali is also part of the Metaversal family, and we're super excited to have him on today. Welcome, my friend.

Oliver Quie (03:43.404)

Dan, Yossi, thank you so much for having me on board. And as you know, I strongly believe that Metaversal are spearheading the digital ownership economy and so happy to be playing a small role in your extremely exciting journey. So very happy to share more about what we're building.

Dan Schmerin (04:04.304)

We love when people come on and offer pleasantries to Yossi, specifically early in the day.

Yossi Hasson (04:04.677)

That's great too.

Dan Schmerin (04:10.778)

because that really sets him on this positive trajectory, which pays us dividends well after lunchtime. So maybe we'll jump in. know Yossi has a number of questions that he wants us to run through, but why don't you kick off by giving us a little bit of an overview of your career prior to Innerworks and then what was the genesis for trying to tackle what we all know is a really large, really thorny issue.

and begin to break it down into sixth grade verbiage so that everyone at home, in at least my household, can easily understand why what you're building is so important.

Oliver Quie (04:51.352)

Yeah, of course. So I've been involved in the tech space one way or another since, since 2016, both as a, as a builder and, and a venture investor. And as you know, the genesis of, of inner works was in a remarkably different place. our journey started as a, as a behavioral sciences concept. we believed there was an important question of what's the most accurate.

behavioral profile you could create about a human being. And the original use cases we were interested in were again, a completely different place to where we are now. So we were, I was a young student interested in entrepreneurship. thought if there's anything I have some credibility and understanding around, it's the clunky way that people find jobs. And I was very interested in the fundamentally superficial data sets that are fed into this increasingly

automated industry that's trying to matchmake people with potential job vacancies. And in particular, you know, I remember very early conversations with you, Dan, where, you know, we, we understood the things like resumes, credentials have two different issues. One is a verification issue. How do you know that those claims are true or false? And I know that that's an area that the metaversals, you know, progressively become experts in.

and what a modern version of a verifiable credentials profile looks like with medallion. But the other side of it is that credentials are a completely distinct data set from the underlying nature or the inner works with the genesis of the name, inner workings of a human being. And so to answer that question of, what does a person really like? How can you accelerate your understanding of that user or that human being?

And we were exploring lots of different data sets that it could inform that question. And one of those data sets, and we had a small team of, of behavioral scientists that were exploring lots of different approaches to answer this question. One of the key data sets were what's known as psychoinformatics, which is making behavioral predictions off of subliminal behaviors. Psychoinformatics that's one word. If you.

Dan Schmerin (07:10.524)

Psychoinformatics, that's one word. Yossi, you have a dictionary for that?

Oliver Quie (07:17.44)

If you use it, you sound very smart.

Yossi Hasson (07:18.885)

I've come across some psychoinformatics in my time.

Oliver Quie (07:22.26)

Yeah, yeah, yeah. And it's, yeah, it's, it's, it is a sincerely cool field and has a cool name. but the fundamentals are simple, which is behavioral predictions off the back of subliminal behavior on a, device or the internet. And, it was, it was a conversation with, with Dan whilst we were, Exploring early MVPs.

Dan Schmerin (07:22.555)

Have you now.

Oliver Quie (07:49.708)

that we understood that the pay, but the behavioral profiles we were generating off the back of how people use their devices. The first use case for those was not making predictions about how they might behave, but instead it was the fact that these profiles were completely unique. So no, two human beings on earth will own the same devices or operate them in the same way. So we've now emerged as a.

behavioral biometrics company working in the field of authentication, proof of humanity and proof of identity. So I hope that gives a useful background on how we got here today.

Dan Schmerin (08:27.176)

And for the layman, the reason that that's increasingly important is because of the rise of AI and the, the, you know, proliferation of deep fakes. And one wants to know, is this the real Yosi? Will the real Yosi please stand up?

Yossi Hasson (08:46.041)

I think we even saw it before the, we saw it before the world of AI and deep fakes. even if you think back to the Elon Musk acquisition of Twitter, one contentious issues was what's the real Twitter count and how many daily active users are there and how many of those are bots. And the number is a surprisingly large number of bot activity happening on platforms like Twitter and Facebook and LinkedIn. Yes.

