Comment by infecto
7 months ago
I still pay for Kagi for its search but this has kind of been the problem from the beginning with their org.
- Search has been a breath of fresh air, I wish they dedicated more time to it.
- Orion...is ok? I use it off and on and it is fine but would rather have better search. The premise of the browser is nice but it feels like this could probably be a whole separate company or a purely open source endeavor. It has always been kind of clunky and not something I want to pay for.
- AI tools, I get the multiple pivots and I do believe that more recent advancements in ML/AI will make search a better experience but I do wish they had a little more focus.
- The t-shirts are weird and made me lose a lot of faith in their org. The way I look at it is instead of making their products better, they flushed a bunch of money down the drain for t-shirts and its offensive to paying customers who are paying because they enjoy the product and want it to become better, they don't want a t-shirt.
- I don't care about email, I don't care about other tools, make a great search experience first. Release all of the AI enhancements that you think will make sense, focus, focus, focus.
Edit: As I was adding my comment this post flagged and marked dead. Sometimes HN is weird.
> The t-shirts are weird and made me lose a lot of faith in their org. The way I look at it is instead of making their products better, they flushed a bunch of money down the drain for t-shirts and its offensive to paying customers who are paying because they enjoy the product and want it to become better, they don't want a t-shirt.
Kagi founder here and I want to clarify the train of thoughts around Kagi printing and giving away 20,000 t-shirts for its users.
- Kagi is not a typical VC funded startup.
- It is company I bootstrapped by going all in (meaning I put millions of dollars of my money into it).
- After all these years building it, we are lucky to have such incredibly passionate user community.
- That community is 100% responsible for Kagi's growth as a business through word of mouth (Kagi does no paid advertising).
- We are also famously taking a firm stance against ad-tech, so conventional advertising is not something I want to do.
- To do something as crazy as to start a company that builds a paid search engine and browser you obviously need to be thinking out of the box.
So combine all of this together and I thought that sending a t-shirt to all the people who supported us along the journey made a lot of sense.
The only thing I did not count on is how difficult will be to pull this off as I did not want to settle with less than premium quality for these t-shirts. As a result they will be delayed (my best guess is July/August) and I apologize for that to our users. In hindsight, we probably should have opted out for something easier to pull off (someone mentioned a billboard on 101, that would certainly be much easier).
This did not jeopardize Kagi's finances in any way at any point, nor I would do anything like that ever (as I said I am all in and have everything to lose, so I run a fiscally responsible business). In fact, Kagi has turned profitable recently.
This has also not impacted our ability to hire (we went from 10 people twelve months ago to 25+ now) and it did not impact our ability to ship a great product (check Kagi and Orion changelogs). I would venture to say that most Kagi users agree that Kagi is getting better and better every week with great speed.
So would I do it again? Well let's wait and see what we have in store for hitting 50,000 members mark :)
I work in CX, you should listen to your customers. Your gut got you this far, but to be a profitable company you are going to need to consider the advice and concerns of your stakeholders. Based on your current description, you have two stakeholders (yourself + customers).
If the venture fails, you will ask yourself if you listened enough. Be proactive, address concerns, do not put yourself in a defensive position. Embrace change, be agile, and most importantly listen to your feedback.
Wish Kagi nothing but success, I would very much like a disruptor in this space. Best of luck to you and your team.
This is a forum where people respond well to practical explanations from thoughtful founders.
I don't know if the OP got what they needed from this reply, but I assume I'm not alone in being impressed by the humility and candor of the response and developed much greater affinity for Kagi from some of the specifics of what were said.
I want more companies to have communicative, principled management that invites a sustainable base of like-minded customers/partners and fewer companies that pretend they can please 7B people by radically changing their product every 3 months.
21 replies →
>I work in CX, you should listen to your customers.
The only way a customer speaks is with money. If people like what you sell, you'll have more customers speaking with their wallet. If they don't then they tell you so by not purchasing what you sell. Internet commenters (such as myself) do not represent all customers or even a majority. People who are happy with a product usually see no reason to give feedback – especially when it's a small purchase. Likewise, people who hate your product wouldn't purchase it in the first place.
