How Ad Blockers Work, Why They Break, and the Future of Private Search (JP Schmetz Interview)

Most ad blockers start by blocking everything—then you become tech support for friends & family when pages break. Henry interviewed JP Schmetz, founder of Brave Search and CEO of Ghostery, about making trackers visible, why there are only 3 search indexes in the world, reinventing the open web with AI, and how to avoid becoming "the family CTO."

🔗 SOURCES & LINKS
• Ghostery: https://www.ghostery.com/
• Who Tracks Me Database: https://whotracks.me/
• Ghostery Private Search: https://www.ghostery.com/private-search
• Brave Search: https://search.brave.com/

⏱️ TIMESTAMPS
00:00:00 INTRO
00:01:32 ORIGINS WITH GHOSTERY
00:03:15 SEARCH ENGINES + BROWSERS
00:06:32 GHOSTERY'S BROWSER EXTENSION
00:09:09 THE PRIVACY DISCONNECT
00:13:02 REAL WORLD IMPACTS
00:16:03 PRIVACY VARIES BY LOCATION
00:16:47 PAGE BLOCKING + BREAKING
00:24:06 ADAPTATION + TRUST
00:28:48 ROLES AT GHOSTERY + BRAVE
00:29:22 OPEN SOURCE
00:30:38 BRAVE SHIELDS OR GHOSTERY?
00:32:36 SUGGESTED CONFIGURATIONS
00:40:13 WHOTRACKS.ME
00:44:30 WHAT IS A TRACKER?
00:47:39 WHAT ARE THE INCENTIVES?
00:50:31 DOES BLOCKING = NO DATA SHARING?
00:53:50 THE INVESTIGATIVE PROCESS
00:57:27 PRIVATE SEARCH
01:00:11 META SEARCH VS. INDEPENDENT INDEX
01:02:30 CONTEXTUAL VS. TARGETED ADVERTISING
01:06:28 BRAVE SEARCH ADS
01:08:16 INDEXES
01:10:53 AI + SEARCH
01:19:43 WHAT'S NEXT?
01:22:22 PUBLISHING + INCENTIVES + THE FUTURE
01:25:30 HOW TO FOLLOW

🎥 VIDEO
Watch on YouTube

🧡 SUPPORT TECHLORE
Keep Techlore Talks independent & growing: ★ Support this podcast ★

Creators and Guests

Host

Henry Fisher

Runner, artist, musician and digital rights activist. Owner of Techlore

Guest

JP Schmetz

Ghostery & Brave

Editor

Tori

Techlore

What is Techlore Talks?

Techlore Talks brings you in-depth conversations with the experts at the forefront of privacy, security, and digital rights. Hosted by Henry Fisher, founder of Techlore and long-time digital rights educator, each episode features meaningful discussions with the people building, researching, and advocating for digital freedom.

From cybersecurity researchers and privacy tool developers to open-source advocates and digital rights activists—if they're shaping how we protect ourselves online, they're on this show.

Topics include: privacy tools and technologies, cybersecurity threats and defenses, open-source software, surveillance and digital rights, encryption, tech policy, and digital sovereignty.

New episodes released regularly. Subscribe and join the community at techlore.tech.

It's very rare to hear someone say, oh, I used to use one of these tools and now I stopped.

I've never met someone like this.

Today, I have the absolute pleasure of having JP on from both Ghostory as well as the Brave team.

Today, we're going to talk about the challenges of ad blocking and why it's difficult for maybe

beginners to get into it and the breakage that can be experienced by new users, as well as what

we can do to help people in our lives actually start utilizing those tools.

He talks a lot about regional differences between ad blocker usage, why people should probably be using an ad blocker regardless of what happens, what the ad industry actually looks like behind the scenes, investigating trackers, and also a huge amount about search engines and why search engines are actually a big part of this battle as well.

I have to say this is one of my favorite episodes I've ever recorded.

JP brought a lot of experience to this podcast, and so I am just thrilled to be here and be

able to share this conversation with all of you.

And I hope that you all walk away from it thinking about one or two things, not just

in your own privacy and security journey, but also for those around you.

Today, I want to welcome on JP from Ghostery, and I would love to just open it up for you

to introduce yourself and also maybe teach me how to pronounce your full name, because

I'm so bad at name pronunciations. Well, thank you for having me on. I'm from Belgium, so I'm a bit

of a mix between French and German. So that explains the Jean-Paul or JP and the Schmetz,

which is very German. So call me JP is probably the easiest way to address me. Perfect. And then,

you know, when did you start working on Ghostory? Why Ghostory? What's kind of the origin story

there. Yeah, so I didn't found Ghost Rea. I'm not a founder of Ghost Rea. It was founded many,

many years ago in 2008, I believe. At the same time, I had founded a search engine.

And we came together in 2016 when we realized that as a search engine, you needed a browser.

And we built a browser and we wanted to make this browser a browser that we wanted to use ourselves.

And that would include essentially privacy preserving things like tracker blockings and ultimately then ads blocking.

And so we became quite close with the Ghostory team and acquired them when they were, their company was splitting between the B2B part and the B2C part.

And we just came up and acquired them in 2016.

So it's been 10 years.

What were you called before Ghostory?

So Ghostory was always called Ghostory.

what the mother company is called Clix

and it's still the case today.

And Clix was building a search engine

which is now called Brave Search.

So we sold the search engine to Brave, the browser in 2021.

Okay. And then we can definitely touch on that.

I don't think people, I'm learning about this as well now.

Yeah, yeah.

It's all a big family of privacy conscious people.

Yeah. And I mean, Brave Search is my default search engine still.

And so I'd love to ask more about it.

That makes me very, very proud because I founded the company that became Brave Search 18 years ago now.

That's awesome.

You mentioned building a search engine required a browser.

Do you mind expanding a little bit into that and why you say this?

So rather, I think, you know, browsers and search engines are basically joined at the hip.

It's a little bit like a car and a steering wheel or a car in a navigation system.

And you, you know, let's forget a little bit about AI because the world is changing a little bit today.

But the only purpose of a search engine is to bring you to a web page that a browser will render.

And if you have a browser, you would rather not type URLs, you would rather type queries and then find the page that you want to go to.

So these two things are very, very related.

And it took a while, obviously, for these two things to come together because in the 90s, the browser companies were not search engine companies and the search engine companies were not browser companies.

But obviously, that came together first when Firefox got a deal from Google to be kind of like the official distribution channel for Google.

And then Google created Chrome.

And eventually, you realize that if you have a browser, you need a special relationship with a search engine.

And if you have a search engine, you need a special relationship with a browser.

Like having a search engine in a vacuum worked for Google for a few years, but this is not,

this is like building a steering wheel company or something.

You'd be like, okay, great.

Maybe I should expand to cars as well.

So these two things are the same thing.

Yeah, it's interesting.

I think people, there was rumors that Apple was developing a search engine for a time.

And I feel like that sounded a bit off-brand for Apple, but also they build a whole browser.

They build Safari.

And I feel like their own search engine would actually accompany that quite nicely, similar to what Google did with Chrome and their own search.

Yes, but then in that particular case, money comes in the way, right?

Because building a search engine is expensive, right?

It's not expensive at the point that Apple could not pay for it.

But it's expensive for Apple specifically because they would first lose 20 and some billion that Google pays them.

And you multiply that by the P-E ratio of Apple, it means the market cap of Apple would first go down by 600 billion before it can go up because they built a search engine that may or may not be as good as Google.

So it's a very tricky thing to get out of.

They tried with Maps.

As you may remember, they had a deal with Google Maps on the first iPhones.

And then they decided that it was strategic and created Apple Maps.

And I think there's still, it was a traumatic event in a sense,

because it didn't work as smoothly as they thought it would.

Yeah.

My friend and I, we call them Apple Maps-isms.

When it makes these funny mistakes.

And I think that Apple learned from that experience that

they should think twice about building a search engine.

And it's very unfortunate because obviously they should.

Yeah, and so I want to ask later too, you know, I have a lot of questions building up very quickly here, but I do want to ask if it costs $20 billion to build Brave Search as well.

But before I...

The answer is no, no, no.

But I did want to, I guess, go back to Ghostory for a second before we dive, you know, later on more into the search engine side of things again.

But the main thing that our community will probably know Ghostory for is the tracker and ad blocker extension.

And so I'd like to start with overall what this is trying to accomplish and maybe how it compares to different philosophies taken by other ad blocking extensions like uBlockOrigin, AdGuard, etc.

Yeah, so the first keyword I think is extension, right?

So we don't pretend to replace your browser.

We are an add-on to your existing browser and habits,

which means that most of the users use it on Chrome

and a minority of users use it on Firefox and Safari and Edge and everything else.

