28 comments

[ 7.3 ms ] story [ 80.5 ms ] thread
"The web is broken", claimed the plain text file formatted for a 1980s era screen and featuring no links or working navigation whatsoever. Someone still upset Gopher didn't survive?
we need a paradigm shift away from search, just as google engineered a shift away from portals in the late 90s. idk what form that would take but I suspect some kid in a garage is going to come up with something...
Rest assured when it leaves the garage Google will buy it up and kill it too.
Or maybe we just need to make better websites, so there more good content for the search engines to find.

We expect everything to be online. But I suspect that for many queries, we get garbage results because that's all there is.

My experience is that good content often is out there, it's just that SEO is so unreasonably effective that search engines aren't finding it.
This post is very shallow on the argument side and at its core just cyncial. I get the feeling the author has some personal beef going on with Kagi.
No beef! I thought the argument of "my experience was overall no better than with an ad-paid search engine" was enough... And how come there's no appreciation for acknowledging I did thank Kagi for existing to let us experience something? The piece I admit is 50% "what the hell Vlad" and other 50% is the actual "paid search engines are no better than ad-paid". I see the comments are divided which probably means I should go back and revise the article a bit
Edit: I've changed the title to "Overall disappointed with Kagi but lessons learned" :)
What a peculiar screed. I don't see anything in the post that validates the title. The author used Kagi for four months before the open beta and seemed to like it.

> A survey was put out a couple months before this announcement, asking users what they'd pay. I said 7$/month considering the current proficiency of the service. To read they went with 10$/month with no justification or reveal of the results just makes me feel completely ignored...

The author understands that it was a survey, but is upset that their personal input wasn't taken as the final decision?

It's fine to disagree with the tone of the beta announcement. It's reasonable to think that $10/month is too expensive. But "paid search engines are dead on arrival" is a big swing and there's nothing here to back it up. I'm trying out Kagi as part of the open beta, and so far I'm liking the experience.

I agree with you about that line in particular coming across as very entitled. This wasn't a scheduling survey for a post-work happy hour spot, it's a business trying to survive. I'm sure they took the author's $7 a month (really, $3 more is not _that_ much) suggestion into account.

The complaint about the call to child protection similarly was misplaced.

And just to put my cards on the table: I ponied up for a year of Kagi this morning when I got the announcement. I'm not sure if Kagi will be around in a year, but that's okay. Over the past few months, I've only had a single search query where Google outperformed Kagi (an odd one: "why did we all stop using ringtones"), and I like most of the features that Kagi has over Google. So I'm willing to put some money down to support a new entrant.

Author here: I guess it wasn't explicit enough! The thing to back it up is Kagi taught me paid-for search engines are no better than ad-paid search engines. I thought it was clear when I wrote this but I'm sorry it wasn't

I'm not upset it wasn't taken into consideration, I'm disappointed there's zero transparency! It's too bad the tone of the writing comes off as upset rather than disappointed, that wasn't my intention...!

Edit: I've clarified things and changed the title to "Overall disappointed with Kagi but lessons learned" to better show my intent. The original title was for sure a silly exaggeration and I have no problem admitting this :)

Good on you for taking feedback so well

However, it seems the title is still called "Paid Search Engines are Dead on Arrival"? Sorry I have no clue how to navigate this lmao

What is the point of this post? Just to bitch about someone’s attempt to make search better?

First of all $10/mo is not way too much if you’re someone who heavily uses search. Especially if you’re working in tech. If you’ve got a 6-figure salary and derive huge value from good search results, $120/yr is nothing.

And the premise that the web is dead seems rather pointless. Sure the web may not be the place it once was, it has evolved and changed over time. You may have preferred how it used to be, but that doesn’t mean it is dead. It will continue to change as people and societies do.

Superhuman is $30 a month and growing rapidly, for comparison.
I've been part of the Kagi public beta for months now. I've been almost eager for Kagi to start accepting payments because the product has been such a great enhancement to my daily life, and I want it to thrive. I agree, $120 a year is nothing.
I have to agree with the sentiments regarding Kagi.

I've been primarily using Kagi search since signing up for the private beta in April. Today, without warning, I was suddenly unable to use it.