9 replies →
The problem is, there are many customers. You should listen, but that doesn’t mean you have to agree with everything or submit to every demand. I for one find Orion useful and it would be a bummer if it was scrapped because of a single comment on HN. Also, “I lost faith in a company because it made T-shirts” sounds a bit hyperbolic to take seriously IMO.
> I work in CX, you should listen to your customers.
...and to be clear, Hacker News is not a representative sample of their customers.
8 replies →
That reminds me of the faster horses quote I'm afraid.
Or you know, that all cell phones had to have a physical keyboard. Until they suddenly didn't.
[Never tried Kagi, but let the man do his thing.]
7 replies →
This is a very weird answer. People who are paying for your service want you to succeed, they want an alternative to Google search. Many of them (like the article above) explicitly say they don't want a t-shirt, they just want a better search product from you.
After all this very valid, very sensible feedback, you're commenting here saying "you need to be thinking out of the box" and trying to justify all the time, money and energy you spent on those t-shirts. Your customers are complaining because they want you to succeed. If they didn't care about you, they'd just cancel the subscription and move on. And your response to it is "nah, what we did was right" and not "yeah, maybe we shouldn't have spent all that resources on a stupid t-shirt that nobody wanted"?
I just don't get it. And what are you gonna do when you hit 50K members? Are you planning to send an entire wardrobe (from undies to a suit) to all your 50K customers (assuming IF you ever hit that many customers)?
I am a little lost here. So correct me.
The Founder of the company decided to pour in his own money worth of millions ( assuming that is true ) into the company. He bootstrapped it.
They were on track to becoming profitable. And now, as the replies shows they are now profitable.
He decided to spend some money to buy everyone a t-shirt as a gift.
And all I see is not thank you but anger and frustrations. Pointing to waste of time as if the product is going downhill or not iterating fast enough. But in fact according to changelog they are doing pretty well.
Spending my own money to say thank you in my own way and all I got was, all these negativity.
I just don't get it.
3 replies →
I don't think it's that weird at all. Nor did they say the "customer" needs to think outside the box (what a great misrepresentation). They were just explaining their thinking process.
1 reply →
Yeah, I love that "free t-shirts" is classified outside the box thinking.
Oh, and nitpick: It's outside the box, not out of the box...
"Kagi does no paid advertising"
I remember Daniel King's PowerPlayChess channel recently started promoting Kagi, doesn't this count as paid advertising or is this deal something else?
You are right, we started doing that last week so technically we are doing marketing now.
8 replies →
Hi, Vlad. We've met before, I'm the person whose 70-year-old mother is using Kagi. I also have actively been trying to move people to Kagi for some time - even paying for their accounts. The biggest block I face is not brand recognition. It's a lack of An Android app.
To move my mother to Kagi, I had to install Brave Browser on her android cell phone, make it the default browser on Android, change Brave's default search engine and create a desktop shortcut to it.
Android is ubiquitous in Brazil. I won't be able to move much people to Kagi with a process this involved.
Happy to chat further about this topic if you wish, privately or publicly.
Appreciate the response. I hope while some of it, including mine might come off as critical or uninformed, it truly comes from a place of love for the search product.
I still don't agree with the shirts and I think the overarching point is the shirts seem like a common theme of trying to do too much. I hope my thinking is not true and I wish the best success because I love Kagi.
I'm not a user, but you must have found a great market because your users are anxious about the company failing. The fact that they spinned the fact that you were able to create a company and a whole t -shirt operation on a marketing budget as a bad thing is telling. You're doing great. The t-shirt op is a great investment an will return a great value.
Keep going! Kagi is great. Years ago nobody thought Google could be challenged and that nobody would pay for search, yet here you are.
do they have a decent subscriber base? i thought it was still very niche
9 replies →
Thanks for your response!
Honestly I get the T-shirt part this way. You got to Doo crazy stuff as a start-up. I also get that you try ai stuff. As long as you keep up de search.
However what scares me is the apparent lack of knowledge about privacy, gdpr and what is PII in a product that, to me, is all about privacy. Have one person in the company be an expert in privacy and GDPR etc and use their insights, since it is critical for your right of existence.