We are available on all platforms.

And the main goal is to, on the one hand, it is to make the invisible visible

because if you have Ghost reinstalled, you will see the trackers.

You will see what's going on behind the scene if you're interested.

Ghostry was, because you asked about differentiation factor,

Ghostry was always the most educational of all the ad blockers

in the sense of showing you a lot, right?

We will talk about who tracks me, et cetera.

So we do have a deep interest in specialization

in trying to make it visible.

All of the stuff that we're taking off,

we want to make sure that you see what's going on.

And then because I've never met a person in life

who has seen what's happening,

without wanting to block it, we also do block it, right?

Because once you notice that the experience of browsing the web

without protection is a little bit like going to a supermarket

and being followed by 100 men in black,

talking into their little earpiece about what you are buying

and keeping a list on you and making sure that when you go to the next shop,

the same guys are there as well and sort of comparing nodes,

no one wants that, right?

And so the goal of Ghost3 is to make that visible and also obviously to block it because, like I said, I've never met anyone who says, oh, I see that.

That's really interesting.

I'd like to keep that on.

Right.

And then a lot of advertising goes away.

So I don't think the primary or at least I know that the primary focus of Ghost3 at founding was certainly not to block ads.

But it turns out that because ads are connected to these third parties, so to these, you know, men in black that are following you around, if you block them, then you also block a lot of ads.

Because the advertising model of the web in the last 10 years has basically shifted to something called programmatic, which depends a lot on the information collected by these third party trackers.

Well said.

So the whole concept of making invisible visible.

What are your personal views?

Because I'm sure you might have seen some of these.

There's been studies where they ask people if they care about privacy and the majority actually say yes.

And then they actually dig a little deeper and ask, OK, well, what do you do for your privacy?

And it turns out they actually have very poor privacy hygiene.

And so where do you think the disconnect is for people?

Do you think that people don't see it, like you say, like it's all invisible?

And then when you make it visible, then they're a little bit more empowered and they do something about it?

Or is it a lack of tools that allow people to reasonably take control of it?

And if those were a little bit more accessible.

And of course, I'm sure there's many factors, but I'd be just curious for your personal thoughts on this.

Yeah, all of the above.

I think that if the invisible was visible, sort of in the analogy that I gave you, right, that you would go to a store and then there would be people actually following you and taking notes, I mean, they would get beaten up, right, basically.

And the fact that online, for some reason, the world evolved a bit too quickly, I guess, for the human mind, and people are not that technologically minded, and they have something to do, right?

They go to buy something.

They don't want to look at the source code of the HTML page.

And even if they did, they wouldn't be able to read it.

So it really depends on a few specialist companies to create products like ours.

And then, you know, how do people find out about us?

I mean, they find out about us maybe in the store, but that means in the extension store, but that means they already know about the concept, right?

So most of the time they will hear from a friend that says, why don't you use a tracker blocker or an ad blocker and I use Go3 and you should do too, right?

In certain countries, this reach pretty much half the population.

I don't underestimate.

It can be very big in Europe, for example.

But I think in the world, it's probably 20% to 30%, I would say, that people will use it.

And people who don't use it, they don't, it's not, I've never met really anyone who said,

oh, I don't use it because I prefer the experience without it.

It's mostly because they don't know, they don't trust themselves.

There is an issue.

We are actually developing a new type of product or a new type of mode in the extension.

because if you block ads and trackers,

sometimes you do break a page, right?

And this is okay if you remember that you have Go3on

and you can disable it for that page just for that particular moment.

But if you install it on your parents' computer

and they have forgotten that they have this

and then they cannot book a parking ticket or a cinema ticket or something

and they think the web is broken, that has the consequences.

So some people do not use it, I think, because they don't know.

Some people don't use it because it breaks.

But it's very rare to hear someone say, oh, I used to use one of these tools and now I stopped.

I've never met someone like this.

Yeah, it's quite rare, especially because it requires someone to then get to the point where they know it's a problem.

They already did the action to stop the problem.

And then they had to somehow go back.

Yeah, it's a tough path to take.

Because the web is clearly faster, right?

The web is also clearly, you cannot say that because Google will kick you out of the extension store,

but your battery on your phone will last longer because it simply uses the network a lot less.

And frankly, the web is cleaner, right?

And if you still want to support your favorite newspaper with their ads, fine.

you just unblock it where it's easy. Right. So you mentioned this men in black example in a digital

world, and I have a few follow-ups to some of the stuff you covered. If someone's new to this podcast

and they hear this example of men in black in a grocery store, it's a really good illustration of

maybe what is going on. But I think a lot of people might still go, okay, so what, right? Because if

it's not happening, I don't know what's happening, but what are some of the real world implications

you might be able to point to that you can tell people, hey, this actually will impact you in these

various ways and this is why you should care i mean first of all what is happening right because

the men in black in the grocery store they unfortunately tend to be also at your doctor's

appointment later and so it's not like it's limited to supermarkets or something right

and then they build a profile of you online which means that somewhere there are a number of

databases that have your full browsing history and they know all of what you have done

online which i i mean just ask anyone like would you be comfortable sharing your browsing history

with me and the answer is no we've done all these things by the way it goes through like going in

the streets and people you know asking people show me your browser history and the answer is never

oh yeah sure like so obviously there is a problem that people psychologically do not like the idea

right and it creates databases that if they were exposed you know no one would find this amusing

and then basically they are used that's probably the least harmful part of it at the moment I think

they are used to target advertising to you which you know if done well could actually be quite

useful albeit a bit spooky right so I don't think advertising is the core problem I think that the

problem is is you have this nuclear waste of data that is being collected just to serve you an ad

that's a little bit more relevant than another ad.

And most of the time, not, right?

Plus, you have a bunch of side effects

that are probably not super interesting for normal users,

but you will find out that, for example,

on some publications are used to collect data

and other publications are used to put the advertising.

So if you read The Economist,

they will deduce from that

that you are probably a bit richer,

a bit more smart and blah, blah, blah,

and then use that information to target you on a platform that is less expensive than the economist.

So you would wonder why is the economist doing that, right?

And a lot of websites are not aware of all the trackers that they have on their own website

because over the year they just had this tag and that tag and that business development guy said,

hey, if you put this on this, maybe you would have a little bit more revenue and then they forget to take it off.

So it's, yeah, it's just a mess that everyone who knows about it, like we did when we acquired Ghostry, we basically look at the browser and we say, why the hell is the browser doing all of this thing on behalf of web pages?

And you could not find anyone, including the web page, that would find it good.

So the only solution is to remove it after its build rather than try to convince the entire industry to stop this nonsense.

Right. And then you mentioned some regions where it might be 50% of people who use these tools. Can you speak to where there's this higher prevalence?

I mean, typically you have countries like Germany, which are, for various reasons, more privacy-centric and also more technology-savvy.

You put these two things together, and then you need one magazine or one website that promotes the use of these things or explain it to users, and then you have an explosion.

So I think Germany comes to mind as being particularly high.

Asia is quite low compared to Western Europe, and the US is in between somewhere.

And then you mentioned page blocking and how some things might break certain elements on the web page.

How has that been going, especially in light of MV3?

So I don't know if MV3 has a huge effect on it, but basically the way these lists,

so the way these blockers function, because all of them, I mean, it's not like everyone has a secret sauce

in terms of what they block and what they don't block, right?

So most of these lists are open source and publicly available lists.

And they are done by volunteers or people working for ad blocking companies.

And sometimes the tooling is a bit rough because you will say,

oh, block this request.

But then maybe a website like an airline is using some specific way

to track this request to avoid bots traffic.

I don't know.

Some engineer comes up with a clever way

to use something,

but you don't realize when you do the list

that blocking it on site X

is going to break site Y

on the other side of the world, right?

So now if you're tech savvy, it's fine.

You reload the page.

You see it doesn't work.

You disable Go3 or Ublock or whatever,

and then it works.

Then you go, okay, problem solved, right?

It's only when you have more, you know, less tech-savvy users that that side looks like it doesn't work anymore, right?

Or a typical example is that you will have a sort of cookie notice popping up, and we remove it, but then the side doesn't scroll, right?

That sort of things.

So if you know how your tool function is not too difficult to take it off, but if you forget you have it, which a lot of people do after many years of having it, right?

Or just at that moment, they blamed the website or blamed the Wi-Fi or something.

And then they cannot use a site.

That's very frustrating.

Plus, you have sites that can break and cannot be easily tested by us because it's after the login, right?

Or you're buying something and the last step is not working.

So these are things that we have to be very careful about.

And grocery was always at the forefront of really thinking and paying attention to that.

