[Kokouane]

Kagi is a tough pill to swallow. Their search is hands down the best around, there's no other way around it.

That being said, $10/mo is also expensive.

The workaround I found is using Kagi Ultimate. I get access to Claude (and I'm still able to attach files + access a dozen other LLMs) for $25/mo, so I was able to cancel Claude and keep Kagi and get the best of both worlds from either product.

Side note: incredible that a small team like Kagi's can somehow use LLMs more effectively in search than a company that has years of search experience (i.e. Google)

[pcchristie]

I love it too, and I do think it's expensive, too.

A lot of people laugh at thinking $10 (USD) pm is expensive, of course it's not huge money for most people. The problem is Kagi is a kind of "vote" towards moving the internet away from the ad-supported crapware trying to spy on your every click and capture your attention non-stop. If you're trying to replace a lot of free services with paid services to cast said vote towards shaping the web into what it should be, these costs really add up.

After paying for search, email, supporting a creator or two (e.g. a podcast), and some software here or there, you can easily end up in the hundreds of dollars, then you look back and notice that at best you've saved yourself from a few annoying ads and maybe gotten a fractionally better service and at worst your experience is unchanged and you're just deriving some intangible satisfaction from having not been spied on (which at the individual level doesn't make much difference unless millions of people follow your lead) or supporting a creator you admire.

It's tough.

[fhd2]

It is tough.

For my company, it's easy: I pay for the tooling my devs need. Overly simplified, I only pay 50-65 % of that, because expenses lower taxes. And compared to salaries, a few hundred $ is not a big deal. If I think about the time it saves us and how much money we can make in that time, it's a no brainer. Even just having people enjoy their work more pays off.

There's opportunity cost. Ad supported services are not overly incentivised to provide a quality product in the long run, so it's a safe assumption to make that they will waste your time to some degree, at least as they mature and enshittify. Some are great, eventually they all become bad, in my experience. It can be smart to use free stuff while it's still good, with an eye on migrating.

I don't know how much sense it makes to apply this opportunity cost thinking to your personal time. I don't really do that, but I do try to reduce time spent on anything that annoys me, and to do more stuff that brings me joy or pride, even if it's not economical. Life is short.

[tasty_freeze]

> That being said, $10/mo is also expensive.

It is more expensive than $0, but if you value your time more than a dollar an hour, the time saved is worth more than $10. I've found I scroll a lot less and have fewer false positive sites where I click in and look around only to find it isn't what my search was looking for.

That is just for the basic search feature. TBH, I haven't even investigated its other features, like lens and the ai stuff.

[ki85squared]

Can't say I understand how $10/month is expensive.

Quality search results ultimately save time digging through poor quality search results. Add up 300+ searches per month and surely you're hitting minimum wage value at least.

The value proposition is absolutely there at $10.

[threeducks]

Less than half of HN users are from the US and wages are lower in most countries, sometimes by a lot. Less than 10% in Turkey or Ukraine for example: https://en.wikipedia.org/wiki/List_of_countries_by_average_w...

[nicce]

Value proposition should be compared to the low cost alternative. Is it $10 better than Google? Maybe, I am not sure.

[ericd]

Just being able to rank domains (and nuke the ones that are usually spam) is enough to make it worth $10/mo for me. This is a tool you’re using constantly, so even small time savings per use adds up to a lot.

[ki85squared]

I am comparing it to the low cost alternative.

[bboygravity]

That (25 usd for Claude + Kagi) is the best sales pitch ever. I'm switching :)

Not sure what I'll do when Grok 3 comes out (I expect it to beat every other LLM out there hands down) but we'll see by then :p

[The_Rob]

What is it about Grok that you expect to be so much better? Not disagreeing necessarily, just want to know your reasoning.

[bboygravity]

Better LLM seems to be about who has the most compute and data for training.

xAI has (by far) the most compute and data now.

[rollcat]

> [...] incredible that a small team [...]

Here, this. Small, focused teams usually deliver more output per person (or even overall) than larger ones. Less management overhead, clear goals and responsibilities, tendency to employ people with cross-disciplinary experience, hiring for talent and not checklists, etc.