I got the same impression - the lack of understanding of the basics of GDPR makes them look as amateurs, not professionals trying to raise the bar for privacy. I was considering using Kagi, but this is a massive turn off.
They'll likely discover that GDPR is not that optional as soon as a customer (or a competitor with a grudge!) reports them to their relevant national privacy/personal data protection authority, after which they'll get to have a very uncomfortable conversation where they will not be able to use those arguments
7 replies →
Yesterday I realized I had 75 tabs open on my mobile web browser and decided to do some trimming. Anything I was confident would come up again and didn’t need to be held onto got closed, including a tab for Kagi. And now I find Kagi has come up again, and I really liked reading this message so now I’m opening the tab again and almost certainly subscribing
Very interesting!
Quick (but difficult) question: do you foresee there arising a reasonably reliable way to filter out the coming wave of ai spam? I’m told that half of Twitter is bots talking to each other at this point, and I’m sure this is coming to other media as well. Eg, massive, massive waves of content marketing, sock puppets, etc.
Is there reason to be optimistic that you or other actors will be able to sift through it?
Yes, significant part of our effort is to build technology that detects LLM spam. We have a working model that detects LLM generated text with 90% accuracy currently. The plan is to integrate in search results and make available as an API.
2 replies →
Sounds like a cool place to work. I'll check back for an opening in TypeScript / Rust backend.
I really appreciate that the $200,000 free t-shirts are delayed. Icing on the cake in my opinion.
Please please please take Hacker News' opinions with a very large grain of salt. Many of Hacker News' users work at garbage AdTech companies and there are often people posting here who say things like "I for one enjoy targeted ads" (that's an actual quote). This place is not representative of your customer base.
I love what you're doing and will continue to support you at your Professional tier as long as you continue doing what you're doing.
For my own business (epaper calendars), HN has been a great source of feedback from potential customers. People here are both direct and kind with their feedback.
The thing you have to keep in mind is that HN is a very specific niche of the Internet. But for a slightly nerdy, not mass-market, product like mine (or Kagi) this niche is a great place to grow.
You just have to be mindful to see the feedback through the lens of the fact that you're talking to a niche audience and keep an eye on what a broader market might be looking for if that's where you're planning to go
If HN is not representative of the customer base for a paid search engine, then what is?
3 replies →
I agree. My look about HN for privacy-focused topics changed after YouTube's blocker war. I realized there are many privacy-truder-tech workers here, and their comments were largely structured smartly to lighten how awful those "tech industries" are.
HN is probably more representative of the customer base than your preconceived notions and hostility. I imagine a large base of the current 20+k users are via HN.
1 reply →
> there are often people posting here who say things like "I for one enjoy targeted ads"
You can find some people on HN who will say anything. That's the nature of a large distributed community.
But I also think the fact that there are people who would say that is a good thing. I despise ads, ad-tracking/privacy invasion, etc with a passion. I won't use anything that has ads. I won't even install most apps, even paid/premium, if they collect personal information. So you know my biases.
Their argument is a whole lot more nuanced (and better) than you are presenting it. The main point is that they find (or consider) ads to be useful for helping people find new products they might be interested in. And that does quite obviously happen sometimes (how often is very much up for debate though(. For them, the privacy tradeoff is worth it. I would guess that a portion of people who agree with that would fall into the "what do I have to hide if I'm not doing anythign wrong" camp. I think they're nuts, but it is an interesting viewpoint to consider that is a lot more nuanced and contextual then just "I enjoy targeted ads."
I prefer internet ads to Billboards. Billboards are disgusting.
I prefer billboards, but I think that’s probably because they are exotic for me. We don’t really have them in Germany (or at least where I am), so when I see them in South Africa, they are always this cool and interesting thing.
1 reply →
I agree there should be less billboards in this world.
> Kagi does no paid advertising
Because of adverse selection?
Users that are users because of marketing are somehow different?
I meant digital advertising, like ads in search engines and websites - stuff that has gone out of control and we are actively fighting against. I would consider a billboard or sponsoring a podcast for example.
2 replies →
Instead of giving away t-shirts, can't you make the AI tools open source? They are clearly not up to the task yet so you might as well build a nice AI community first.
We are not developing LLM models.
AI tools that Kagi uses like vector database search and LLM connectors are already open source.