Whereas other ad blockers tend to assume that the users are very tech savvy.

So we all have different positioning in sort of the ecosystem.

Got it.

Yeah, I was mainly asking because I actually can't point to one example, because I installed

uBlock Origin Lite in Google Chrome for a family member, and they started dealing with

some breakage.

You become the tech support, and that's no fun.

Exactly.

And they actually uninstalled.

So they are the person who took that wrong path, but they didn't uninstall it.

They moved back to their old ad blocker, which is fine, but it's not my favorite ad blocker.

So we have a mode for that that we are going to launch soon in grocery.

So it's probably good that we talk about it.

So at the moment, if you install, first of all, you have this particular pattern, right?

You install it for someone else that may not be as tech savvy as you.

And then second, the experience with the ad blocker is always, oh, something doesn't work, I need to disable it.

So it's a kind of a negative experience.

And the default is I ignore it because it's doing its work, but I don't really notice.

We have a mode that we call Zap in Go3 that we'll launch soon, where you start by blocking nothing.

So it's a bit ironic that you start by blocking nothing.

But every time you go to a site that bothers you because it has too many ads or it looks like sluggish or something,

you click on the zap mode, and from then on,

this site will always be ads-free and trackers-free.

So basically, you can tell your family member

that you install it for, whenever a website annoys you,

just click on zap, and then from that day on,

that will be it, right?

And they will never zap their banks.

They will never zap something where they do a transaction on it

because it doesn't bother them.

There's no ads.

There's nothing visibly problematic.

But then they will go to a new site that is particularly overloaded with ads and pop-ups and whatnot.

And then they will just simply zap them.

And from then on, life will be like an ad blocker.

So we don't know if that's going to work.

But this is an interesting twist on the concept of ad blocker, especially in the context where you install it for someone else.

And it can never break by definition because what's going to happen, they will zap the website.

If the website doesn't scroll afterward, they will obviously unzap it because they did it 10 seconds ago and will remember.

and if they keep it on then we know that it works right everyone knows that it works they know that

it works and they're happy about it and it turns a the experience of adblocker into a very positive

sort of agentic experience where you say wow i've taken the decision i clicked the button and now

life is better so we'll see you know it's one of these modes that we will have and then next time

you install it for a family member you can try them on the zap mode and see what uh whether they like

it or not yeah on paper it's it's kind of brilliant because it really puts them in the driver's seat

now when you zap a site is it is it essentially like enabling your typical added tracker blocker

for that specific site exactly so it's it's instead of of creating a list of exceptions you create a

list of um of positive uh sort of um turn on instead of turn off got it it's obviously has

some effect on privacy, but in fact, it's quite limited because if you look at how people use the

web, I mean, they have a set of, I don't know, I would say 90% of what you do on the web is

repetitive in the sense that you're going to go to the same sites. And then 10% is random because

you click on a Google link and you go somewhere. But seeing a tracker or an ad on the 10% is never

going to overly bother you right uh it's more like the repetitiveness uh of of something annoying

happening on something you do every day that gets you yeah i mean it's a great middle ground and i

think i would love to see it and my my guess and my hope is that as people started taking more

ownership they would actually be frustrated at the manualness and then they turn on the whole thing

Exactly. So I would hope that it would be some kind of way to make the tool a little more accessible to people.

And I'm sure. Yeah, I'd love to try on that family member.

Yeah, well, we'll tell you as soon as it launches, it's going to be very soon.

But they, yeah, if you stop where you want, because some people will say, I love the zap mode.

It's fine because I've zapped, you know, the first week I zapped like a lot of things.

And now I'm basically set up and I don't need to do it anymore.

And other will say, well, now I get the principle.

I just switch to full mode and I'll go the other way. Right.

yeah it's um and at least you don't turn yourself into the family cto if you install it for a family

member or you don't get blamed and and whatnot right it's uh it's true we've interviewed our

users and we've asked them very specifically just before thanksgiving like like what prevents you

from installing grocery or brave on your on your family member computer and basically that's the

reason like i don't want to i don't want them to blame me afterwards because it broke their stupid

you know, banking app or health insurance thing or some work tool.

Yeah.

You know, this reminds me, I've been thinking about this quite a bit because I use Brave

as kind of my go-to browser.

I have a few others installed, but it's definitely where I spend 90 plus percent of my time.

And I had to set up a new desktop recently, a Mac mini, and I learned that Brave Sync doesn't

seem to carry over to Brave Shield settings.

through Brave Sync. I don't know if I did something wrong or if that's just a design,

you know, limitation. But I actually, one of the reasons why I was kind of late to our interviews,

I had to re-log in to Riverside here because I always enable forget my, you know, the forgetful

feature in Brave Shields where, you know, you log in somewhere and then, you know, a few seconds

after you close the tab, it wipes all the data from that tab. It's a fantastic feature. But I

I already have all my exceptions that were on this other computer.

Now they're not there and I have to go through all of them again.

But it got me thinking.

I feel like there is a very obvious use case here where if someone's logging into a service,

there is probably some degree of trust they have in that service.

And I feel like there could be a more automated way of approaching things for not just ad blockers,

but also things like Brave, where it's like, oh, if I automatically log into a service,

then add it to the exceptions of, you know, forget me mode.

And then it's a little bit more adaptive and easier on users.

I think you touched upon the right point is,

do you start from an idea that you trust nothing,

which is, I guess, the aggressive mode of ad blockers

and privacy trackers and generally these tools?

Or you would start from the extreme opposite,

I trust everything except what I don't trust,

which is, I guess, the new zap mode?

Or do you have some kind of intelligent adaptive system

to say, well, you seem to be trusting this thing by logging in, for example.

Because privacy becomes super, let's say, relative if you are logged in Amazon

and you are shopping under your name and your address and you say,

oh, but I don't want, like they don't have third-party trackers

because they're smarter than that and they don't want to share the information.

So there's no reason, like you're already giving them anything.

And you've done this very willfully because you want the product to come to your home

and not to a random address.

So I think the browser, yes, it's right.

Unfortunately, and that's probably the reason for sync, et cetera,

you have always to balance privacy,

some kind of idea of privacy with the feature you are building, right?

And this is a tricky balance because you have a bunch of users

that are very idealistic about it and they say,

well, sync should work without accounts

and Brave should not know anything about anything.

But I still expect them that I get a new Mac Mini and somehow everything carries over without any problem.

Now, there are methods to do that, but they're not bug-free, right?

And it's a little tricky to do.

Let's say it's trickier to do than having the Google way, which is,

trust us, we'll just put everything under your Gmail address and then it will seamlessly carry over.

So you have a balance and it's hard for companies like us to strike it correctly, right?

And when it comes to these adaptive things and AI, it becomes even more tricky because some users really hate AI.

So the last thing they want is an AI to run in the browser to observe their behavior and then take decisions.

And other people find that extremely cool, right?

So again, lots of different people.

And when you look at Brave and Grocery level, you're talking about tens of millions of users at Grocery and hundreds of millions at Brave.

And you go like, oof.

These are a whole world of different types of people, right?

yeah and it's hard to find it's hard to do a feature where everyone goes yeah yeah that's the

right way to do it yeah yeah i wasn't uh i hope my question didn't come across as uh like hard

criticism because i i do love it's one of my favorite features but it was more to just it is

reasonable and i will definitely look into it or tell the team it was more to illustrate automated

more way of doing things where like, you know, if you log into a service.

Shields and generally exceptions in ad blocker is something that you spend,

you know, a lot of time by definition building, right? It's hard to replicate because like if

I would point a gun to your head and say, tell me all the exceptions you have in shields,

you'd be unable to answer, right? Like it's not possible. Like it's too many and over too long

of a time and too random, right? Yeah. My best bet would probably be to open up my password manager

and just go through all the things there. What are your roles actually between Ghost3 and Brave?

If I can kind of dive a little bit more into you. So on the Ghost3 side, I'm the CEO and they use,

remember that Brave Search used to be the same company as Ghost3, right? They were sisters

and then they separated five years ago. So you have Ghost3, I'm the CEO of Ghost3

and I am at Brave.

I'm a board observer

because I sold in the search engine

and I'm also taking care of Brave search ads

in particular.

Got it.

So how to monetize the search engine.

Got it.

And then when it comes to

Ghost 3 going open source,

I know this happened quite a while ago,

but what was the idea behind making it open source?

I think that in general,

if you do a product that relies on,

you know, having a strong privacy promise,

it does help to do it in an open source way

because people can then verify,

I mean, at least some people can verify

that what you are saying is true.

In my experience, the reason for my companies

to do open source stuff was very often

to be able to tell people,

you can look at the code,

we are not doing what you claim we are doing.

You know, people have theories and conspiracies, etc.

