How Kagi is fixing search w/ Vlad Prelovac

To kick things off, why don't you just tell us a little bit about who you are, a little bit of your background and.

And maybe a little bit about what Kagi is as well.

And that has allowed us to basically build a search engine that has best search results in the world, the best search experience in the world, and people are willing to pay for and also have privacy as a cornerstone of our service because there's no incentive to do anything but user data. Actually, user data is liability to us. It's of no use. So I personally been working in the tech space for the last 20 years and previously had a company in the website management space.

We managed about half a million of websites doing security updates, backups, and things like that. It sort of gave me necessary domain expertise to go into a little bit wider problem, which is the problem of search. And Kagi was launched to public in 2022.

Like, of all the things that you could choose to focus on, like, you seem to clearly have a desire to preserve privacy, to give people good user experiences without the kind of perverse Monetization that we've seen take over the space. Why was search the thing that you wanted to jump in on?

And it just hit me really hard that we are consenting to that sort of state of things. And there was really no alternative. All search engines out there were ad supported in one way or the other. And I did pay for YouTube Premium for her and her siblings, but there was nothing similar you could do in search. You could not pay for a service that will be user centric and have your best interest in mind. And at the same time I just noticed that Google search was getting really, really bad.

People talk a lot about it today, but even six years ago was getting worse. And what really triggered me in a way was I think I was searching for something like GitLab or Target and the poor company had to buy an ad on top for their brand name. And then there was the organic result right below it. But if you don't do it, somebody else just buys your keyword and people are misled to your competitor. And it just striked me that that was incredibly intelligence insulting for me as a user.

You're showing me two identical results. And this is when I understood that the search engine that we are all using is not built in my best interest in mind. It's built with, you know, advertisers best interest. And so just having just sold my previous company, having some funds, I decided to tackle this problem. I had a fairly good idea how to do it.

I knew from day one that the only way to do this right is to have a paid business model which will automatically align all incentives and actually incentivizes you as a search engine to innovate and create better search results, better search experiences, otherwise the people will not pay for it. And you know, especially there are so many free alternatives outside freeing air quotes. So yeah, that's how Kaggle started. I assembled a small team and bootstrapped the whole thing.

And in June 2022 we launched and had thousands of paying customers in month one. And it just kept growing from there.

But when you think about the way that the incentives actually work, they're incentivized to serve you the results that they're paid to serve rather than the results that you actually need for your search. And I wonder if that's a lot of the reason why it's just become widely accepted that Google's results are just absolutely awful now.

And that's something like even my non technical, non privacy loving, they don't care that Google is an awful company who abuses data and ad monopoly, all those things, but they just realized the search sucks and I can't use it anymore. So it's interesting, I wonder if that perverse financial incentive is what has kind of led to their downfall in the actual usefulness as a tool.

They're just good enough to be, you know, trying to be better than anyone else is what Google, you know, has as a goal, basically. And they are optimizing for what puts money in their pockets, which is advertising money. Right. You're not putting the money in their pockets. As a user, Kagi has it completely inverse. Like one of the most important thing for us is time to answer. Like we want you to spend the least possible time on Kagi.

We want you to find what you're looking for and talk to your results most of the time. And you know, anecdotal evidence is that that is actually the case. And we had some people who counted their search history in Google before they moved to Kagi and they reported halving the number of searches they do in a month, for example, like they would go from thousand to 500 just because you're able to find things faster.

And really it's the alignment of incentives is if we are not able to do that, Nobody would pay $10 a month for Cognizant. So that alignment and incentives is what drives creation of a user centric product. And Google is user centric also. Or should I say customer centric also. The only problem is the customer is the advertiser obviously.

But is there anything else that kind of jumps out at you that you feel like separates Kagi from especially like I'm thinking especially for like the audience of opt out separating you from things like DuckDuckGo or the Self hosted options like C and Google or Google, these ones that are at least a little bit more privacy preserving if they're not necessarily fixing anything wrong with the actual results being served.

