Search engines alternatives now that Google isn't Google anymore
techcrunch.com
techcrunch.com
If you're questioning the AI features, know that I am only barely aware they exist. I have never, not even once, accidentally or otherwise, engaged the AI features without going out of my way to do so. I've never seen what their AI is like. I have no idea what it's for or why I'd want it.
It's beautiful. Kagi has AI I suppose, but it's over there and not in my face. I don't think I've ever seen an AI nag in the UI, but their UI itself is also over there and out of my way.
Thank you, Kagi, for staying politely the hell out of my way. I love you.
Something like "how can I access the root filesystem of proxmox?" and you'll get an instant answer instead of a link to the whole ass Proxmox documentation.
An ending question mark enables fast answers, like Google’s AI summary.
!ki sends your query to the assistant on light research mode. It runs a few searches against their index and summarizes the results.
I typically don’t need more than that. Most stuff I just find through search.
Maybe shopping is the weak area, as Google does get product feeds and Kagi doesn’t. I don’t think this bothers me at all.
I am truly baffled (and annoyed) about this fact.
Third-party keyboards, still not usable but browsers are basically ok.
The one third-party keyboard that seems to work is the one from Google, if you want a better experience than Apple’s.
There's more to device quality than whether a monkey can operate it and looks shiny.
Obviously this has security implications, but I don't ordinarily search for anything sketchy on my iPhone so I'm personally not too worried about it.
Anti-drm tools are a big case in point. And so is Bypass Paywalls Clean Firefox plugin. All of these have been purged from the "Great American Corporate Firewall".
Russian people != Russian govt != Russian companies.
I just use the tools that work.
Not at all.
Ethicists, in my experience, are hacks that cant actually do, but instead loudly pontificate and deride others who can. They basically run gish gallops and whataboutism to get the real creators to stop.
> "Who am I to refuse medical breakthroughs because they came from monstrous human experiments on prisoners of war? If it works it works!" /s
You can sarcasm-tag this all you want, but the medical experiments the Nazies did against Jews, Romani, Gay, mentally disabled, and others WERE used in this way for the US space and rocket initiatives and the creation of NASA. They were already dead, and the data collected. So yeah, we did use them. And we got further ahead by not having to run more experiments. And the data helped save our people.
Its not a great situation, sure. But whats done is done. No sense in whining and bemoaning "ethics" of corpses. Might as well get some good use out of their abject slaughter.
Many things on the web use Yandex.
It sounds like they use everything to give their subscribers good results. Which is what it sounds like I am paying for.
It's not prejudice when it's based on a post-facto assessment of the Russian government's mobilisation of their companies for obscene and evil goals like clamping down on free speech, persecuting LGBT people, or trying to destroy Ukraine.
"It's Russian and therefore it's out" is a valid stance in light of all the known consequences of using a Russian product. If Russian people do not like this, they are welcome to break their links to the Russian state by emigrating and founding companies elsewhere, or stay and overthrow their government. Either works for me.
Many Russians have learnt to keep quiet about politics so they can get ahead in Russia, and they seem to harbour some deep-seated delusion that everyone from abroad should play along with this for their own convenience and profit. They whisper 'no war' to one another, by which they mean Ukraine should surrender already, so that this whole unpleasantness (to them) blows over, and the rest of the world goes back to accepting their blood money. No.
Germany has been grappling with its own horrific genocide for a hundred years and still hasn't quite figured it out. Russia is, as of the time of writing, currently undertaking one. Come back in a hundred years, maybe we can talk about Russian products then.
Even when Google gets it right (often they don't) I have to wade through a bunch of AI slop and countless ads. After that, it's dozens of SEO referral link sites trying to sell me garbage.
Kagi gives me better results by default, no ads, and I can customize results by prioritizing or blocking different domains. Very much worth the small price.
A good search engine shows you the results for exactly the query you entered. A good search engine does not discard your entire query to show you what it thinks you meant.