You can find these in our Github:
https://github.com/kagisearch/
I'm a deep technical nerd, but I approach Kagi from a basic user perspective.
Things I love and can't live without:
- When I search for something, I don't have to deal with weeks of whatever I searched for coming up in ads on every web page I visit.
- I don't feel like "the man" is snooping on me in some sort of weird dark social credit score thing. (I literally got a call from Google once offering me a job based on what I'd been searching for. Flattering, but totally freaked me out)
- The quality is good for non-local things
- I'm the customer, not the product
- That makes things like blocking or enhancing sites possible
What I'd like to see improve:
- I don't want AI. I don't want summaries, I don't want hallucinations, I don't want assistants. I don't want it.
- Local results and map integration. When I click on a local result, actually having the map go to the result I clicked on. Currently this doesn't work well.
- Hours for local businesses.
I find myself still going to google for these things, and while it doesn't seem like a lot, aside from work stuff those kinds of searches are probably 80% of what I need. Where can we go for dinner tonight that's near by and still open? Who has all-you-can-eat deals near by? Where can I find some floating shelves to put in my office near by?
Those are all examples of things that Google does really well, and I don't have much luck with on Kagi.
I agree with the author that I'd rather see the quality there improve before AI features.
Personally I love the AI tools. The summary tool is what got me converted from the trial to a paying user.
Quality is there for the most part, IME, but I definitely agree that their local features need a LOT of work.
> I wish they dedicated more time to it.
What changes you have in mind to search functionality? I feel like the core search is rock solid as is, but they address search quality reports on their feedback forums all the time.
To me, the AI features (and specifically how they are only used when you opt in per query) are enhancing search, and the time they have been allocating to those features has continuously improved Kagis utility to me.
Note: I subscribe to Kagi Ultimate, so I use some AI features that are not available in the base plans.
I love Kagi, I'm an early paying subscriber, but I think the quality of their results is overstated. Anytime you get past result #5 or so, the results just get _weird_. If you have to do deep research on something, you'll often get pages that seemingly have nothing to do with the query, or these class of pages that seem to be poor answers to common queries aggregated together.
I hope not to sound like I'm blaming you, but do you actually use the features that are unique to Kagi? Over time my manually configured block/lower/raise/pin list has continuously grown, quickly leading to higher and higher quality search results. I also have integrated custom lenses and bangs into my workflow more and more over time. I often end up searching seemingly very generic things and getting exactly what I'm looking for in the first or second result. Maybe my results after the first couple are weird too, but it doesn't really matter to me because I don't actually get very deep into results most of the time.
Google results were rather like this in the golden age of Google searches. But the bad results can give you an idea of how to tweak your search to get what you're actually looking for.
The modern, more polished Google search can make it literally impossible to find things that exist, because Google tries too hard to provide what it thinks you're looking for.
DDG is like that. If it can't find any more matches it seems to spam random results.
I tried Kagi and really enjoyed it but the pricing tier doesn't sit right with me, it's not that much better than DDG for my purposes. All these monthly subs start to add up. I'd be happier if there was a lifetime tier.
3 replies →
Ok, maybe it’s not that the quality of Kagi results is so high, it’s more that the quality of Google results is so poor. Just not having to deal with all the spam is a winner to me.
- Localized search is not a great experience
- Business listing search via maps is not a great experience. Maps and searching on maps are a more important endeavor than browser and email when thinking about the ecosystem.
- AI is definitely important but so far none of those features (afaik) have trickled down to non-ultimate users. From what I have seen, features have been removed from the regular plans.
- Remove reliance on using bing/google searches.
- Search is not a one and done operation.
> What changes you have in mind to search functionality?
Reverting the changes from around December that made it next to impossible to search for language-agnostic or English terms in another language.
Also reverting the changes over time that brought them closer to google or DDG and ignoring search terms unless you use verbatim or quote everything.
Kagi used to be about being explicit, but it’s slowly turning into the same "we know what you want to search for, so STFU" that all the other search engines are.
User since December 2021.