And it's very easy if the code is open source to point it to it.

Plus, as an extension, you are almost by definition open source

because extensions are not compiled.

Now, you can program them in a way that they look like

they basically become unreadable,

but then you have software that can make them more readable, right?

So in a way, we were open source by design

because that's how extensions are.

But we were also making promises about privacy,

and we felt that it was very important to make them verifiable.

Got it.

And then if you have a Brave user,

are they supposed to be using Ghost 3 or Brave Shields?

I mean, I would have my own answer to this,

but where do you see Ghost 3 fitting in someone's workflow

versus just using the Brave browser?

I mean, I can tell you what I do

and I can tell you what I see other people do.

So obviously the majority of Brave users use Brave Shields,

but there is a significant minority that uses Brave Shields

plus another ad blocker for double protection, I guess.

It's a little bit advanced because there are some interactions

and obviously how do you create exceptions

if you have two things running on top of each other, right?

That sort of thing becomes extremely complex

and there are some talks between the ad blockers

that they would share exceptions

or at least a signal of an exception between one another

when they run on multiple ad blockers on one computer.

So we know that there are multiple people using Go3

on top of Brave Shields.

I've been testing the Zap mode for myself.

So I've disabled Shields and used the Zap mode.

And I like it a lot, by the way.

They have a few things that we need to fix,

but it's quite enjoyable.

So I alternate and I check.

But I've seen pretty much every combination, right?

So people using Brave Shields,

people disabling Brave Shields,

using another ad blocker,

people using two ad blockers on top of one another.

And we have, I mean, that's less of an issue now,

but 10 years ago when we acquired Ghost3,

we were surprised that many people never turned it on.

Ghost3 was somehow set off by default

and was only showing you the information

and not blocking it by default.

And a lot of people never turned it on,

meaning they were looking at the information

and that's why the ads industry was always of heavy, heavy Ghost3 users

because they would observe their own trackers and competitors and whatnot.

but a lot of people were not blocking.

That changed because we made the blocking by default

and then it obviously changes the distribution.

Got it.

So maybe some avatars here to throw your way

and maybe a general suggestion that you have for each of them.

So if somebody's just a regular person,

they have a full-time job,

they maybe have four or five hours of screen time a day,

they're not chronically online

but they do basic web browsing in the evening.

Do you have a general tool or browser extension you push them to?

Well, I think that the first consideration is between are you ready to change browser or not, right?

And some people just can't because, I don't know, the company does not allow them to use anything but Chrome or whatnot, right?

Like there are many, or there's such a vested interest because, I don't know.

I mean, we know, right, that you can import your bookmarks, et cetera.

But some people are just basically married to the browser, right?

And if you are in that situation, I would definitely download an extension on top.

If you are not, I would definitely consider switching browser to something like Brave.

So it depends, right?

If you use an extension, you take the implicit decision that you are not willing to switch

browser for whatever reason.

If you switch browser, you might as well switch to a browser like Brave because there's no

reason not to, in a sense, right?

Because you use a browser that basically looked at certain key problems like tracking and

solves it instead of being just working for Google, which is all the other browsers.

There's really no other browsers.

It's really like all the browsers are basically having a Google deal, which we don't have,

and we have our own search engine.

But if you must use Chrome or you must use Edge or you would like to continue, then you

should use an ad blocker.

And then it depends a little bit on who you are, because if you want to use Ublock,

origin then obviously you can only use it on some browsers because mb3 etc

and then it's a kind of a tech savvy slash nerdy slash very advanced kind of thing

and grocery we position ourselves clearly into the more communicative easy ui friendly

to normal people etc including the zap mode which is then super friendly because

it's a very light introduction to ad blocking with a good experience in the sense that you

click on the button and you go like ah great right i've done something and it's better

so so i think yeah first decision is whether you change browser second is if you don't put an

extension on it i would really not recommend using two extensions on top of another because that's

very undefined behavior there's no real reason but we see it in the wild a lot right so it's

people do that. But it's a bit undefined what happens, I think.

So would you overall agree with these following avatars as well? If you're the family CTO,

as you put it, and you have to install something for grandma to use, I assume you recommend an

extension or maybe even something like ZapMode if that's live by the time this goes live.

It depends how much time you have to spend to teach and how receptive they are to teaching.

And our bet at the moment is that the Zap mode is easy to teach.

And if they don't use it, it doesn't really change what they have today.

Like it doesn't break anything.

It's just the same, right?

Whereas if you install U-Block Origin aggressive mode on grandma's computer and then move to another state, it's not going to work well, right?

Like you know that.

So these are two extremes, obviously, right?

But yeah, it's really that, right?

And as you said, it could be that next Thanksgiving you come back and grandma is ready to upgrade

to full blocking.

Or maybe not.

Maybe it's perfectly okay, right?

Because normal people do not actually go to millions of websites.

Like, you know, they have their habits.

And once it's set up correctly to what they like, there's probably not a real problem.

Got it.

And then the last avatar, you know, someone is looking to do things as optimally as possible.

They want the absolute best privacy protections, but I assume also evidence-based.

So they're not just going to install five ad blockers in their browser.

So what do you generally recommend as like the most evidence-based approach to ad blocking

for users that are a bit more advanced?

Well, I mean, obviously I'm pitching my own products, but if you move, if you change browser,

go to Brave for sure.

and if you can, just put something like Go3 on top of it

because these are the most reasonable ones.

They don't go too far.

They don't break too much.

But clearly, you don't underestimate the complexity of breakage.

Simple breakage is actually easy to fix

and it is generally quickly fixed.

So if a public website starts showing ads

or a public website doesn't scroll,

you can assume that these things will get fixed.

where it gets more complicated is you realize that one of your credit cards only works 30 percent of

the time on amazon right why because in 30 percent of the case they have to call do a call to your

bank and do some verification in the background and for some reason this fails right or you cannot

get your boarding pass on united or you can't book a flight on air canada but you only find that out

at the end of the booking process, right?

So this is, first of all, insanely difficult to debug

because you can't reproduce it, right?

Like the people who do these things do not book flights

just to see if they can optimize the ad blocker.

And then second, it's really frustrating

because you've gone the whole way and then it fails

and you're not sure whether you should blame your browser

or blame the websites.

It's really unclear and not fixable.

And those examples are all real, right?

Like these are not invented.

There are things that we struggle with.

And so if you're the type of person that finds disaggravation worse than a little bit of privacy leak, you know, you have to decide who you are exactly, right?

But, I mean, there are some people who claim that ad blockers never break, but that's just simply not true, right?

That's just simply not true.

Or they just spend time in, I mean, obviously, if you spend all your day on Reddit and stuff like that, you know, eventually you optimize for that.

But if you do transact in the real world and you interact with websites built by governments in 1999 and stuff like this, you're going to hit some really problematic things.

Where you have to disable, you know, not only shields in Brave, but sometimes also more, right?

And things that you, like fingerprinting avoidance, et cetera.

And these may be difficult.

I mean, if you're a specialist, it's not a problem, like you unclick this and that because you understand what it is.

But for normal people, it may be a bit overwhelming to say, why do I have to disable seven things?

And then it behaves like a stock browser.

Then the next thing is obviously the more people use Brave or the more people use Ghost Re or Adblocker in general, the more it is tested against them.

So a website can be very, very lazy and I don't even have to use Brave.

But I mean, a long time ago, people stopped testing against Firefox, which is kind of not great.

because it means that stuff breaks more than on non-standard browsers.

Brave is Chromium, so on the one hand, it's not a real problem.

On the other hand, then Shields could be maybe a problem.

Got it.

Okay, and then to kind of pivot away from the extensions,

who tracks.me?

I didn't know this existed until I met Adam in Cypress,

and I really like the tool.

The concept here is that it allows you to get insights into websites, the trackers on those websites, what they actually are used for.

So what was kind of the inspiration behind that tool?

And, you know, how does it feed into that invisible and visible mission?

It comes from a long history of tools.

Unfortunately, we cannot replicate these tools that used to work very, very well.

And we would we called it tracker map, which was a you would just enter any website and any location in the world.

and it was like a movable map of all of the trackers.

So you would see it loading, let's say, some tracker,

and then that tracker would load other trackers,

and then these trackers would load other trackers,

and you had this really immense tree that was moving.

And every time you would show it to people, they would go like,

what? Is this what's happening when I load this page?

And you go, yeah, exactly that.

We cannot do it anymore because browsers have become a lot more complicated

and it's difficult to say exactly who has loaded what

and so we cannot build that tree anymore.

But what we can do is to tell you

this tracker is present on this website.

And then if it is there,

is it also in this other website?

So you can see the visibility that they have of your behavior.