It seems that with the launch of the public beta today, I was immediately switched to the free tier and limited to 50 searches for the month! That limit was obviously reached within just a few hours.

Bear in mind that I've been using it as my primary search engine across all platforms for the past 1½ months and all of sudden my latest search result returns a request for payment..

The email I received when I was accepted to the private beta in December said, "Kagi is free during the beta, but will rely on membership in the future." More recently, an email in mid-May notified users about the specifics of the shift to a public beta and the introduction of free/premium tiers. I feel they've done a decent job keeping their users apprised of the changes.
None of the announcements or even the billing page makes mention of what the "limited free tier" offers.

They should have been explicit in saying that it was going to be limited to "X searches per month" especially since it was going to be so restrictive at 50

None of the emails or announcements provided a link to the pricing page and the link to it isn't found on either the kagi.com homepage when logged in or the billing page.

In fact this is the first time I'm seeing a link to the pricing page.

Ironically, even searching "Kagi.com pricing" on Google doesn't bring up that page. Maybe if I could still use Kagi to search for that term...

Also, while I don't expect early adopters to be rewarded, it would have been a nice touch to offer them a discount or a grace period.
To be fair Kagi is often worse than Google in my experience.

If they are listening my perks are

- for non-English results it is significantly worse than Google

- the map is nearly useless and a joke. Names in Hebrew and Arabic appear inverted

- (browser's fault) required login in private tabs is a mayor annoyance. I wish I could pick say DDG for private tabs and kagi for normal ones

- (Safari's fault) the extension is nearly useless, it doesn't really register as an alternate search engine but it tries desperately to intercept requests to other search engines instead

Moreover for private browsing the

some of these criticisms are probably fair. I only speak english so can't comment on non-english results, but it wouldn't suprise me if the Kagi creators were primarily or only english speaking and step one is optimizing for english.

I'll also say that Google maps, while it has made some missteps in my eyes is the one good feature of Google (well, google translate is pretty good too.) and while I haven't kept up, last I heard, Google had over 1000 staff on the maps division, so, yes, I expect any mapping service is going to be worse than Google.

> required login in private tabs is a mayor annoyance

I tried to give Neeva feedback about that, suggesting to use either vhost auth or in-URL auth such as 67f3a96e-2805-4fb5-9f52-f049d918c584.example.com/?q=whatever or search.example.com/67f3a96e-2805-4fb5-9f52-f049d918c584/?q=whatever so the normal browser's query mechanism would work fine in both normal and private browsing mode

The map is currently still… suboptimal. They know it.

For non-english results, did you search with !reg or changed the region to something fitting whatever language you are searching in? I only speak German and I have no complaints with the results I get with !reg there.

Private browsing: There’s a URL in the settings (account -> session link), that contains your private session token. Using that, you can search in private mode as long as you stay logged in elsewhere.

The business model for search is sort of stuck between a rock and a hard place. Funding through ads demonstrably does not work very well. The incentives for providing good quality search just isn't there.

Subscriptions may work, it obviously won't be for everyone, but maybe it doesn't have to be either. Maybe it can be a luxury product for people with pockets as deep as their dissatisfaction with what's on offer for free.

Third option is what, donations? As someone running something like a donation-funded search engine and makes about $100/month doing so, it doesn't seem to work out very well. Admittedly it's a limited offering that focuses on a fairly specific niche of websites, and I'm not really going full Jimmy Wales in pleading for money on every other page.

People in general don't seem to want to pay for free services.

Like what's left, crypto microtransactions? I don't really feel that would work either if nothing else because most people just don't own cryptocurrencies to pay with.

If it were easy/cheap enough to host, can a model like shared game servers or web/email hosting work? People pay $20 a month without thinking for web hosting. What does it take to make "search hosting" a thing, where cheap search hosting companies can crop up both at the low end with bare bones offerings and others climb up the value chain with offerings like squarespace...
Cheap is probably a bit out of reach. I think right now, for hosting something like my search engine, you're looking at either a one-time cost of around $5000, or $200/month in server rentals. Maybe you can bring that down with the economies of scale, but it's never going to be anywhere close to $20.