I paid for the 1 year plan, because I was excited about a company that was only for search, and provided good search by nature of people paying for it (so no ads and fair ranking)
I have noticed the search has gotten worse in the ~7 months I've been using it, I started using Google more and more, and I was not planning to renew. I still use it, but after reading this article I'm definitely not going to use it again after my 1 year ends.
They aren't prioritizing the only feature I care about.
> I feel like the core search is rock solid as is
Certainly not. I still get a decent amount of AI-generated blogspam in my results. Yes, it's great that Kagi offers me the option to manually block sites I don't want in results, but that's a workaround, not a solution, to the AI-generated spam problem.
I don't know if it's possible to detect this sort of crap automatically, but IMO this is the biggest threat to web search today.
Search quality requires maintenance and continual tuning. It's not a one and done "add more functionality" kind of product.
Fair. I was trying to make the point that they are already dedicating time to continuously tuning based on feedback on their forums.
The search quality tanked massively about two months ago, and is now almost as bad as Google.
I mean they could focus on actually building out their own search engine as an example? (i.e. moving further away from using Google and Bing APIs)
It's just a matter of focus with a team of that size.
Re the t-shirts: last time I checked the were private equity not VC and priced their product for profitability not growth.
Would you be upset if they had just donated that excess to charity?
So if I understand your comment, you are suggesting that they went and raised money to make t-shirts?
Not upset in the slightest, I love Kagi search and want to see it continue. Merch is a solved problem and there was no reason to bring it in-house and make such a big announcement around it.
I'm suggesting that self owned companies are allowed to and often do spend absurd amounts of their spare money on pointless things like marketing or internal transformations.
The difference they don't tell you about their internal accounting so you don't join the dots.
Start ups burn money on silly things like offices way too nice for what they need all the time. That's much closer to unethical than a company with no real duty to outsiders throwing away money.
4 replies →
Maybe they wanted to make T shirts. Who knows. I don't get why people are up in arms about it.
False dichotomy. They should be plowing that money into more R&D, or, absent the current ability to do that, saving it for a rainy day.
As a paying customer, I want Kagi to succeed. I want Kagi's search offering to keep improving. Spending a couple hundred thousand of the company's money on t-shirts (one that I would receive, as I was a fairly early customer) sounds foolish to me, regardless of how much the founder is personally invested in the company, and regardless of whether or not he'll invest more of his own money to keep the company growing in the future, if needed.
I'm still bullish on Kagi's future, but things like this (and things mentioned in the linked article as a whole) make me a little worried.
> Would you be upset if they had just donated that excess to charity?
Probably! When I was at Twilio, we participated in GitHub's charity dodgeball tournament a few times (early last decade, I think). The cost of admission was $3000 per team, and would go to charity. After a couple years doing it, finance started getting uncomfortable with it. We were a private, unprofitable company (now Twilio is, of course, a public, unprofitable company), giving away money that our VCs had invested in us.
Initially I rolled my eyes, "just the bean-counters doing what they do best: whining about every bit of spend". But later, looking back, I realized they had a point. While $3000 wasn't a lot of money in the grand scheme of the company, what benefit was spending it actually providing the company? Ok, so 12 or so employees got to go and have a fun day at a rec center, boosting morale for them. We got our logo on the website for the tournament, which was maybe a little visibility/marketing. But was that really worth $3000 of our VC money? Maybe it was, but I don't think it's an obvious "yes".
> Would you be upset if they had just donated that excess to charity?
Yes.
Would you be upset if they just paid themselves a higher salary?
Do you think they charge too much?
Do they charge more than you think it's worth?
1 reply →
> Edit: As I was adding my comment this post flagged and marked dead. Sometimes HN is weird.
I'm a full Kagi shill. But I also want the stuff I like to remain stuff I like and reasonable criticism is the path there.
orion is the only browser i use on ios as it supports uBlockOrigin and a bunch of other extensions.
i’m glad they spent the time and effort on it.
Orion is my daily driver and I hope they don't crush that. It has bugs, but it works.
As a "hard-core" Kagi user:
1) I legit can't fathom going back to Google or any other search engine. I don't know what I'll do if they go under.
2) Investing in integrating AI into their search is absolutely vital and I like a lot of what they're doing there
3) Everything else, including the insanity of the t-shirts thing, is a complete waste of time and money. I don't understand what their strategy is if it isn't to set money on fire.