A good search engine supports and respects modifiers and advanced queries: AND, OR, NOT, quoted strings, plus and minus. It gives you advanced parameters like publish/retreival date, categories, origin.
If there are no results for your exact, specific query, a good search engine shows you zero results. A good search engine does not waste your time with literally infinite fake results tangentially related to what the engine decided you meant instead of honoring your query.
I gleefully pay for Kagi because it's a good search engine. It does all the things a good search engine should do. It does not do the things that bad search engines do. Kagi has put great effort into designing a search engine that maximises your ability to find the thing you're looking for.
Because I pay Kagi with cold hard cash, I do not have to instead pay Kagi by consuming thousands of fake "sponsored" search results. I enter a query, and Kagi gives me precisely the results for that query and nothing more. It's really that simple. I pay for Kagi because it's a good search engine, and most other search engines are not good.
Google isn't even a search engine anymore. It's a heuristic ad delivery mechanism that happens to also surface indexed web content. Your query is a suggestion and is usually disregarded completely and replaced with a more lucrative query. Google hasn't supported advanced search in over a decade. No plus or minus or quoted strings make any difference.
I heartily encourage you to learn about search engines and try one. I doubt very strongly that you've ever used a proper search engine. You've most likely only engaged with heuristic ad machines before. Try a search engine, you'll never want to go back.
I'm also REALLY happy to see that you can actively disable not just adds but also suggestions to install tools/apps and newsletter requests. Thank you!
I only use Google to search for reddit posts.
The rest is ChatGPT or Claude.
Search is always faster than asking an LLM if I have a general idea of what I am looking for. I may consult an LLM if I want to compare things or kick off deep research, but most of the time I find myself having to go back and forth with it and correcting assumptions it made.
According to my Kagi stats, I am averaging around 3k searches per month.
I can’t help but feel that you are really missing out on a lot of results when just relying on LLMs for search.
I'm with the GP. To me LLMs are just better search engines. In the most literal sense. They have their own index and can generate links if you want them.
https://news.ycombinator.com/item?id=48269481
By “their own index” i assume you mean training data? If so, thats a big part of the issue for me. As an example, if I ask an LLM for some part of the Zig stdlib, I will get an incorrect answer 10/10 times because it will refuse to look up the latest documentation.
ChatGPT has effectively assimilated Google search. You can tell it to look up the latest version of the zig docs and do whatever.
Regardless, if we stick with the example of the Zig docs, using the search and then opening the actual docs is a much better and faster experience for me. I get the context I need, I don’t get the verbose llm output that packages it and I get there faster.
Mind you, I use AI a lot and have subscriptions to all of the major models, but this is just not a use-case I find useful in my workflow.
I also keep trying to use it this way by looking for things such as decent libraries or frameworks for something new im trying out. I’d say that at least 50% of the time I end up searching for myself and finding popular ones that were completely ignored by whatever LLM I searched with. Always a frustrating experience…
Looking through some of my history of today:
- “github rate-limits”
- “oriental hornet”
- “riva 88 florida”
- “logistic map”
- “zeiss euv mirror”
- “authentik helm”
Just to name a few…
For a lot of these, LLMs would slow me down significantly. Most of the time I already know exactly what im looking for.
Shopping and finding locations is definitely one.
I participated actively in the Google Local Guides. I depend on the content from there to discover stuff geographically. But I’m also aware reviews and interactions and even images that look human can’t really be counted on anymore.
Slow gradual growth before, large increase in the daily growth rate since.
It'll be interesting keeping an eye on how that growth rate goes over time. :)
Astroturfing implies that Kagi is paying for people like me to praise them, it's just a good product (for my personal use at least), and I'm glad to recommend it while it stays good.
I want Kagi to have enough customers to run a business and earn a profit but not so many that they need to make the product worse to continue growing.
Then again maybe you're the only correct person on the entire internet and everyone else is shills and crazy people. That definitely sounds more likely.
