Hacker News Reader

> What makes you think you can do better than a company with over 20 years experience, 10s of 1000s of engineers working on their product?

I've tested it since December.

What made them think they could do it I don't know but they already outclass Google by a wide margin.

The shortest and most precise explanation I can give is:

Kagi is to mainstream search engines today what Google was to mainstream search engines 20 years ago.

And that maybe settles your question too?

> No-one has even come _close_ to taking a dent out of Google in the last 20 years.

Unluckily for power users (and maybe ordinary users) and luckily for would-be Google competitors Google has rested on their laurels, told everyone the problem was webspam while their quality declined down the the levels of earlier search engines. Edit: Kagi shows the problem is not webspam but Googles insistence on 1.) not letting people block domains 2.) not respecting peoples queries.

Competing with Google a decade ago would be hard.

DDG still isn't anywhere near peak Google quality and yet they've seen (actual) exponential growth for years. That wouldn't have happened if Google still was leagues ahead.

Until December I started in DDG and used !g to see if Google had a different result that was better. Often it did not.

Since December I have used Google twice to verify something from Kagi or to look up something Kagi couldn't find (I'll try to get a bug report underway freediver, I just need to recreate using non personal data.)

>DDG still isn't anywhere near peak Google quality and yet they've seen (actual) exponential growth for years.

Google gets the total of DDG's lifetime searches (over a decade's worth) in about half a week. "Exponential growth" is overselling it.

Maybe they’re using “exponential” metaphorically to mean “a lot”. But if they’re using “exponential” literally, then maybe the exponent is 1.03.

Look at the graphs.

To me it looks literally exponential until half a year ago.

or look here: https://news.ycombinator.com/item?id=29852783

Seems I'm not alone in thinking it has been almost exponential for a good while and until recently.

Other people may have different opinions about their personal data.

The advantage Kagi has over Google is the different business model. Google would never for example introduce a "non-commercial" filter as it would go against their advertising-based business model. Kagi doesn't give a shit about ads and will do whatever that keeps its users paying and so has a non-commercial "lens" (as they call it). I'm not sure how good or accurate it is, but it's at least there which is already more than I can say about Google.

Kagi has a feature to prioritize or exclude some domains. Google doesn't have that despite their size, resources and decades of search experience. That makes it pretty clear to me that Google's weakness has nothing to do with experience or resources and thus a new entrant could potentially compete without either.

If Google went completely add free, they would need to charge $57 per year to entire US population. I bet vast majority of population will rather be willing to endure ads. HN bubble often forgets how small it is relative to whole population.

The price point of Kagi is anywhere from $120/yr to $240/yr, based on their current FAQ.

Let's say they manage to take just 1% of Google's daily search traffic, which based on a quick Google search (hah) would be about 35 MILLION search queries. If we do a very very rough ballpark estimate of 50 queries per user, per day, it would be about 700,000 unique users... or between 84,000,000 and 168,000,000 USD/year gross revenue.

Now, do I think it's reasonable to expect them to take 1% of Google's search traffic anytime soon? Probably not. I do however think that the sheer size of this market means there is a LOT of missed opportunity by the incumbents in the market that can be leveraged to create niche products that perform successfully by any other market's standards.

So, even if the HN bubble is small there are other bubbles out there. A LOT of other bubbles out there.

>which based on a quick Google search (hah) would be about 35 MILLION search queries

Public numbers are a decade old. You're off by an order of magnitude.

That's fair enough, but the option should be available for those who are willing to pay. $57 per year sounds like a steal for anyone doing any kind of productivity work with computers.

I think it is worse. The people who are willing to pay $57 to opt out of ads are probably the most profitable (I am guessing here: no citation available). Seems like a case of Adverse selection https://en.wikipedia.org/wiki/Adverse_selection

That doesn't seem obvious to me. I'd be more inclined to think anyone willing to pay to avoid ads likely isn't very receptive to advertising in the first place. If someone is so annoyed by ads, they're probably not going to be clicking on them, right?