Dan Schmerin (09:10.682)

You're talking about when he tried to back out of the deal, asserting that so many of Twitter's accounts were bought accounts, fake accounts, computerized accounts, and he didn't want to pay such a large number for nonsense. Is that right?

Yossi Hasson (09:18.961)

100 percent.

Yossi Hasson (09:24.379)

Correct. Yeah. And we used, if anyone who uses Twitter knows that it's also just not nonsense that a bot problem on Twitter, as an example, is huge. But problems in email, phishing attempts, et cetera. The internet is plagued by these bots that are ultimately trying to get access and be able to use that for financial means and gains. And Ali probably knows that better than anyone else in the world. And this is a unique way to solve that. it's, you know, it just,

translates into real dollars that are being spent for fraudulent activity effectively that's there. So Ali, how do you go from this kind of behavioral design that you just mentioned and pulling this kind of, this is how I click on the mouse and this is where I would go if I was clicking here and this is the routine that I do time and time again to that then allowing.

you to know that this is either a bot or it really is me. And that sounds like a lot of work and technology to be able to do that. Maybe talk through some of the inner works of how that happens.

Oliver Quie (10:32.621)

Yeah, sure. So a quick word on what behavioral biometrics are. So broadly speaking, it's data that can help inform the identity of the physical device and then how that device is being operated. So

On the operational side, we were particularly interested in very physical aspects of the user's behavior. So things like gyroscopic data, how do you physically hold the device? How do you touch the screen? How do you type? How do you move the mouse or touch the track pad if you have one? And just from the way that you move your mouse, for example, we can identify you with 92 % plus accuracy, even on a device we haven't seen you on before. So there's a completely unique.

pattern or fingerprint of behavior that can be gathered from very minimal time, like from just from your interactions on a landing page of a service, we're able to determine proof of humanity with very, high accuracy. And if we have time series data about you and we don't need very much, just a few seconds of data a few times. Yeah. Yeah.

Dan Schmerin (11:38.555)

Feels very intrusive. Feels very intrusive. This data is being collected by whom?

Oliver Quie (11:47.274)

Sure. So businesses are the, are the funnel for us. you know, a business that's concerned about are my users human? And then are the users that are accessing the service, the true account holders. So,

Dan Schmerin (12:03.132)

So you're saying that the way I navigate around a business corporate website is going to be readily identifiable and specific to an individual. That the way I navigate the other day, I was looking to purchase Thorlo socks. Do you know Thorlo socks? They're very padded. They're very comfortable. It's delicious for your feet. You're walking on clouds.

Oliver Quie (12:17.719)

Yes.

Oliver Quie (12:25.964)

Are they a sponsor? Are they a podcast sponsor?

Dan Schmerin (12:30.204)

Thorlo's a sponsor, they're not a sponsor. I just happen to like, I happen to like padded socks. Yossi, if you're generous and kind, I will introduce you to the world of Thorlo socks. So what you're saying is that the way I navigate around the Thorlo sock website leaves a particular digital footprint, a signature.

Oliver Quie (12:32.076)

You

Dan Schmerin (12:51.152)

that is going to be unique to me because I may look at men's socks and then I may do some filters and it's different from how Yossi would navigate around that page. Is that correct?

Yossi Hasson (12:59.769)

You would particularly spend a lot more time on the site than I think I would on the Follow Zox website.

Dan Schmerin (13:05.314)

If you have happy feet, you're gonna have a happy person.

Oliver Quie (13:10.709)

So, yeah, you touched on some really important points and we can determine proof of humanity before you've done any complex UI interaction. So, Interworks as a service can be installed behind the page of any native, any page of any native mobile app or website. So we could be on login page. We could be on a more...

a page that has richer, more complex UI. let's say it ranges from the complexity of a sign up sign in page to rich interactions that could happen on something like a game. You touched on privacy. We only release signals based off of the user's behavior. The only thing we release to a business is these are the users that we think are bots. These are the ones we think are human. These are the accounts that are either being hacked or shared.

So we don't release any behavioral data.

