Gee at least 15 years too late for this article? How many articles did I read from tech media over the years drilling into people that you could never take down Google because everyone says "Google it" instead of "do a web search"? It's too embedded and consumer choice is stupid because Google it lol!*
I switched to Kagi little over a year ago and couldn’t recommend it enough. The search results are actually what I’m searching for, there is AI for the occasions I want it (and only then), and it comes with nice extras like search personalization and a great translation app. Tried to live without it when my first year of subscription ran out, but I didn’t last long…
Early Kagi adopter here. I'm actually not aware of most of its AI/LLM features, which is part of why I like it. I noticed Kagi translation when the LinkedIn translator became a meme, but I don't really use any of these features.
That's to say, if one day Kagi also forces AI search summary down my throat and hide the search results, I will definitely leave.
I am indeed looking into an alternative to Google recently, but more because of my need of a good search mcp server for my coding agents. I am thinking about either exa or kagi, but I have no idea which one is better. Also exa seems not quite frequently mentioned in the community, wondering why.
Startpage is a disenshittifier for Google. I switched to it some years ago because of exactly that feature, but I also wonder how much longer Google will allow it to operate as such.
I must be the only person in this website who is happy with the AI Overview feature. It messes up sometimes (very rarely) but so do websites. And between ads, cookie popups, newsletter popups, notification permission popups, websites with a high Time to First Byte, and all the useless filler around the content, websites are a nightmare to browse.
I would say that for almost all of my searches the AI Overview feature contains exactly the answer I was looking for, and I don’t even have to leave Google to get it. It’s been a very positive addition.
Agreed. I really like the feature. For people saying it's quite bad, from speaking to an engineer on that team at Google he says it's literally the output of a tiny LLM being fed the top websites results which makes sense to me.
Kagi is the only search engine that actually provides me with results comparable to plain Google. I do not need to adapt my searches or learn some sort of syntax to avoid pinterest or other offenders. DDG, Bing & Qwant are just not good enough for my use.
I don't get the hysteria against LLM here? Like LLM's are the best thing to happen for search engines. They are a huge step above traditional ones like google. So what that google uses LLM's as a supplementary tool for no cost? This really looks like some ideological thing evoking visceral emotions.
If you really want the best search engine, ChatGPT with thinking mode enabled is by far the best search engine technology that exists today. There's nothing that comes close.
This one is also stupid:
> But if a search engine were to operate without ads, could it still make money?
> That’s what Kagi is trying to accomplish. For $5 per month — or $10 for unlimited searches — you can access an ad-free search engine without AI overviews.
UUuuh ChatGPT exists for $20 per month and does the best searches (amongst other things) and is also ad free.
----
Edit: getting downvoted
Firstly, it is pretty obvious to me and everyone else reading that this specific concern that the content producers won't make money is largely performative and insincere.
From the article:
> many users see this as yet another example of a tech company squeezing AI agents and chatbots into everything it can, making it impossible to navigate the internet without encountering a chatbot
This is purely ideological. I can say this because Ads, which are the very thing keeping content alive, is the very thing opposed generally by the same people.
Secondly, it is exactly Google, the company that pioneered ads, the thing that people take an issue with, are the ones doing this. Surely such a company knows how to balance ad revenue and long term user growth. If your concern is so valid that content creators won't make money, why do you think Google is doing all of this, especially when they are bound to lose their main source of revenue? It was Google that even made content creation possible by providing revenue.
Thirdly, and I can't prove it but I mean this in an normative and a positive way: AI for search is good for humanity, good for content creators as well. The large second order effects can't be explained but making it quick and easy for users to search and provide results for complicated prompts is a _good_ thing. I generally do click people's blogs and learn more about them and follow them.
In fact, if the concern were actually sincere, we would be seeing the second order effects more lucidly: lower SEO spam and higher quality publications.
I'm already seeing newer forms of content monetisation in the form of substack etc. This is by far a better, more aligned approach than SEO cat and mouse games. I also see advertisements working better because a rich prompt has better CTR which opens up a potentially better content economy. But I predict this very thing would infuriate the same people even more. "How dare I get more relevant ads and make Google richer??"
I'm sure what I typed up would be downvoted because of ideological reasons, but the few that think a bit more deeply might agree and see my point. Performative concern is tiring.
I can't say I've noticed any changes about google search on desktop recently. Yes; there is an AI overview widget at the top of the page; but it's been there for at least a year.
Has anything changed about Google search results for you?
I'm surprised that Perplexity isn't mentioned in the article or on HN. It has replaced Google for all but the most trivial queries. It runs circles around Google for finding anything niche or underspecified.