In terms of purchasing power they might look desirable, but surely they’re also likely to be massively overrepresented as users of ad blocking solutions (also pure speculation on my part). That could make them less profitable in practice.

In my case, Google doesn't work too well.

I usually have something in mind when I search, and Google doesn't offer any customization for me to put it in. Instead, it shows me what it prefers.

For example, I frequently need to switch languages and/or regions. You can't do that with Google (except maybe logging in and switching some settings but lmao, I need a switch). No matter what I do, Google reads my IP range and "decides" what results I should view. Very frequently, this means I literally can not find the thing I am looking for.

For many people - I assume anyone who isn't American - Google simply isn't working very well. We use it, yes. But DuckDuckGo has a Region switch that already makes it 10x more useful than Google and usually offers far superior results. A new search engine could expand on that, offering more customizations.

I'd go so far that for every human, except 350 million Americans who never switch language or region, Google is ALREADY the inferior search engine and people use it because it is what they know.

The infuriating thing is that Google used to do this right. It used to be possible to go to Google.de and see the German results for instance. Now I always get the Swedish results (I live in Sweden). I often have searches where I want results from Sweden, Germany Australia/NZ or general English. I have to become increasingly elaborate with my search terms to get them (essentially search with whole sentences in the desired language), it's really annoying and I'm surprised how Google could get the so wrong.

I appreciate the sentiment and I can certainly understand it. This is of course what I thought too in the beginning, but as curiosity and then determination led one thing to another, we learned that things are doable, as much as that sounded as science fiction. Certainly there is evidence in user feedback that, at least for some people, Kagi is decent at what it does.

The fundamental premise that enables it is the focus on the user. It allows truly wonderful things to happen in the product (and we sometimes forget that search is still a product). Aligning the incentives between us and the user (which is also the customer) is really freeing from a product management perspective.

We are still users of Google products - we utilize the paid search API for results and GCP for infrastructure (hello Google!). We think there are some things that can be built with these truly amazing assets, that Google's business model simply does not permit, and we are thankful for the opportunity.

In the meantime we also built our own index which focuses only on noncommercial content and are going to invest more in this direction in the future. We call it "the more humane web".

Another thing is that we do not need to get _very_ far in the market. There is no shareholder pressure, just angry (or happy!) users. This is not, nor will ever be, a Google scale product.

This is a paid search product of premium quality (or so I'd like to believe it is shaping to be) for a tight number of individuals (and families!) who are a little different and want something different for their search experience. Few tens of thousands of subscribers will make Kagi sustainable and I hope we will have what it takes to get there. That is all I can wish for right now.

Thanks for considering!

Towards that end, I wonder if you could partner with or incorporate some of the things Marginalia is doing. https://search.marginalia.nu/

I do plan on offer some sort of API access, but I just haven't had the time to work out the details yet. In general I do believe in collaboration within the alt-search space, but I'm a bit reluctant to have my non-profit work end up being the back-end for someone else's for-profit operation. Still thinking about it.

In part I've been postponing this because I'm planning a bigger overhaul of rate limiting and access stuff, in part to get rid off cloudflare because I don't want a creepy MITM spying on my users if it isn't absolutely necessary.

Thanks for the reply. Let's see what happens and I wish you all the best!

> If it uses my data to help target advertising - so be it

What a great thing that data leaks and power abuse don't ever happen. /s

> What makes you think you can do better than a company with over 20 years experience

The severe decline in quality? I've been using DDG for a while now; in about 2 years I've looked at Google results maybe 5 times and was disappointed every time. If you think Google is the best you can get right now you're missing out. Your argument is based on the assumption that a company with more experience cannot possibly be bested at what they're doing.

> google could implement site muting in an instant if they wanted

Google could do a lot of great things, but unused potential is worthless. Remember how Google got big precisely because they did what competitors could do but didn't?