And then it turns out that universities

and a lot of other people are very interested in this data.

So we opened open source stats.

And we spend a lot of time putting names on trackers, right?

Because trackers, when they load a request, they don't have a name, right?

Like you don't know what the company is called and you don't know who they belong to, etc.

And we spend a lot of time mapping that to entities, right?

And by the way, they change names all the time.

But you had companies called Blue Kai and then suddenly it was owned by Oracle.

And you would not know that by just looking at a tracker that goes to bk.com, right?

And so we tried to link that to websites.

And that's also in the UI of the grocery, right?

You can click on it.

And if you're so inclined, you can actually learn a lot about who's doing what on the website.

And that's also why the ad industry was always a heavy duty user of grocery,

because they could see, like if you work for a website

and you try to see all the trackers that are running on your website.

Or let's say in Europe a few years ago when they had to implement GDPR notices

just to find out what was running on their own website.

Like how it's retroactive.

Well, yeah, unfortunately, because it has been built over the years

and suddenly they had a requirement to list that properly

and get the privacy notice correctly, etc.

So they had to do an inventory.

And one of the best ways to do inventories was to use something like Ghost3

or look at who tracks me.

And a few audit companies use it, so they download the data and help us keeping it up to date because they do audits.

I'm not sure exactly what kind of audits, but they do audits of websites and probably for GDPR or stuff purposes to see is your privacy policy in sync with the actual reality of your website, that sort of thing.

Yeah, it's not, obviously it's not a consumer product, right?

It's not something that, well, it's interesting for half an hour for a normal consumer,

but for the industry, it's quite interesting.

Yeah, no, I definitely see the value for the industry,

but I would almost push back a little bit

because I think that for a consumer to actually see what these things do and are,

that is still extremely valuable.

Yeah, what I mean is that most people see it and then block it, right?

And then once you block it, then obviously it disappears, right?

So this is one of the weirdness of the UI that we are struggling with,

is that if you block everything, it looks like the numbers are low.

It's only when you unblock everything that you see, oh, God,

like this is a lot of trackers, right?

So sometimes, I mean, obviously people don't do that,

but you can disable Go3, and then you watch the tracker con go up,

and then you click on the info, and then you see them all.

But then again, I've never met someone who said, oh, I'll leave it off, right?

Then they could block it again, and then it all disappears.

So I want to dive a little bit more into these trackers and how they work on a technical level.

You used a really good analogy with the men in black earlier, following you around.

But what is a tracker itself on a website, right?

I think people hear this term.

I use this term.

The whole industry uses this term of tracker.

But what is a tracker?

So one big distinction that we make is first party versus third party.

and I simply mean that if you're in the domain cnn.com, right,

everything that goes to cnn.com would be called first party

because it belongs to the same organization.

So if you download an image, for example, there will be a request.

Because, well, first of all, people don't understand how a web page is built.

It's not like a page in a magazine.

It's more first you get a bunch of HTML that will mostly contain the text,

and then you're going to download a bunch of pictures,

And then you're going to download a bunch of code that will maybe create a poll that you can answer and get the answer, etc.

So there's a lot of files that are being downloaded to create and render that page.

And some of these files do not go to CNN.com.

They go to something else.

And something else simply meaning another company.

And that could be for different purposes.

Some of them can be super legitimate.

Like, for example, it goes to a CDN, which is a content delivery network, because let's say there's a video on the site and Cloudflare does a better job at delivering that quickly to you.

So it will go to a Cloudflare domain.

Although, you know, people have become much smarter.

So the Cloudflare domain will look like a CNN domain because it's proper to do it that way.

But sometimes it's going to be companies that will basically just record, oh, this user, which I know because if you plant a cookie on a user's browser, it is connected to that domain name that plants it.

So CNN can plant a cookie and recognize you when you come back and maybe personalize it for you.

But also these advertising companies can plant a cookie on you.

And every time you go to CNN, they will get, ah, Henry is reading this article.

And then you go to another site.

Oh, Henry is reading this now.

And then you go to Amazon.

Well, Amazon does not have third-party tracker, but I assume you go to a shop that is a bit more naive.

Oh, now he's shopping for this, right?

Oh, he's been looking at gloves now on this website.

Then next time he goes to CNN, that company will give the information to an advertising

server that says, oh, Henry is in the market for gloves because we just saw him 10 minutes

ago on the gloves website.

So maybe we show him gloves, right?

Or he's looking for a trip to Paris.

And then suddenly your whole web is filled with ads about Paris.

And you wonder, how the hell do they know?

And the mechanic is simply these third party things, because third party means that they

can be on many different websites as a third party.

and they can plant cookies so they actually know who you are.

The cookie may not say Henry,

but the cookie is a number that is basically you, right?

And it doesn't take a genius to link that to one single person.

And what incentive, right?

Like why?

Because if I build a website, we have a privacy first community.

So I actively try to choose services that will not,

like we use Ghost for our new website.

It's open source and they don't use trackers by default.

But when we embed a YouTube video into our website,

technically it does show as a tracker.

because it embeds the YouTube video and then YouTube serves ads, et cetera.

So why does someone like CNN embed these trackers?

What benefit is there for them?

So you have to think of it in terms of organizations, right?

Because it's not CNN.

There's not a CNN.

So at one point in 2012, some salesperson came to a BD person at CNN and say,

if you put this tracker, you can get this dashboard that will record,

that will give you this benefit.

which can be statistics or counts or a widget.

I mean, let's take Facebook.

So Facebook at one point started to litter the like button

all over the internet, right?

So the like button was a very interesting thing

because it was a little picture, right?

And if you clicked on it, you would like it in Facebook,

which some people used to do.

Not anymore, but it used to be a big thing.

But every time, like, you could assume,

okay, that thing is cool when I click it,

but it was also very interesting for Facebook

when you didn't click it

because they would learn that Henry,

because they knew who you were,

like they knew he was Henry

because Henry had a Facebook account at one time.

And they would know you

because they had a cookie planted

with a Facebook.com domain.

So when you were in CNN

and the like button basically appeared,

they knew you were reading that article.

Like the like thing was a complete side effect,

Like maybe 0,1% of the people would click on the like button, but 100% of the people would be tracked.

So Facebook built this amazing amount of information about your preference by following you around with these like buttons.

And so why do they do it?

Well, obviously they do it because they wanted to have a huge profile of you and then do ads and other stuff.

Why does CNN put the like button?

Because they want it to be liked on Facebook, because it gave them traffic.

So that's actually a pretty good caricature of what's happening.

the provider the third party comes and say i have a widget i have a thing or it's going to make you

more money it's going to make you more likes it's going to make you more this or that and and whatnot

and then some business development person says yeah okay that helps my goals and and i'll put it on

and then these things just stay for years and years and years i'm pretty sure you can still

find like buttons on facebook like buttons on some pages but now it's not a like button it's going to

be a tweet thing or it's going to be follow me on Instagram, et cetera. But all of these assets

that you see, it's not when they click, that you click on it, that they become active. They're

active even when you don't click on them. Got it. And then you mentioned earlier, you know,

the example of Amazon and how Amazon might not have third-party trackers built in. But how,

and maybe this does happen, but users might hear this and go, well, okay, so if everything is

blocked on a website, then there is no way for there to be data sharing. So is that true? How do

we know that Amazon can't have, I guess, ways of still submitting user data to the same companies?

Let's take the biggest tracker of them all is Google, but you will not find a tracker on Google.

And Amazon, you will not find a tracker on Amazon, but they will track you wherever they can with the

Amazon ad network, right? So these companies have learned that the economics of data is about

collecting and making sure that other people don't collect on you right so you you i mean if you go

to facebook there's not gonna be any trackers but they track you everywhere else uh so so the

i mean there are other ways to do data sharing and it tends to be probably a little bit better

in terms of privacy but but basically there are people who understand the game and they are on the

collecting side and they don't share and there are other people that are on the sharing side but they

They don't benefit much from it.

Yeah, it's an interesting world.

And I never understood why they participate, but they do.

Because they like that like button.

And they like to have some Google traffic extra.

Because they got Google Analytics, for example, is a huge tracker, right?

And Google Analytics was basically, if you go to the beginning around 2000,

well, it was a company called Urchin, but they got acquired by Google.

And people really loved to have a free analytics tool just to know how many visitors they had.

But the side effect was that they were giving all the data to Google, right?

And so Google could know, for example, let's say you had a cool blog about technology and your competitor had another cool blog about technology.

So both of you would know what your traffic was, but Google would know what the traffic for both of you were.

So obviously it's not balanced, right?

Like Google knew way too much because you gave them the key to your data.

The power dynamics are pretty interesting to think about and who actually has the power in these relationships.