Dan Schmerin (14:03.494)

So you're not getting into the weeds of making a judgment call on whether I prefer white socks or black socks. That's not in the remit when it comes to privacy. You're simply identifying, we think that Dan is actually human because of the following criteria. And these other accounts don't appear to be human. They appear to be automated intruders.

Oliver Quie (14:26.368)

Yes. So, so there's two really important themes there. One is from a product mechanics perspective, you can label certain behaviors that are indicative of a bot and indicative of a human. And one thing that's been incredible to observe is, one of, one of our clients is one of, if not the largest crypto platforms in, on the market.

And four months ago, and they had an immense number of bots, you know, the majority of users that we observed in this particular subset of traffic that we analyze were not human. And that was post the authentication that the majority of users that cybersecurity leaders had approved were not human, right? And they were invested by bots.

and those users looked like bots. They had behaviors that we can label as like a bot. Four months later, when we've released another major report, most of the users look like humans. However, the proportion of bots is still the same, around the same ballpark.

So over that four month period, the field of behavioral biometrics has really changed. So what do you do when all of the behaviors look the same and look authentic? You start having to become a much deeper tech company looking at things like, well, are there hidden algorithms behind this behavior that is human -like?

And has this been generated by a device that actually exists? So we started doing checks, like running a mathematical equation on the client side's device to check as the processing speed really like an iPhone or is it as fast as a service station? So we've very quickly seen that limitations within the field and how the methodologies are having to react to generative AI techniques literally on a daily basis.

Dan Schmerin (16:26.747)

And let me ask, when I've been prompted historically...

by Google and others to try and identify that I'm human, right? It's click on all the traffic lights and they give me the nine boxes and I have to click on, and then I click on all the traffic lots and they do it again. Maybe a miss zone or maybe I'm just training their AI system. So they make me do it twice. And then, and then I see on Yossi's screen, he doesn't get the traffic lights, but he has the little puzzle piece that he needs to drag and he needs to come over and ooh, he beat 99.

9 .8 % of the people based on his skill dragging the puzzle and I really want to meet the people who are terrible at moving the puzzle piece because it's quite simple to do Why are those solutions, you know capture recapture? Why are those solutions not going to prove satisfactory? such that your System your solution is going to to be necessary

Oliver Quie (17:28.312)

So those types of bots are more effective than humans at passing those tests. So if you install a capture -based tool, in all likelihood, it is no longer an indication of whether that user is passing through the check as a human or a bot. And yet, they're the market -dominant tool. So you can see.

how there's a multi -billion dollar hole in that industry that we're looking to fill.

Dan Schmerin (17:56.454)

So it's not much of a mark of achievement at how good Yossi is moving the puzzle pieces, what you're saying. It really is.

Oliver Quie (18:00.6)

No, so, so it's, it's not a, it's not a complex task for a bot to complete. And when we benchmark those tests against the sophistication of generative AI tools, it's in the wrong ballpark. So the types of bots that we've been speaking about in the podcast so far are something that could pass clicking on photos of traffic lights or dragging a puzzle piece into a gap. However,

The types of bots that we think about that we know are emerging are ones that could log in to this very call and look, sound, seem like me or you. And that's the nature of the threat that's emerging. The sophistication of this technology that we're facing. And so that...

Yossi Hasson (18:51.247)

you think about that? What you're describing is in some ways a major threat, but other people are viewing it as the major opportunity where you can create these AI agents that could do the cold calling for your organization or outbound email.

Oliver Quie (19:07.799)

Yeah.

Yossi Hasson (19:08.893)

And bring more effectiveness and efficiency into your team to augment the work that is happening today. So there's this kind of universe of good bots or good agents that are out there. And there's what you're describing this as a new universe of bad bots and bad agents that are, that are out there. How, how do you think about being able to over time, delineate between the two to almost kind of let the good guys in, but the bad guys out the bad bots out.

Oliver Quie (19:33.654)

Yeah. So it, I mean, it's a hundred percent clear that AI based users and content will comprise the majority of the internet. It's already the majority of internet traffic that's bots and the majority of content online that is generated by automation tools. It's a question of quality and authenticity from my perspective.