> [...] can somehow use LLMs more effectively [...]

LLMs are an incredibly effective tool for the few areas where they do fit the problem. But there's so much "AI" hype going on, everyone is trying to cram it into anything and everything, running around with a hammer trying to smash things just in case they turn out to be a nail. Even the old-time players (who should know better) can't resist the urge.

It's almost like oligopolies faced with changing markets tend to start collapsing under their own weight.

[viraptor]

Unless you're a really heavy user, you can possibly save a lot on those LLM bills by using the API and some third party app. (Like Msty for example)

[SV_BubbleTime]

This is true for me, except I don’t want to run another app, and I like using Claude on mobile.

But the API is tempting for a cost savings.

I put $5 in API for “Continue” extension in VS Code and it’s been months and haven’t used it all up yet.

[kevin_thibedeau]

There was a time when it was incredible that a small team like Google could somehow implement search more effectively than a company that has years of search experience (i.e. AltaVista)

[BryantD]

In that case -- and probably in the case of Kagi vs. Google as well -- it was entirely dependent on focus. In the hypothetical situation where your goal changes from "provide the best possible search" to "beat Yahoo," your available resources will be used on different things, and then...

[nottorp]

> Side note: incredible that a small team like Kagi's can somehow use LLMs more effectively in search than a company that has years of search experience (i.e. Google)

But Google is not a search company. It used to be, but now it's an ad company. I'm sure their LLM use serves their purposes right.

[Octplane]

I'm discovering that Claude is included in Kagi Ultimate and this is also slightly blowing my mind... Probably going to do the switch.

The fact that it's expensive is true but for me largely compensated by the niceties of the service. I don't really like the fact that they use Yandex either but the other search engines are not really satisfying for me anymore.

I've used Kagi Pro for several months now and it's working great for my personal and professional needs. The only thing I'm missing is the shopping features but I can with that and switch back to Google when I'm looking to spend my money on physical goods...

[david38]

Is $10 really expensive?

For most people, no. Can you think of $10/m you spend on something less important than search? A couple coffees, a sandwich, HBO, Netflix, a drink, using two gallons of gas recreationally, etc

[LUmBULtERA]

I almost never buy coffee. Rarely buy sandwiches out. I get HBO for $3/month now for 6 months and will cancel after. Netflix is $7/month, though the whole household uses that. Two gallons a gas can buy a lot of transportation to necessities for the kiddo. Though we have an EV and $10 gets us maybe 250 miles more or less of driving -- that's a lot.

[mrbigbob]

No, $10 a month is not expensive. However, the problem is the amount of products and services that in the past you could buy once or buy and if you wish to upgrade to the next version you had to pay again have become less and less and subscription services have skyrocketed.

[chronogamous]

Most people I know can afford some of the luxuries you list, but only barely. If you have to choose between having a drink once a month at a place other than your own home, and having an ad-free search engine that actually works, you'll find that many people are thick enough to go for that drink.

For context, this is speaking from the Netherlands, where housing is relatively expensive.

[carlosjobim]

A couple of coffees isn't a luxury, not even for the worst penny pinching Dutch miser imaginable.

[nottorp]

How many coffees can you drink in a day? How many services and even standalone apps that don't use server resources want a measly 10/month from you?

[xigoi]

I don’t drink coffee or alcohol, watch movies or drive a car.

[dsissitka]

> That being said, $10/mo is also expensive.

Back in the beta they planned on launching with a $20-$30/month unlimited plan and they didn't think they'd be able to bring the price down. That was a little too expensive for me so I moved on. I like what they're doing and I'd pay $10/month but I just don't have a use for it anymore.

[tomrod]

> a company that has years of search experience (i.e. Google)

From experience on the figurative side of this reality, I can attest that it is hard to build a track while the train is running on it.

[SparkyMcUnicorn]

Agreed on Kagi ultimate.

The agents (optional with a toggle) hooked up to the LLMs are fairly decent. They'll search, grab YouTube transcripts, read online dev documentation, etc.

[stormfather]

How does Kagi Ultimate compare to Perplexity?