But the rule of all evil is convenience, right?

So people say, well, I don't have enough money to create my own analytics, so I will just

take one for free and I won't pay for it, right?

Because Google, Urchin used to be something that cost money.

Google Analytics did not cost money, so they made it free right away.

And then the other thing is when Google AdSense came in, which had a small role to play in,

also the same thing, like I'll just, I need some money, so I'll just put the Google AdSense

tag on my website and then suddenly money flows.

there's tons of side effects to that but the the primary benefit was i get something for free or

even something that pays me and i don't care about all the side effects so google probably the best

in the world they basically realized that people were driven by convenience or web publishers or

or people who build websites were driven by convenience and started to to give them the

drugs right say hey you want traffic we get google search you want analytics we got google analytics

61 monetization, we got Google AdSense, right?

And the side effects was basically that Google would know everything about your website and

your users and then would use it for their own purposes, obviously.

Wow.

Yeah, really well said.

And then just a couple of final technical questions about the who tracks.me webpage.

First, how do you figure this out?

What's the investigative process looks like?

do you find this domain and then you have to figure out like the certificate authority and then

well so yeah exactly so basically when we when we do our work which is basically uh you know

observing the traffic that goes out to third-party domains on a given web page we basically catalog

we basically count right we can't oh there's 17 requests to this domain and that domain etc

then either we already know the domain so we know exactly if someone has taken the time to take a

look at this, or we go into a little investigation.

We say, who owns that domain?

What does it look like it's doing?

Is it a CDN? Is it something?

And then we put it into categories,

and then we find out who owns the company

by looking into, basically, it's web search, right?

And AI is actually pretty good at helping in these days,

trying to help you.

And then we just put it in the Who Tracks Me database,

which then benefits all of our users

and all the people who use the WooTrackMe database or products.

Very nice.

And then you mentioned this tree image of how it used to work.

Just to clarify, is what you're saying that tree still exists?

It's just hard to verifiably prove that's how it works because of browser technology?

Yeah, it's harder to build because in the old days,

whenever you saw a third-party request emanating from the browser,

it had a clear parent, meaning that you could tell who was requesting it, right?

So let's say you land on CNN.com and it calls Google Analytics.

The Google Analytics call would say CNN.com did that, right?

So you could say, okay, there's a connection.

But then you will have more complex things like T-Alium,

which is a kind of tag manager or Google Tag Manager for that matter, right?

So a tag manager does is call a bunch of other things.

And in the old days, you could always say,

ah, this company, Blue Kai, has been called by Google Tag Manager.

But in the meantime, the browser is much more efficient

and tries to do a lot more preloading.

You have a lot of optimization techniques within the browser

that doesn't allow us to say,

it was definitely this thing calling that thing.

So then you cannot build a tree that you can 100% guarantee

is exactly what happened.

Even partially between websites,

because sometimes you're in Google search

and it's already preloading the next page

because it thinks that you're going to click on the first result.

You have a lot of optimization in browser to make it fast

that don't allow you to create a perfectly great tracker map.

But the tracker map was an impressive thing to look at.

You would look at it and say, I don't like this one bit.

Because people are still stuck in the mental image

that the web is like a magazine that you turn to a page

and the page is one item.

Like they don't imagine the thing being the combination of 100 requests to the Internet.

Because then the next question is like, why 100?

Why not three, right?

And that's a good question because why do you need so many requests?

You don't.

You need one request for the text and then one request for each picture.

And then maybe a couple requests for some JavaScript to do some animation.

And then you run out of ideas.

So why are they 100?

Well, there's 100 because there's Google Analytics and then there's this.

And then there's Tialium and then there's Tag Manager.

And then there's this and that and that and that.

All the stuff that if you use an ad blocker, you can block and you see not only no difference, but actually everything goes faster and it's a lot cleaner.

Very nice.

And then, you know, kind of where I'd like to finish off in terms of categories is search.

So on the GoStory's website, I saw there's a private search tab.

Is that just Brave Search?

Like, what's the difference between that and Brave Search?

It is using the Brave Search API, which is kind of, what do you call it?

Well, it's a search API, so it allows you to turn it into whatever look you want, right?

Including some of the biggest AI companies in the world use Brave Search API to help create whatever AI companies do in their answer.

So it's a really huge product.

Claude uses you guys, right?

Claude?

This is one of the things that we cannot talk about officially, but people look at the privacy policies and they figure it out by themselves, yes.

So, yes, there's lots of big companies that use Brave Search API.

There's only three big indexes in the world, right?

There's Google, there's Bing, and there's Brave.

So these are the only three that can answer every question at a level of quality which is nearly the same.

And Google does not license their search results that easily.

Some people find loopholes in how to get them.

Bing has stopped licensing it in August last year because they want to keep it all to themselves.

And Brave is the only one licensing it to pretty much everyone who wants to have them.

And private search is important because, coming back to the grocery private search,

is if you really want to be private, it's very tricky to search in Google

because Google will basically put your name on every search that you do.

and even if you say it's private,

they will know that you did this search.

And they probably know every search I've made in Google

for the last 25 years.

I haven't used Google Search for probably the last five,

maybe even more, because of Brave Search,

but it's hard to care about privacy

and use Chrome and Google Search, let me put it this way.

Or basically, your privacy stance means

I don't trust anyone, but I trust Google totally fully,

which is, I guess, reasonable if you want to.

but it's hard to be

to say Google is the biggest tracker in the world

but that's okay because they're the biggest

so I'll keep that one's okay right

that's a bit contradictory

and Brave for example never

and obviously not grocery as well

we never associate a query with a person

nor do we keep track of

any kind of identifier between queries

so we just don't know what Henry did yesterday

and what Henry does today

and even in five seconds after you do one query

and you do another one, they are not connected.

Got it.

And so the terminology I've used and I've heard

is a meta search versus like an independent index.

So something like DuckDuckGo uses a lot of Bing.

And then we have like Startpage uses Google.

And I feel like it's pretty well established

what the benefits are to using even those services,

which is they are, you know, it might shift trust,

but they are guaranteeing a layer of privacy by shifting the trust to them instead of the native company.

There's a trick to that because it breaks with advertising.

So if you use DuckDuckGo, that's fine.

I think the claim to say they are not tracking your queries and they're not associating it with your name is totally there.

But the problem is they also use the Bing ads and the Bing ads will try to track you to see if you convert on the websites that you go to.

So the privacy promise you have to...

So by the way, the web was remarkably private

until advertising became programmatic.

Like Google Search was actually, I would say,

very private until very late, 2012, 13, 14.

Because the only component to targeting and advertising,

I would say even up to today, is whether the query that you have is relevant to the ad that you see.

So if you look for lipsticks, they don't care if you're a man or a woman.

I mean, that absolutely plays no role in the targeting.

Whereas, you know, if you look at the more Facebook type of advertising that started around 2008 or 2009,

it's all about who you are rather than what you are looking for.

And then these two things started to, unfortunately, merge a little bit too much, and tracking started to go completely out of hands.

There's a reason why Adblocker started in 2008 and not in 2002, or anti-tracking things.

But in any case, if you use Brave or any private search from Ghost3, so Ghost3 is a little bit, Ghost3 private search is a little bit more like DuckDuckGo, I guess, because it uses another index.

Brave is the only one where you know your query is not getting anywhere else than Brave.

Yeah, I would like to just clarify here the nuance.

So because if I read DuckDuckGo or StartPages technical information,

they kind of lean into contextual advertising as their main way to advertise,

which means if I look for baby diapers, then it might serve me ads for baby diapers,

but it won't conclude that I'm looking for baby diapers because of an unrelated search term.

So it only uses that one search term.

So where is the privacy concern come into that question?

And how does Brave Search compare to that model?

Then it becomes like you have to look into the details, right?

So if you have one of these search engines,

typically they will load the ads from the client side.

It depends if they load it client side or server side.

But what they will all do is to try to track you after conversions, right?

Because they need to, the customer that buys the ad for a lot of money,

because search ads are very expensive,

needs to know that the conversion that you received

actually emanated from the click that was paid to, say,

Dr. Go, Start Page, or whatever.

And that's where there's a privacy concern,

because then you say, well, we are an ad blocker

and we block everything,

except we don't block what is a conversion channel.

And that's a problem.

Now, the problem is less important in recent years than it used to be because pretty much everyone, including Google, is recommending big shops, etc., to do attribution by first party only, meaning that they should not use external tools to do their conversion tracking because these are likely to be blocked.

And then, you know, in countries like Germany where you have 40% ad blocking, then obviously all these things get blocked, you know, and then you don't really get a good attribution anyway.