I think there's a clear issue around the nature of those users, mostly being malicious right now and the nature of that content, not being like meeting the needs of the humans are supposed to serve. So I'm a, I'm a fundamental believer that these AI agents, you know, will be deployed for useful things for humans. However, they can only be useful if we're able to draw a distinction between a human being and a digital tool. And I think.

You know, it's incredible how history rhymes, right? Like, I think the types of, there's always been a fundamental trust issue between the physical and the digital world and all of these key moments of transition, know, whether it's been

currencies or communication, you the level of trust we all build day to day online with people, know, COVID was a great example, like, how much of real world communication or say real world, how much physical world communication was able to transition into the digital space quite successfully. But trust is a fundamental issue that will challenge a lot of the progress that's been made, I think already.

Dan Schmerin (21:14.608)

I mean, gets on a big issue because you'll see, and I've spent considerable time talking with our advisors about the fact that we're increasingly moving into a low trust society. Many argue we're already there and there's these data points of not just separating what's fact in fiction, not just trying to determine how to authenticate media, but that some of these societal

Oliver Quie (21:26.946)

Mm.

Dan Schmerin (21:44.43)

Norms are beginning to break down and you're observing this transition from physical interaction to more digital interaction. You point out that the salesperson that used to drive to meeting, to meeting, to meeting in order to close business development opportunities is now increasingly behind the camera.

doing it virtually by Zoom. And we're still big believers in human interaction, by the way. Gets us into a whole different topic around the loneliness epidemic. So fast forward to where things stand today.

Have you guys been able to test the product across a number of different businesses? You mentioned at least one example. What is the feedback you've been receiving over the course of 2024? Where are we in the journey?

Oliver Quie (22:44.364)

Yeah. So as you know, we're a young company, but our most exciting breakthrough is as things stand, we are the leading product in terms of accuracy of broth detection. So if you benchmark us against any major cybersecurity providers tool, we will detect more bots than them. The other

key value that we bring is that we're invisible. So we sit invisibly behind a service. Authentication is, there's off, there's historically been a choice of fraud resistance and friction, right? So if you insert more friction, you get a higher level of fraud resistance. For example, signing in with your Google account is a bit quicker than going through facial recognition plus KYC plus uploading.

proof of address, this type of thing, right? And you see services where fraud could be higher risks, as other than like a financial platform versus where fraud is seemingly lower risk on something like a social media platform. However, what we've observed as a company is that there's a huge opportunity to have a low friction product that delivers the level of fraud resistance that's affiliated with.

deep cybersecurity enterprise companies, ones that are doing KYC with facial recognition, which is not going to work much longer. Because if I find a few photos of you online, I can generate an image that will pass facial recognition. I could open a bank account in your name if I'm able to steal your centralized government ID. no.

Dan Schmerin (24:24.132)

And neither will voice recognition, right? We know that all these financial institutions are freaking out because they ask Yossi to come in voice verified. Hello, I'm Yossi. And he says it with that elegant accent and fidelity or Schwab lets him in. Those days are over because we're going to take a couple of clips from his online commentary and you can access his financial accounts. And so we know that this is a major issue.

Oliver Quie (24:34.679)

Yeah.

Oliver Quie (24:49.88)

Correct. you know, the, the status quo has been, if we get on a zoom call, for example, my likeness and my, sound of my voice and my behaviors are indicative of my identity. And that's no longer going to work in, I would say 18 months to two years, the, the, the, scale of the, of the issue and the sophistication of technology.

will be such that observing my behavior digitally won't be enough to know it's really me, especially when we see how much content, images of ourselves we're all putting online. It makes it easier to feed into these models.

Dan Schmerin (25:32.742)

But you made a pretty big claim just a few minutes ago, which is that the major companies that are in this space that are known for providing such services, many of which are publicly traded and have monster valuations, are not actually proving as accurate as little old inner works. How is that possible?

Oliver Quie (25:42.832)

Hmm

Oliver Quie (25:53.868)

Yeah, I mean, from our perspective, this is very much the blockbuster Netflix moment for cybersecurity. I think there's an entire perspective that there's a status quo that won't change. And there's an underestimation of the fact that almost all of the technical strategies used for authentication and identity will start to fail.