We actually have a project called Mole Web which is a collection. It's a manually curated collection of small blogs and personal website that anyone can contribute to. It's open source and it drives a portion of our index and we actually surface content from these small blogs that otherwise never surfaces in mainstream search engines. So the quality of search is one thing. The other one is just innovation in the search interface. We stagnated for 20 years.

Even non Google search engine have not produced really any innovation in the space. All the modest minds that these companies are tasked with optimizing for ad click through not really the search experience. So we knew from early days that we would also need to innovate in the search experience. So for example in Kagi one of the favorite feature of our users is you can block a website you don't like and it will never ever again show up in your results. It's going to be dead.

And we actually publish a. There's a Public leaderboard of most blocked and most, you know, like, hated websites. And there's like also you can promote websites so you can find all of that kagi.com stats and so that, you know, this sort of search feed personalization where we are not trying to impose what we think are the best results for you and our own biases and whatnot, but we let you moderate and be in full control of your search feed.

If you don't like, you know, particular website, you know, political affiliation, whatever, right, you are in control and you can blog these websites and promote the ones that you like. I'll give you another example that's very popular. We have a feature called lenses where basically you can create a search lens with a subset of websites and search only through them. So people create a lens for their hobby sites, for recipes.

We have a couple of built in like for programming or academic research. So basically you search only within the space of these websites. So these are very, you know, two very simple, I would say innovations. Now looking back, but it took 20 years for somebody to bring them to search because nobody was really interested in what the user wants. And so Kage really has, you know, at least 50 features like that and most of our users don't even know or use all of them.

But we are really investing a lot on, on bringing the search experience to the next level.

So just being able to pin those pin my own blog to be able to find the things when I'm quickly trying to share a link with somebody instead of trying to go to my own blog and find it. I can actually find it faster in Coggy by searching for something that's within it. And then I know that my site's going to be pinned and then the block feature as well.

I did just look up the stats because I was curious and I'm not at all shocked by most of the ones that are up here, like Pinterest, which is polluted searches for a decade now. Quora things that generally are the most useless results. It is a fantastic way to get, get those things just purged from your list.

And I think the most powerful part of it is it's not some supreme overlord deciding this site or these sites shouldn't be seen in ads, or these should be deranked, or these should be ranked higher. But you, the user, get to build your own filter. You get a really good base set of search results and then you get to say, I really appreciate things from Reddit more than other things, or I really don't want to see Reddit replies and I want to focus on things like small blogs like that.

That type of ability to craft your own algorithm is something that has really stood out to me when using Kagi to make it something that's much more personalized. And that takes like, I will say it's quite a barrier of entry for me at least because it's this whole novel idea. Like I'm used to. Just the search engine will do everything for me.

And while Kagi does out of the box, it's interesting because there's so many features and so many toggles you can play with and so many new things that you can tweak to see what gives you the even better than the default Kagi results for your very specific set of things. So it's definitely kind of like a kid in a candy store thing where you don't know what to turn on next or tweak next or add to next.

But that's a great problem to have rather than just saying please Google, please give me the result that I need here.

You can completely change the interface and people have published like numerous themes and whatnot, so you can really make coggy your own. And that's the whole point of putting user in the driving seat and letting them decide. I mean, even the personalization feature that other search engines could easily adopt, they cannot because of this conflict of interest. And probably the advertiser wouldn't like people blocking them or whatever. Right.

So, so that, that simple thing, aligning incentives goes very, very far.

And I know it's been talked to death, but just this idea of if you're not paying with your money, you're paying with your data. And everyone's just become so used to that that it is like when I'm talking to friends. I was talking to a friend last night about Kagi and was just sharing the pricing and he was like, I've never even thought about spending $10 a month on search. But he's really interested because of all the things that we've been talking about.