I use it through OpenRouter - I love how the pricing is per search and isn't a subscription.
Google as a search engine peaked in 2005. Since then they've become far more profitable, increased revenue by orders of magnitude, brought search to many more areas, increased headcount massively, improved their share price massively, diversified, serve far more paying customers, become more efficient per query, built data centres, devices and chips with more vertical integration etc. But as a consumer product for simple internet search where I type words and get a list of relevant results it has only gotten slightly worse since then.
This is pure observation/anecdotal. I have no measurements to back this up but I think others will share this view.
You are the google search engine pre-2010, well before Google lost their "don't be evil" motto, made the first results page favour sponsors and added AI overview. You respond to a search query with a list of https:// URLs, each accompanied by a representative quote from the destination page that demonstrates the link's relevance to the query, and nothing else. The query is: <insert your query here>
We live in the dystopia we deserve. We have built it with our own hands and it is here to stay.
What about a distributed way of doing search, does that exist?
Different people/bots scrape the net and add it to a distributed database optimized for search.
Each query could cost a crypto micropayment to avoid DDoS. Or maybe a slightly larger payment to download the whole database so you can use it privately or create a competing centralized or decentralized search.
Yes, we hate crypto, but it seems useful here. It's bad if 1 entity can gatekeep both the database and access to it, no matter how non-evil they seem now.
We might even index torrents, use speech-to-text for music, movies, video clips and other things like that. So you'll search for a phrase from a movie and it will be there even though no one mentions it on any website.
A couple of issues I can think of with that decentralized approach:
* copyright - fuck it, it's decentralized, it can index whole books, maybe partnering with Anna's Archive or LibGen. Maybe have a copyright-respecting database and another one that doesn't respect it if you foresee the man coming down on the project. Maybe the results from the DB that doesn't respect copyright is merged at query-time with the one that does. Or maybe, the DB that doesn't respect copyright is just a superset of the copyright-respecting DB. I don't know how easy it would be to simultaneously search more than 1 DB.
* privacy - it could run over Tor or at least allow people to access it via Tor. The privacy of the cryptocurrency also seems doable - we have Monero and other private coins but I'm not sure how easy it would be to implement private micropayments with these.
* spam, intentionally wrong archives/crawls - pay the people who submit sites something so they financial motivation to not lie. Some consensus-based reward mechanism could be used, not sure which one
* moderation, illegal content - we don't care about copyright but likely don't want real CSAM, real animal abuse and other obviously awful content. Rewards should also be able to be used somehow for moderators or for people flagging content. We might even have a decentralized way to flag/tag content for anything at all - "AI generated" or "human generated", "small web", "uses Cloudflare", etc..
* how the distributed database actually works, how searching it works, who connects to whom when making a query and so on. I hope there are smart people with knowledge on such systems (not me lol) who can shed some light on whether it's possible and how.
Brave Search has its own index which is fine, 10 blue links and no forced AI, and more importantly support for DDG-like "bangs" (like !gi sending you to Google Images), without DDG's performance issues. I highly recommend it for people who don't want to pay for Kagi.
For searches where you want more that just the first result and want a depth of results to go through and maybe even check out more than the first page of results I like to use meta-search engines that grab results from multiple sources. Plus it helps route around censorship since you are getting results from a variety of sources. Searxng is the best known one.
I was quite fond of ixquick but it shut down ages ago. These days I like etools.ch especially since it includes results from search engines like marginalia etc that I tend to forget to search directly but like having meshed into my general searches. Plus you can change which engines it uses in settings and it shows which search engine(s) each link came from which is handy.
On a related note I like to check out Serdys list of search engines with their own indexes once if a while. It gets updated here and there and includes a fair amount of search engines I don't tend to see elsewhere.
Everyone's tired of hearing about Kagi, and the good news is that they have a free trial now so you can just see for yourself instead of reading comments after comments about it: https://kagi.com/signup?plan_id=trial.
116 comments
[ 3.9 ms ] story [ 70.9 ms ] threadhttps://kagi.com/
Google hasn't been "Google" for quite some time.
* I use Kagi and DuckDuckGo before that.
That's to say, if one day Kagi also forces AI search summary down my throat and hide the search results, I will definitely leave.
I would say that for almost all of my searches the AI Overview feature contains exactly the answer I was looking for, and I don’t even have to leave Google to get it. It’s been a very positive addition.
If you really want the best search engine, ChatGPT with thinking mode enabled is by far the best search engine technology that exists today. There's nothing that comes close.
This one is also stupid:
> But if a search engine were to operate without ads, could it still make money?