And we've started to, you know, we've seen, huge indications of that. mean, the, Twitter acquisition was an example of unveiling of the scale of fraud online, and on a use case that people underestimated.

the impact of right so people get scared about identity theft or authentication when it comes to something like Losing their money on a financial account. They think what's the big deal? or they don't think that they're as Manipulated by bots on Twitter than than they might be but you know we

We observe that people can be living in essentially fish bowls of isolated fish bowls of AI -generated content online that are dishing out disinformation at huge scale that can have an impact on social stability, political stability, electioneering.

Yossi Hasson (27:19.525)

Yes, state actors are implanting, you know, stories for just to cause social unrest.

Oliver Quie (27:24.682)

Yeah. Yeah. I mean, one of my co -founders was posting, you know, earlier this year we had, know, riots across the UK. often where, you know, physical violence was, occurred because of disinformation online. so when, when we think about the biggest, as a company, when we think about the biggest threats, of, bot generated attacks, I think about

social stability because of disinformation, and then the other one being the security of people's digital assets, namely their money. These are as big as issues can get. So much of our world is now digital. Ownership and security is meaningless without an effective identity solution.

Yossi Hasson (28:19.653)

And speaking about that, you know, so much of the world is digital and, it becomes an even bigger part of our lives as we, as we progress. You guys have used a pretty interesting approach, leveraging the latest technologies, know, web three blockchain, ZK proofs, et cetera, in how you tackle this problem without going too deep into the tech. Maybe just touch on how that has enabled you to do what you do today, where you're saying, you know, you're outperforming.

any of the other bot detection providers that exist today, which is pretty impressive for a young startup, tell us a bit about the underlying technology that enables that.

Oliver Quie (29:00.864)

Yeah, so there's two key elements of our IP. One is the way that we process behavioral data. So we have a unique and a differentiated approach to looking at the behavior that we can collect off of the hardware device sensors on your device. However, the second element is that

You see many cybersecurity companies deciding to specialize with one niche. know, so for example, they might do a capture based test. They might do some behavioral biometrics. They might do tests related to checking to see whether users are using tools like VPNs or blocks on their browser that can stop browser fingerprinting.

But they haven't brought all of these tools together. So, you we have what I would describe as a multifaceted approach where we don't see any reason why every factor that you can include in a solution should be included. So we brought together a lot of different strategies and different verticals within cybersecurity that we've observed, but not at the expense.

of their user experience. we believe that...

organic continuous authentication is the answer. So what you organically do online, the way that you behave, the transactions that you execute, the content that you consume, the way that you are is the most powerful indicator of your humanity and your identity. So from a technical perspective, think what's, should have said this first really.

Oliver Quie (30:50.102)

What's really important about us is that we haven't created certain categories of checks where we say, if yes, then bot or if no, then human. We are constantly adapting using adaptive learning technologies to the individual behavior of the user and the place that they've emerged online, the device that they've come from. And, you know, yeah, sorry, go ahead.

Dan Schmerin (31:14.554)

I love it. want to, I want to bring all of this back to the thesis for this podcast, which is re -imagining ownership. What does this technology and the differentiation between legitimate users and automated bots.

have to do with digital ownership at the end of the day? What does it have to do with integrity at the end of the day? And I say that because for some platforms, having bots means larger top line numbers. And they use that for funding rounds. They use that for promotion purposes. We have a good billion users on our platform. Do you really have a good billions? This is the Twitter issue, but it's not limited to Twitter.

Oliver Quie (31:39.522)

Hmm.

Oliver Quie (32:01.9)

Hmm.

Dan Schmerin (32:03.514)

We've seen this in web two. We've seen this now in web three. We believe being able to distinguish and differentiate between authentic and inauthentic is important, but talk to us about the impact it has around digital ownership.

Oliver Quie (32:21.848)

So the best way I could explain it is public versus private. If you're the only ones with the keys to your house and you store the things that you value behind that door, and if it becomes the case that actually anyone can come through and it's too easy to simulate being you or too easy to steal those keys, ownership becomes meaningless. Privacy becomes meaningless. Where could you keep that stuff?