And I think when you realize how much of your life you spend searching, it's a, it's a very worthwhile investment. So it is just a fascinating area how just shifting the financial incentives makes it so much more beneficial for the end user rather than, like you said, the customer of Google, who are the ad companies, not the actual end user. Yeah.

And it's one of the most entrenched habits in our society that free. So the search is somehow free and God given and you know, granted to us where, you know, that's not true. And it's one of the most profitable and biggest companies in the world and the world's biggest advertiser behind it. And so it takes time, it takes time for people to understand that and really start valuing access to information and specifically access to information that's provided in their best interest.

That's a very, very important thing for everyone and the society in general.

So that is a much broader topic as well that I think we were talking a little bit before we started recording here about just kind of how the human freedom side of privacy alongside the view of privacy. Just so that I don't have to be part of this crazy corporate ad monopoly world that we've fallen into. Those things so easily intertwine.

Because I think that's an interesting example that most people don't think of as well is if everyone's using Google for search results, guess who has a complete monopoly on the flow of information on the Internet? Because unless you know the specific website you want to get information from, you're only going to be able to get it through their search results and they have complete black box control over that. Which is kind of a human freedom angle of the search provider arena.

So this whole ad based economy has not created the largest surveillance system in history and the largest behavior changing system in history, but also have ruined what the basic idea of web was. To be this place where we fair and self express and be creative and things like that. And it's 99% driven by this ad based economy today which creates very, very bad incentives for the overall healthiness of the open band.

And I was just catching up, getting, getting my computer set up back to COGI and looking and you'll have a lot of things outside of search. So I do, I do want to touch on those as well. But something that's kind of always bugged me from the privacy perspective is there is one downside that I see to the approach that Coggy takes over a free search engine, which is that you necessarily have to have an account and you have to pay for that account.

Y'all do have some like some non ID to payment methods like Bitcoin and Lightning that you support, but it still means that all of your searches are potentially tied back to a single account. I know that y'all have a very clear no logging policy and you have mentioned as well that just the financial incentives are misaligned for you to collect user data.

But I'd love to hear you speak more on that kind of that concern because I know that that's one that I have had kind of since I started using Kagi and was a little bit of the reason why I backed away from Kagi initially, even though I'm coming back and starting to use it again now. But that's also one that I've heard repeated by several people in the kind of sphere of influence that I have.

So I'd love to hear more of your thoughts around the privacy expectations and the pros and cons of having an account for your search.

So this is the reason we have an account and sometimes people, you know, forgot a password, they contact, support, things like that. So in a paid service you kind of need to have that. So over the years we have introduced several options. One of them is, first of all, you don't even need an email address. Like you can use a fake email address. Again, we don't care about email address. That's the thing you use to log into your account. It can be anything.

If you use a fake email address, then obviously you cannot restore that. But other than that, it's up to you. The other thing we added, I think was last year was support for Bitcoin and Bitcoin Lightning to further sort of distance yourself to any payment methods that could be trackable and things like that. I'm not an expert in that community. I know that Bitcoin mining is pretty good.

I know there are some other Alternatives like Monero and things like that, we haven't explored them yet because frankly we don't want to be in the billing business, we want to be in the search business. So we thought that these two combining reusable email address ping with Bitcoin Lightning provides you, at least from a user standpoint, regardless of what do we say in our privacy policy, a pretty good sense of security.

The thing that we are working on right now that I'm happy to sort of announce is we are working on a blind authentication token method. So basically privacy paths for search and we're hoping to launch this in the next few weeks. Basically it will be a browser extension and when you use that there will be a technical guarantee then that your search cannot be associated with your account even if you want it.

So that is, I think will be the last piece of the puzzle that we have worked on the last six months from proof of concept to now fully integrated browser extension. And after that the users will have 100% secure method of knowing that what they do on Kagi is going to stay private, even if Kagi somehow wanted to know what they're searching for.

So we went like the full circle there and listen to our user community and with the launch of this new technology, I think we are accomplishing, you know, the older requirements by this community.