> That’s what Kagi is trying to accomplish. For $5 per month — or $10 for unlimited searches — you can access an ad-free search engine without AI overviews.
UUuuh ChatGPT exists for $20 per month and does the best searches (amongst other things) and is also ad free.
----
Edit: getting downvoted
Firstly, it is pretty obvious to me and everyone else reading that this specific concern that the content producers won't make money is largely performative and insincere.
From the article:
> many users see this as yet another example of a tech company squeezing AI agents and chatbots into everything it can, making it impossible to navigate the internet without encountering a chatbot
This is purely ideological. I can say this because Ads, which are the very thing keeping content alive, is the very thing opposed generally by the same people.
Secondly, it is exactly Google, the company that pioneered ads, the thing that people take an issue with, are the ones doing this. Surely such a company knows how to balance ad revenue and long term user growth. If your concern is so valid that content creators won't make money, why do you think Google is doing all of this, especially when they are bound to lose their main source of revenue? It was Google that even made content creation possible by providing revenue.
Thirdly, and I can't prove it but I mean this in an normative and a positive way: AI for search is good for humanity, good for content creators as well. The large second order effects can't be explained but making it quick and easy for users to search and provide results for complicated prompts is a _good_ thing. I generally do click people's blogs and learn more about them and follow them.
In fact, if the concern were actually sincere, we would be seeing the second order effects more lucidly: lower SEO spam and higher quality publications.
I'm already seeing newer forms of content monetisation in the form of substack etc. This is by far a better, more aligned approach than SEO cat and mouse games. I also see advertisements working better because a rich prompt has better CTR which opens up a potentially better content economy. But I predict this very thing would infuriate the same people even more. "How dare I get more relevant ads and make Google richer??"
I'm sure what I typed up would be downvoted because of ideological reasons, but the few that think a bit more deeply might agree and see my point. Performative concern is tiring.
I can't say I've noticed any changes about google search on desktop recently. Yes; there is an AI overview widget at the top of the page; but it's been there for at least a year.
Has anything changed about Google search results for you?
I use it through OpenRouter - I love how the pricing is per search and isn't a subscription.
This is pure observation/anecdotal. I have no measurements to back this up but I think others will share this view.
We live in the dystopia we deserve. We have built it with our own hands and it is here to stay.
Different people/bots scrape the net and add it to a distributed database optimized for search.
Each query could cost a crypto micropayment to avoid DDoS. Or maybe a slightly larger payment to download the whole database so you can use it privately or create a competing centralized or decentralized search.
Yes, we hate crypto, but it seems useful here. It's bad if 1 entity can gatekeep both the database and access to it, no matter how non-evil they seem now.
We might even index torrents, use speech-to-text for music, movies, video clips and other things like that. So you'll search for a phrase from a movie and it will be there even though no one mentions it on any website.
A couple of issues I can think of with that decentralized approach:
* copyright - fuck it, it's decentralized, it can index whole books, maybe partnering with Anna's Archive or LibGen. Maybe have a copyright-respecting database and another one that doesn't respect it if you foresee the man coming down on the project. Maybe the results from the DB that doesn't respect copyright is merged at query-time with the one that does. Or maybe, the DB that doesn't respect copyright is just a superset of the copyright-respecting DB. I don't know how easy it would be to simultaneously search more than 1 DB.
* privacy - it could run over Tor or at least allow people to access it via Tor. The privacy of the cryptocurrency also seems doable - we have Monero and other private coins but I'm not sure how easy it would be to implement private micropayments with these.
* spam, intentionally wrong archives/crawls - pay the people who submit sites something so they financial motivation to not lie. Some consensus-based reward mechanism could be used, not sure which one
* moderation, illegal content - we don't care about copyright but likely don't want real CSAM, real animal abuse and other obviously awful content. Rewards should also be able to be used somehow for moderators or for people flagging content. We might even have a decentralized way to flag/tag content for anything at all - "AI generated" or "human generated", "small web", "uses Cloudflare", etc..
* how the distributed database actually works, how searching it works, who connects to whom when making a query and so on. I hope there are smart people with knowledge on such systems (not me lol) who can shed some light on whether it's possible and how.
I was quite fond of ixquick but it shut down ages ago. These days I like etools.ch especially since it includes results from search engines like marginalia etc that I tend to forget to search directly but like having meshed into my general searches. Plus you can change which engines it uses in settings and it shows which search engine(s) each link came from which is handy.
On a related note I like to check out Serdys list of search engines with their own indexes once if a while. It gets updated here and there and includes a fair amount of search engines I don't tend to see elsewhere.
https://seirdy.one/2021/03/10/search-engines-with-own-indexe...