But it was just a note that if you control everything, I think you can demonstrably prove your privacy stance.

If you don't control because you use yourself third party to do your search, then obviously it's hard for you to prove that you are doing a safe job.

And I don't want to criticize all of these meta-search, as you call them, right?

It's a perfectly legitimate type of business.

It's just that when I founded what became Brave Search,

it was very important to me to not go that route

because once you go that route,

I think it's nearly impossible to build

because it's too convenient to continue using that third party, right?

It's cheap.

Sometimes it even makes money right away.

And you have different versions of that, right?

Because Google Search in Firefox,

you could call it Firefox Search as well, right?

but it's very difficult for Firefox to remove themselves from that relationship

or Safari from Google relationship or Doug Doug Go from the Bing relationship.

It becomes nearly impossible after a while.

And we took the very strong decision from day one

to start to build everything from scratch and everything by ourselves

without external dependencies.

I wouldn't say that it was for privacy reason, right?

Like, I don't think that was the motivator.

it was more for engineering pride or something like it's like we might as well do it right

and then it became clear that if we were when we started to build a browser which was around 2016

around the same time where we acquired ghostry we were like well if we build our own browser we're

going to build it with privacy in mind because we've seen you know because when you do search

you collect a lot of data you you look at a lot of data and you go like oh that's pretty bad what's

being collected. I don't want to be part of this. And then you start to think about privacy.

But the main driver for what became Brave Search was purely engineering pride and the desire to

build it all to be fully sustainable and independent. Got it. And then I do want to

quickly touch on the indexes. I only have a couple of questions left here, but I did just want to

briefly ask brave search ads how how are those served and how is privacy respected in those ads

and your version of those well in in different ways right so first of all they are only targeted

by the query so we don't there is no sense of oh this user is this and therefore they will get this

ad versus this other ad so that's that doesn't happen at all the second thing is we don't create

the profile, so we don't know who has seen what ad, know who has clicked on what ad.

And there's a limitation to our customer.

Unless they use first-party tracking, they will not be very happy with Brave ads because

they will not be able to attach the attribution to the click, right?

But luckily, most of the biggest spender in search ads have learned that lesson in the

last five years or so and have all pretty much very solid first-party tracking.

which is really not a problem, right?

Because if you go book a trip at booking.com,

obviously, you know, at one point you're going to register

and you're going to have your flight ticket under your name

and not someone else's name.

And the only thing that they do is they plant a booking cookie for you,

you know, which is not tracking because it's not tracking you

between booking and then CNN and then your doctor's office and whatnot.

So we have only rarely issues with customers

when they use third-party tracking,

and this is mostly very, very small customers,

but we don't, like for Brave, it's not a good match

because if you have 1% or 2% market share,

it's very difficult to target very small customers

because for them it's going to be very small as well.

So we function very, very well with huge customers

and not so much with small ones.

I was just going to ask, why are there no major indexes?

There's Google, Bing, and you guys.

It seems like the major ones.

So what are the challenges?

why did it not take 20 billion dollars for you guys to do it so first of all i think the main

issue is that you really really really need to be passionate about building it because

because you can go get it all for free right so so in 2008 if you wanted to do a search engine like

ecosia quant the go more or less all appeared in these years right 2008 to 2016 they were you could

just pick it off the off the shelf there was even yahoo was doing something called boss which was

build your own search engine right so so you could get something for completely free and then you

could focus on the fun part of packaging putting your name on it marketing all the soft skills

right so so you need to find crazy people like me and joseph pujol and other people that basically

said no no i really want to build this stuff i don't want to just get it and paint it which was

crazy because no vc would finance it because they were like come on you're going to compete against

Google, that's crazy, right? So it was essentially mission impossible. And I would say that only in

the last year, it became, I think there's like five companies in Silicon Valley that I know of very,

well, plus all the big major AI lab, they're all pretending to build search, which is ironic to me,

because you do something for 18 years, and everyone thinks you're crazy to even think about doing

something like that, because you can get it for free without doing the work. And then suddenly,

everyone needs it. So, you know, I don't know if you can call that good timing, because 18 years

is pretty bad timing, I would say. And very frustrating, but it is very difficult, because

you need to essentially learn all the web functions, you need to learn how people talk to the internet,

you need to have enough users to see enough queries, but unless you have a good search engine,

you cannot, you know, why would they use you? You have to solve an insane amount of problems.

and for most of the time without any economic benefit

because neither VCs nor big companies like Apple would,

and they were always respectful, right?

So the nice thing is that everyone knows

that building a search engine is hard.

So you would get a lot of sympathy from everyone saying like,

wow, this is really courageous and really hard, et cetera.

But you would not get that much money.

Except now, right?

Now everything has changed about a year and a half ago.

That's the realization that every AI needs a search engine to talk to.

Yeah.

And it's quite fascinating, too.

And we're starting to see this play out.

I think the tide, there's been a huge messaging shift in the last month where beforehand ChatGPT was kind of seen as the king of AI.

And now in the last month, everybody's convinced that Gemini is kind of the way forward because of Google stuff.

But now it's, you know, Google is probably making it very hard to access their search.

And so Google's reserving that for their own AI and all these other AI companies now have to scramble.

You have a lot of movement.

So Google has decided to sue SERP API.

SERP API is a kind of a company where you can send them a query and they tell you what Google showed, right?

Because they will just basically query Google from some botnet or something.

And so a lot of the big AI were using that.

I don't want to name names, but they were using it.

So now they're on the, at least there's some stress from Google because you can feel that Google is not essentially, will not make it easy in the future for people to, for a machine to get Google results by just acting like a user.

But they can get it, exactly.

And we've done some internal experiments at Brave using many, many thousands of queries to check how much of the, like when you get a really good answer, how much of that is coming from my AI is super smart versus my data is super complete, right?

And it turns out that the data is super complete is quite a bit more important than the smartness.

Both are important.

But if you take the best model, let's say something like HRGBT or Entropic, combined with a SERP API type, I just get 10 blue links and a few snippets, it will not beat Brave Ask or XAI, which uses Brave quite a bit, because they get the full scope of data.

And Google is, and unfortunately this is a game that they know how to play, can increase their quality on demand.

So they can decide to ramp it up or ramp it down.

And they just decided to beat ChatGPT at the moment by a little bit.

But they could go a lot more.

But obviously, it costs a lot of money because if you start to ramp it up, something that used to cost them probably a few cents per thousand is starting to cost them like a few tens of a cent per thousand, right?

And the economics are not great.

And we've seen that in different countries.

Like I ran or I'm still running a search engine in Vietnam.

And as soon as we were getting good, Google became better.

And then we would get better.

And then Google would beat us again.

like five seconds later.

Like they did it in Turkey with Yandex.

As soon as they have competition,

they know how to ramp up the quality.

And so that's going to be the name of the game,

I think, in 26,

is that people are going to realize

I need a really good search engine,

sorry, a really good AI

combined with a really good search engine

in order to beat Google,

which has basically both, right?

Right. Really well said.

And it's going to be interesting

because, you know, two years ago, I remember, I mean, I'm friends with a lot of these AI guys,

and I wrote an email to probably one of the most famous one on the 31st of December 2022,

saying, you're going to need a search engine. And they were like, yeah, maybe, maybe not, you know.

And then at the end of 23, it was like, yeah, probably. At the end of 24, it was like, oh,

shit, can we get access to your search API? And so the realization is that the AI needs contact

to the world and a contact to the world is first to the search engine and maybe later to the browser

because then you bring user which is the user agent is the browser is called the user agent and so

the user the open web the ai and search come together and unfortunately google has all the

pieces already so the other ones have to kind of put them together i say unfortunately because i

don't want google to win again everything right yeah yeah well i i think that you know when you

put it that way really demonstrates the importance of what you're doing and why there's a lot of

dependence on it now because people need access to that data and they're not going to get it from

Google and Google doesn't want to share it. And I've seen this myself. I think people can actually

test this themselves if they use even the highest quality model. Like if you, you know, go to the

highest quality chat GPT model or use Cloud Opus, but you just, you know, paste 10 URLs into it and

rely on it to fetch data from the URLs. It's horrible. But if you actually copy and paste

the article content into the AI model,

you'll actually see it's far better

in terms of its performance.

And it won't hallucinate.

You can simulate that.

And since you use Brave Search,

I don't know if you know the trick.

I mean, you can click on the ask button,

but I don't do that personally.

What I do is I add two question marks

at the end of queries

where I want a more AI experience.

And then it gives you the AI experience.

Try it.

It's fantastic.

Because you can tell, right?

Sometimes you say,

well, I just want the link

because I'm looking for the menu of that restaurant.

I don't want an AI to summarize the menu or whatever, right?