You can have your account in a way that doesn't link your identity or link at least your real name, et cetera, from your credit card directly to your account. So that at least was kind of like the minimum bar that I feel like y'all hit. Well, and that at least allowed you to have a pseudonymous account and not have everything tied directly to your id.

But Blinded Auth will be a really, really powerful move forward and I think will be a, a huge part of especially the communities I'm in being willing to jump in and use Kagi regularly, especially with already having Bitcoin and Lightning. Because as we talked about before, Bitcoin and Monero specifically are kind of a key part of the things that I focus on for enabling human freedom.

And I think there's a lot of aligned people in the Bitcoin and Monero space that want more privacy preserving search and want to be able to pay for it directly. So that's really exciting. And I know Blinded Off, I'm sure that was a lot of work on the back end because that's not a straightforward thing, but that will be a.

And I want to show you on Monero on the podcast, but I think it would be a good, a good fit and there's a lot of Monero users who would love to be able to spend Monero on Cocky.

That you've launched recently because I know you'll have some AI products, you have translate. There's a lot of different things going on. So what's going to get you most excited that you'll have out right now outside of search?

So basically the idea is we would introduce an ecosystem of tools that will be cross platform, cross device, where you can really, you know, trust that the product has your best interest in mind at all times no matter what product is. And there are a couple of products that, you know, people use to, to consume the web. Search is one of them. Browser is another email. In my mind these three are sort of pillars of consuming the web.

And interestingly, at least for browsers, browser is even more intimate. In search, we are in our browser like 10 hours a day. It's probably the most intimate piece of software we have on our computers yet. Most of us let companies that are, you know, advertising funded one way or the other and you know, they, they let the advertisers pay for our browsing. And that is also something completely unacceptable for me.

So this is why we started working on a browser that will align incentives, be zero telemetry, really privacy friendly, like not none of the mainstream browsers are zero telemetry for example, while they claim their privacy respecting. So that always sort of was astonishing to me. And yeah, so we want to create this ecosystem of tools where people can trust.

People will have a vote with their wallet, people will be able to contact support, people will be able to request features and we will be listening to that user community and building better products together because all the incentives are aligned. And I think the consumption of the web lacked something like that. We used to pay for browsers up to 2005. I think maybe eight was the last paid browser out there. You sort of bought your browser and had it like it was yours.

That sort of vanished 15 years ago and I don't see any reason for that not to be a thing. Like, you know, you buy many things in life, you buy that coffee at Starbucks, you don't expect a stranger to give it free to you every single day. Yet for something much more intimate like search and browsing, you're completely fine with that expectation that somebody else is paying for a coffee.

Right. So I think there will be a major shift in how we view consumption of web in the future and we will want to trust companies that have our best interest in mind. Simple as that. So COG is building this ecosystem. These are the three main products that we build. We are building a lot of sort of secondary products like the Translate that we launched the other day. You know, I believe you should have access to high quality translation without jeopardizing any privacy.

And not only that you get with Cognitranslate, but it's beating Google Translate and DeepL on testing that our users do. So we have the universal Summarizer which summarizes any content on the web. It's a browser extension. You can also use it in Kagi search results to summarize any search results. So we are slowly building this ecosystem of tools. Again, we are not lacking ambition. It's just the team is small and we have to pick our battles.

And most recently we have started working on an integrated AI assistant, which basically is an AI experience, chat like experience that has access to all the top LLMs in the world, but it's grounded in Kagi search. So you get not all, not just older models, but when you're searching for something you get actually accurate results that are grounded in COGI search results. So there's a lot of things going on.

And yeah, sometimes people tell us like we have these release notes every week and they tell us like it's very hard to keep up with the amount of stuff we are producing. But we are, we are incentivized to do that. And they're all good.