So in the physical world is very obvious. We, we have a clear understanding of the things that are ours and the clear understanding of who should be able to access those things. And then now we're in this strange limbo period where so much of our behavior, our transactions happen digitally, but there's a blurred line between what's my stuff, how do I access it? And how do I put a gateway that I can control?

And Web2 simply doesn't enable that. So fundamentally, if you're a believer that the economy will increasingly progress to be digital, then you have to be a believer in the ownership economy. by virtue of that, believer in decentralized ownership.

technologies that is the only technical framework that we can conceive of that enables true ownership of something digital. So, you know, that's a real world issue, like, know, Web3 infrastructure.

we believe is the only thing that can enable the access control and ownership that I just mentioned. And what plugs into that infrastructure is a slightly different question, right? So.

Oliver Quie (34:13.492)

what mechanism can we use to anchor the relationship between a physical human being and their digital world? And the answer is no longer a picture of your face and centralized government ID, it's no longer a username and a password, all of those things no longer work.

Yossi Hasson (34:33.893)

Lots more that I'm sure we could dive into, but we're out of time. We'd to end off Ali on a more emotive question on, can you think back of what was the first thing that you felt like you owned and why was it something that was important to you?

Oliver Quie (34:52.28)

Yeah, I think I mentioned when we met up recently that I grew up in Japan. And you know, the gaming scene in the 90s in Japan was amazing. Like, you know, it was a sense of community and energy around these physical hardware products and these games that I didn't experience here in the UK until much later. But I remember being given a Nintendo 64.

and let me assure you that that's cool. And so the majority of my life transitioned to a digital world at that point, I would say.

Dan Schmerin (35:23.036)

advanced compared to my Nintendo. I had a Nintendo 6 .4.

Yossi Hasson (35:37.431)

As for many of you, Nintendo clearly a big part of many of our lives, especially in the Web3 world. Go for it. I'll allow it.

Dan Schmerin (35:44.656)

Yossi, I just want to ask one bonus question. Okay, you permit it and it is a little bit technical, but I just need you, Oli, to refresh for me. On the topic around privacy and individual security, you have been an innovator and a leader in the use of zero -knowledge proofs.

Oliver Quie (35:44.982)

Yeah.

Dan Schmerin (36:11.59)

Can you just recap again in sixth grade terminology? Why is that so critical? What is the ZK proof and why is this going to be such an important unlock as we go forward for all the things we've covered today?

Oliver Quie (36:28.066)

So zero -knowledge proofs are a mechanism to only disclose a specific set of information and keep everything else private.

It's the fundamental technology to enable something we understand of a selective disclosure. I have a pool of information about myself. I only want to tell you a little bit. An example would, you the best analogy would be, you're at the, you've gotten to the end of the line, the front of the queue at a nightclub, you show the bouncer your ID. The only thing they need to know is that you're over a certain age. They don't need to know your name, your address and all that other information.

So zero knowledge proof is a technical mechanism to know that there is no way this third party can find out anything else besides what you intend to share. So in our case,

we're increasingly going to start collecting some pretty invasive stuff. want to provide a also zero trust mechanism, sort of like don't trust inner works, don't trust the third party. Trust this trust that is architecturally impossible for me to access any of your information apart from the fact that you're a human being and you're the right account holder. I don't want or need to know anything else.

However, you know, on web two, I'm disclosing a bunch of sort of ancillary information that is leading to a level of data mining where I'm being forced down a line where I'm only needing to disclose a certain set of information to access a service, but I'm being forced to disclose loads more. And most of the time I don't really know what that stuff is. I'm just enabling some privacy preferences, opening the door, giving the keys away and the stuff away.

Dan Schmerin (38:08.048)

Well, we'll end on that point and going forward, people will not know whether Yossi wears Thorlo padded socks or not. They need not know. They should just know he has socks and he's got shoes and he is comfortable in those shoes. Great to be with you, Ali. Thanks for taking time to join us. We're very excited for what you're building and looking forward to much more success from you and your team.

Oliver Quie (38:28.459)

Likewise.

Oliver Quie (38:35.715)

Thank you for being a part of it.

Dan Schmerin (38:39.866)

Yo's, that's

‍