But sometimes you say, hey, how do I reset the tire pressure monitor on a Toyota some XYZ?

Then you put two question marks and then you get like a detailed instruction,

like, you know, just make sure your tire pressure is okay.

Go to here, go to there, do that, do this, right?

I didn't know that existed.

Is it like a search bang for like triggering the AI response?

Yeah, because you have different modes, right?

You have normal web mode.

You can go to images or video or news, you know, the usual modes.

But you also have the ask mode, which is this experience, which is AI experience.

And you can click that after you type, right?

But if you already know you want an AI experience, just put two question marks and you'll get it.

Got it.

And out of curiosity, on the opposite end, because now I know this, that's actually very, I just tried it out.

I just typed in hello there and did the two question marks.

And yeah, it just goes right into the ask mode.

Do you guys, I know DuckDuckGo has like a no AI URL if people don't like the AI at all.

Do you guys do something similar with Brave?

Yeah, well, we're aware that some people just don't like AI, right?

So we need to provide some, you know, opt-in, opt-out, et cetera.

And we still sort of, I mean, at the moment we have two types of AI in the browser.

We have the thing that comes in search, which is mostly trying to guess whether it's a normal 10 blue link kind of thing.

or whether it's a thing that should have a little summary on top,

or it's a thing that should get full ask mode, right?

Which is the two question marks.

I don't think, I'm not sure whether we have a place

where you can opt out of these things,

but it's been discussed for sure,

because some people just don't like it,

but it's a very small minority, of course.

And then we have Leo, which is a thing that you trigger,

so it's by definition opt-in, right?

You go like, I want Leo to take a look at my webpage,

so that never happens automatically.

But it's clear that there will be a subgroup of people that basically you don't want to force AI upon.

I don't think it's wise for any company to optimize specifically for these people, except by giving them the choice, right?

Because if not, you end up with a pretty strange target group.

And we know that the future is, for better or for worse, I don't think you can take intelligence out of the world, you know, artificial intelligence out of the world when it's in the world.

Like, it doesn't work that way, right?

But we're aware that some people don't like it and you get some pushback.

But, yeah, the question is really what to do about it

and how much resource to invest into providing tools for these people.

But, frankly, I've never really met someone who doesn't like these things.

I mean, these things that we do in search, right?

I've met people who don't like ChatGPT or don't like other apps

or like AgentTik browser or cannot imagine that the browser

will buy them tickets etc which is by the way also something that i would not do because i don't want

the browser to do something slower than i can do myself you know that sort of thing but i just heard

an example from a colleague that he that was doing some research or some building permit research in

the uk which apparently is a very tedious process of going from website to website and he just asked

the agent take a mode of brave to just go and do it for him and that worked out very well and that

I can imagine, right? A very tedious process that you have to do, but don't want to do.

Booking your vacation, in my experience, people actually like doing that themselves, right?

Ordering food or something is not good examples because people enjoy the process, but,

you know, filling your taxes maybe, or collecting, filling documents that are totally tedious and

completely not relevant, maybe, right? Right. So, what's next for you? What are you looking

forward to in 2026 that you're able to share for ghostry or brave or anything else well i think on

on the on the whole you know growing these products because we believe both on the brave side and the

grocery side that is definitely not reached the limits of what we can do by far and second i think

more on the ai part is to discover what is the true killer use case in the browser for ai

because we have some ideas,

but all of the things that are floated at the moment,

like buying your trip automatically and stuff like this,

this is not real.

This is not changing the world.

And it also misunderstands what people do with browsers, right?

What normal people do with browsers

ranges from watching porn to watching sports to betting

to doing a very enjoyable thing, let's say.

I don't see why the summary of an enjoyable thing is enjoyable, right?

And so we need to move past that and figure out what is the web of tomorrow, what is the open web of tomorrow, rather, with AI in the world, right?

Like, is it a world where every small service provider, like a hairdresser, is able to create an incredible website that can talk to man and machine equally good and you can make reservations and pay and do complicated things?

And the hairdresser does not have to spend $25,000 with some kind of provider that screws them and doesn't do the work right.

You know, like the whole process of building the web used to be very easy 30 years ago because it was that simple and couldn't do anything anyway.

Right. It was very basic.

And then it got ultra complicated and it became, unfortunately, too difficult for most people.

And then they started to go into, oh, I'll create my Facebook shop or my Facebook page.

And then that became Instagram.

And now it's maybe TikTok.

And I think we need to find a solution as a browser to say, no, we want every single business person, influencer, whatever, to be also on the open web.

And to be equally accessible to man and machine, depending on what they do.

Because if you can say, hey, book me an appointment at the hairdresser for next Wednesday to your browser, and then it can basically interact with the hairdresser site.

And then say, okay, confirm or say, well, Wednesday doesn't work, but Friday at four

would work.

And I just checked your calendar and it's actually like you're free.

Then you say, okay, you know, that starts to make sense.

Yeah, this goes back to the convenience point you brought up earlier, because as a publisher,

and this goes for any video content creation, the default is YouTube.

It's free.

And not only is it free, but if you get enough views and engagement, there's ad revenue as

well.

And there's no maintenance.

You don't have to have any technical experience or knowledge to get started on it.

But that's the default nowadays.

And same thing goes with blogs, everything out there if you're starting a business.

And it takes so much active effort to get away from those systems.

I cannot tell you how many hours I've wasted.

It's not wasting because it feels like wasting in the middle of doing it.

But just maintaining a pure tube server.

I'm not extremely technical with this stuff.

But weekends, it just randomly breaks.

The auto sync is totally broken.

I ripped my hair out just trying to maintain this alternative platform.

And I think it's worth it for our audience because I want something to be accessible away from Google.

And that's open source and that we host the videos ourselves.

But I can't blame publishers in this current state.

Of course.

And I think the web has become, you know, before 1999 or so, it was not really a publishing.

Like it was not SEO, for example, was not a thing.

Or even advertising was not fundamentally a thing at that time.

And you would have a lot more websites that had other interests than making money from ads, right?

And then it became this loop, search, SEO, brings traffic, make ads, make money, and then, you know, you create a gigantic washing machine.

And we have to sort of reinvent that, right?

We have to find out the motivations that people have to build open web sort of presence and the economic incentives to do so.

Now, if you're a hairdresser, it's clear, right?

you want people to come and get a haircut if you sell words and pictures then you will have to have

some kind of payment in advertising or something else right and and i feel that we will have to

reinvent all of that because it's all being disrupted like the the old search model of

you know just just optimize your search engine for the queries of my users and then you will

get traffic and then i give you analytics and i give you uh financing is probably broken not so

much on the video side, I think it's not broken at all because it's about time of spent, right?

It's not the same to listen to this interview as to get a five-sentence summary of it, right? It's

not the same. But for a lot of written content, maybe unfortunately for the content, the summary

is same or if not better than the content itself. And so that's why video and entertainment in

general is a lot more, you know, you don't want to, you want to watch the football game. You don't

want to have a summary of it, right? It's not an AI. I mean, maybe AI will create football games.

I don't know, but are they going to be as exciting as the real one? Probably not. Right. So anyway,

we, we, we have an interesting 10 years ahead of us because we're going to have to reinvent

pretty much everything. Wow. Okay. Well, thank you so much for taking the time to be here. I know I

kept you for a long time so i really uh thank you for you know that commitment and how can people

follow you and what you're working on if they want to check you out and we'll be sure to leave

links in the description so that's an interesting thing i'm remarkably unsocial so i don't really

tweet i don't do linkedin i don't do instagram i don't do many of these things but um just just

keep using brave search and that sort of thing and use ghostry if you want to stick with chrome and

you'll be in good hands you'll be benefiting from whatever i do without me talking

Or you invite me back once a year and we talk.

Yeah, I'd love to have you on.

I mean, I'd love to see, you know, maybe every January what you kind of are looking forward to and what actually happened in the year prior.

We can do that.

Yeah.

Well, thank you so much, JP.

Thanks.

There you have it.

I want to thank JP for his time and for being on this podcast.

I definitely appreciate everything he did and all the information he shared, which came from clearly years of experience.

I also want to thank you all for watching and showing up and taking your digital rights

journey seriously in this day and age.

You being here is also empowering those around you as well.

Now, if you want to be more direct, definitely share this podcast and share the work that

we have with your friends and family.

And also make sure you're subscribed and leave a rating and or leave a comment or a like,

depending on which platform you're on.

All that engagement really does help grow this podcast so we can get this information in

front of as many people as possible.

So I want to thank you all for being here.

If you want to really support what we're doing, you can become a Techlorian down in the description,

which is a paid membership to support what we do.

Otherwise, I'll see you guys all in the next one.