Investing in better search is absolutely vital, and AI may be the right tool there, but I don't care about the AI. I pay Kagi to be a better search and informational retrieval tool, not to do AI.
It's not like they've gone all-in on AI though. Going through their changelog https://kagi.com/changelog it looks like they regularly make improvements to their core product and there've been a lot of significant QoL improvements in recent months. Just the Wolfram change alone has cut my need for Google significantly.
The one thing I really hope they put more work into is searching for local news. That's one of the areas where I still have to turn to Google.
I can't speak to the t-shirts. I was on duckduckgo before Kagi and also can't imagine returning there. I don't know what they're doing there but it's not improving. And yeah I am so with you on 2).
It seems like (again, t-shirts aside) Kagi is throwing a bunch of stuff at the wall to see what sticks. I hope they're having fun because I sure am.
I considered investing a small amount in them when they were raising a round from customers since I loved the search product so much. I too can't imagine going back to anything else, especially now that I have prioritized and blocked domains set up perfectly and added lenses, and this stuff works across desktop and mobile!
I've been mildly regretting not investing up until 5 minute ago when I read about spending 1/3 of that on the t-shirt factory.
The claim that's made in this blog - that Kagi 'owns a t-shirt factory' seems disingenuous, or lazy at best. Kagi's own blog says that instead of going with a major branded merch manufacturer/distributor, they chose to work with a small print shop instead. Nothing about blowing funds on an actual factory/print shop. "Owning a merch operation end to end" just means they're not paying some manufacturer to do production, warehousing, order fulfillment/drop shipping, etc.
4 replies →
I agree, but not necessarily that AI will make results better. Search engines already rely heavily on heuristics, and I really doubt that LLMs or vector databases are going to improve results in any combination. At best, they will overfit results to the lowest common denominator.
What I want is a search engine that supports full-text queries with exact matches. This quite literally no longer exists anywhere, and maybe that's because it just doesn't scale. Nevertheless, I would find a lot more value in a search engine that returns exact matches. Someone will probably reply saying that Kagi, DDG, or The Google do exact matches with quotes, but this is not true. When it works, you've just gotten lucky. At best, it will filter out inexact matches, but that doesn't mean it will actually return every exact match in the index.
Agreed, I want that + the page ranking that Kagi has, and literally don't care about any other features.
1 reply →
I agree pretty much verbatim. I don't see how anyone could criticize them for getting into the AI game as well or at least using a 3rd party AI software for some results. That would just be silly these days. I like Orion browser but to be honest firefox does what I need.
Totally agree on all points. I don't believe I have the technical capability for it but both the fear of losing great search and the lack of direction has made me think about what it would take to replicate the search experience.
> Everything else, including the insanity of the t-shirts thing, is a complete waste of time and money.
Presumably the tshirts are a marketing cost that they hope will lead to greater brand exposure and more subscribers.
They should've spent it on a marketing agency, because I don't know how a shirt which doesn't even have the name Kagi on it is supposed to give them brand exposure.
3 replies →
I agree with a lot of your sentiments, but we’re just in peak ai I think.
I'm actually looking forward for the t-shirt...
Kagi’s killer feature is somehow managing to get literally every post about them featured on the front page of HN.
If they fail with all of the free marketing they’ve been gifted by this community I can only shake my head.
A lot of us either use them or have used them in the recent past
I've been using Kagi only for a couple of months, so I'm still very much in the honeymoon phase. Perhaps they're still searching for their identity. Very much hope they rest independents and good at web searches.
I can't stand the randomness of how posts seem to be getting flagged more and more on HN. Seems like if a post is flagged and killed a reason should be given somewhere on the page by the flagger. Educate us on why our discussions should be off-limits, please. It would also be interesting to see if certain topics are always flagged by the same individuals and patterns emerge.
FWIW, this one got unflagged pretty quickly.
I didn't flag it, but I came close just because the tone of the piece is so sensational and needlessly aggressive. I left it put because it's the first negative Kagi piece I've seen and I didn't want to silence an alternative perspective, but the quality was definitely below what I hope to see on the front page.
It’s a pretty good thing to discuss though?
[dead]