I'm not on Mac OS anymore, so unfortunately I can't use it today. That is a quick question that just comes to mind while I'm thinking about it. Do y'all have plans, I'm assuming to expand Orion to other platforms than macOS and iOS?

We're basically building it from scratch. It's not using chromium, it's using WebKit. So there is no, you know, browser browser app framework like Chromium. You have to create from scratch every item, every menu, every interaction. So we've been doing that for the last five years. We are nearing the launch of V1 and once that happens, we are planning to start building Orion for another platform. So I can't say when, but there's the will.

Like rather than me being a free user of Google and having no attachment to Google and not feeling at all like the things that I'm doing with Google are creating things that will benefit me in the future. There's this really cool aligned incentives that happen when I'm paying for Kagi Search and I see that Kagi has now been able to build all of these new tools that you're building a browser.

It's also a very different feeling for me as the user because you get that sense of like I helped a little bit to get it there and every bit that you do, helping spread the word on these kind of programs, like obviously I'm not working at Kaki, y'all are the ones doing the work.

But when you're funding something with your money rather than funding some far off corporation with your ads, with your ad clicks or your data, there's a much more interesting kind of set of, I don't know, it just feels very different. It's a very encouraging, very more human centric approach to the web where it feels a little bit more like you're a part of something and that you're helping something else. To become better rather than just a part of the system.

So I don't know if that's a little bit too philosophical.

And you will learn about Kagi through podcasts like yours and people spreading the world on social media, on Reddit and hacker news. You know, the major tech media is not interested. So we are very, very grateful for having such users that help us build this together. And really, you know, there would be no cognitive, literally with all these users in every sense of the word. So thank you for doing that and thanks to all our user community. You mean really a lot to us.

Just any closing thoughts or last things you want to say before we wrap up here.

I would love to see my children being able to use Kagi, their school versus Google and not be forced to change their search engine because they have how you said, and then the teacher changes it back to Google and they change it back. Like, that's so wrong on so many levels.

And I always thought like we would get to the point where, you know, this will be a reality. And it seems like we got a bit closer with LLMs, but it also feels like it's not bad. And the technology seems to be limited. And I'm not 100% sure it will get us to that vision because nobody check the output of a Star Trek board computer. It was always right. With LLMs, it's like hit and miss.

Right. And so I'm not sure that technology will get us there, but I'm really excited about possibility of having that sort of technology more readily available. And you can sort of talk to a computer and it will sort of be efficient and accurate. There's a very interesting vision by Apple from 1987 called the Knowledge Navigator. And you can find this on YouTube and you can see what Apple envisioned in 87. So that's something I would like to try to build one day.

But I think we are very far from that. And I say I don't think LLMs will get us there. Unfortunately, they will need to be another breakthrough. But important part of that breakthrough will be that incentives will be aligned. Again, that Star Trek onboard computer, you cannot imagine it showing it an ad before or after it said something to you. Right. It needs to be in your service.

So again, this idea of information being in your best interest I think will be a big idea or a bigger idea in the next five to 10 years.

And when the data that you feed it is overwhelmingly data that's driven by ads or that's driven by a kind of clickbait economy, there's a lot that would get better if the actual information that people incentivize to put out into the web improves. So I think there's an interesting thought that perhaps as we slowly fix the web and slowly build in better incentives and slowly break away from the ad economy.

Things like LLMs may naturally get better as the quality of the info that's out there on the Internet actually improves as well. So certainly a tangent, but an interesting one where I'm curious how that that interplay works out in the future. Well, thank you so much for coming on, Vlad. I know we kept this one relatively brief, but I wanted to make sure to let people know what you're doing at Kagi.

Like I said, I've been a user for a couple years now and really, really enjoying it and excited to jump back in, especially with the blinded auth that you're working on. I think that's going to be a huge breakthrough and fix the only real downside that I saw with COGI at this point. So really looking forward to it. Thank you so much for your hard work. And then last question, just where can people find more about you? Find more about Kagi? Where do you want to send people?