Google has been getting worse for the past few years, probably in the demands for making a standard product to suit the average person. But now the first page is full of results that exclude one of the search terms. And my strategy for years to find obscure results- of finding the oddest set of relevant terms to describe what I want to find to weed out commonplace results that share a couple terms — has stopped working almost entirely as Google's algo seems to have upweighted popularity considerably.

At this point I don't think it would be that hard to provide a useful alternative, especially with a business model where the customer is the searcher and not the advertiser. I'm not sure that all of those engineers are having an additive affect on the UI. Quite the contrary, I imagine many are working toward improving the experience for advertisers, who are the customers.

The quality of cadbury chocolate has been getting worse over the years since Kraft bought them out. I still buy it because it looks like the same product, kinda tastes like it, and ultimately will give me what I want. I don't disagree the quality of Google's results has gone down - but you are talking about changing user behaviour here and for the majority (not you and I) google still does what it needs.

It may even be improving for most people's needs. But I think it does present a scenario where a niche search engine with a different business model could be successful (and they probably wouldn't need 10k engineers to reach parity with Google's, at least in that niche). Maybe there isn't enough of mes though.

I haven't changed search engines, and Bing won hands down when I did the Bing/Google challenge they promoted in 2012. But it is getting 'bad enough' and I need search results for projects enough where I'd try the beta if invited and would probably stick with it if it's a good experience.

Hell, if they gave me a few more Scopus-like tools for more precise queries on a version of Google scholar, I'd be sold.

The problem I have is that Google is not optimizing for the best results - they are optimizing for homogeneous "best" results for the average person. And a lot of that comes down to optimizing for showing product placements (even if they are not ads) and difficulty filtering out garbage paperclip sites and repetitive blogs and whatnot. It used to be that I could fine tune my google searches using just quotes or other simple "advanced" methods and get great results but that has significantly decreased as the years have gone by. If I search for a string in quotes I dont want google automatically "suggesting" I meant something else and then I have to click another button that I am "sure" I meant my original terms.

This argument always comes up. What makes you think you can... Bla bla bla.

When Google search came, they started their search engine by providing a clean search page with no ads, at a time when all the others had cluttered, ad-filled, shitty pages.

Users picked Google because it was the better experience back then. Today, if a search engine comes along that gives you great search results and privacy, it will become very popular quickly, because a lot of us are fed up with Google and want them to go away completely.

Google came along when the search industry was in its infancy, with multiple providers vying from top-spot. We have now had over 20 years of dominance, refinements to get it to what it is today. Entering the market today is on a whole different spectrum to what it was back then.

If anyone is able to take a large bite out of Google's search market share, it will be by figuring out how to not drown results in ads, SEO and blogspam. If they can successfully pull that off, it has obvious value.

Easy to detect ads and affiliate links, just down rank all commercial sites. The good sites are all free anyway.

Quality of google results, has been going down for quite a while.

There have been some speculation that it's down to SEO and google trying to make as much money as they can.

It's possible that a company which does not derive their income from search ADS could create a better search engine.

If someone can make a better search, I am all for it, even if its just for limited topics.

Subjectively google's results have got dramatically worse over the past 10 years. Whether this is intentional on google's part because they have decided to prioritise ad revenue over useful results, or if it is because the problem has got much harder because of the increased number of pages and the increased sophistication of spammers is open to question.

If the answer is a) then Kagi or another disruptor could undoubtedly produce better search results, and probably in a short period of time, with a small number of decent engineers as there is now a vast body of research and praxis to build on that didn't exist when google was starting up. If the answer is some variant on b) then they're on a hiding to nothing.

You don’t care what Google does (aside from search) because you consciously installed a tool which hides their work from you. Glorious self-deception!

By hiding ads, you’re forcing everyone else to pay for your search. I’m glad I use DDG, and I?3 signed up for Kagi’s beta.