All that's to say: when I paid for Kagi, I thought I was investing in additional search infrastructure, and didn't realize Kagi had no aspirations to build their own general purpose index, and instead primarily aggregate results from other indexes, either adversarily (Google, Bing) or not (Yandex, Mojeek, Brave, Apple, etc.) I understand they do maintain their own small-web index, but I thought their aspirations were higher when I first jumped on that train.
I stopped using YouTube 10+ years ago, so no clue if it still the case.
Also, I don't really understand what your catchy phrase means. Google's management is being malicious by not forcing their employees to not be lazy?
There was this idea born in the late '90's/early 00's that everything digital should be free. The internet was dominated by teenagers with no job and no credit card, so it made sense.
But the result of that has been a whole generation with an allergy to compensation, and the inability for anyone to compete with "free" services, even if everyone hates that service.
Congratulations, this might be the single most trivially-disprovable statement I've ever seen on this site
Google won because it was a single text box. Yahoo lost because it full of ads and pretended to be a phone book. Linux won in the server world because it was free and superior, Windows lost because it's shite and expensive.
I could go on, but before I do that I'd have to be convinced I'm not replying to a 27 year-old who just graduated business school.
You can get some development and innovations built purely on "free", but without actual professionals who can make a living by developing these systems, they never take off to reach the masses. The best example is social media and the Fediverse.
But even if the data point is not good to support the argument, I don't think one could argue that Linux succeeded by "being free". If Linux was a "serious threat" in 1998, it was because there already companies looking into it and willing to make back it up financially to help its development.
It was getting so much commercial and academic engagement that we had the idioms (cliches?) of the "LAMP stack" for basic web servers and "Beowulf clusters" for high performance computing. Even SGI was already revealing a Linux plan, before 2000, when they still seemed like a fixture of the HPC industry rather than an also ran.
TANSTAAFL does not need a qualifier to apply. "Nothing is really free, so whatever you got 'for free' from a community member or some non-commercial effort was bound to have limited reach" is more like the point I'm trying to make.
This is the way FOSS is meant to work. I got jobs where an employer was happy to run other people's FOSS software "for free", happy for me to contribute bugs/requirements/patches back upstream, and happy to release our own projects under FOSS licenses.
It is a win-win for all involved. That's the whole point of it.
This is like trickle-down economics for FOSS and it doesn't work.
To me, it is no different than management, planning, logistics, marketing, etc. which is done for the purpose of supporting some other objective.
It simply means that you perform software development as work-for-hire in support of that other objective, rather than for the purpose of licensing revenue. It provides wages for services rendered, just like the vast majority of other job types.
It just doesn't provide for scalable virtual rent extraction for a "publisher" or other middleman. To me, that is a benefit of it. It removes a bunch of perverse incentives from the table. Incentives that tend to harm the developers and users for the benefit of those middlemen.
The one way to get out of this mess would be to have the market paying a premium for companies that do R&D in FOSS directly. It can not be a secondary goal, and we can not be telling them they shuold find some other way to make a living.
And prior to whatever IBM did in 2000, I already had a job deploying Linux and BSD systems in production at a corporate job.
Yeah, and it was a BBS ran and backed by a software development company that used it as a channel to promote and sell their software. IOW, they were not offering the infrastructure "for free".
> I already had a job deploying Linux and BSD systems in production at a corporate job
Which means that there was someone paying your employer to support it. Again, not doing it "for free".
I think you and the others responding to me are just trying to disprove the specifics of my comment but entirely missing the meat of the argument: I am far from being "a 27 year-old who just graduated business school", but I agree with GP said: people will not pay for digital services unless they absolutely have to, so companies that try to make a living by offering a quality service in exchange for payment will invariably lose to someone that offers their product "for free" but exploits their customers elsewhere.
I remember a colleague around 1998, he said: "how will they ever make money? Its just an empty website?"
LOL
When large portion of moneyless teenagers grown up into indebted to death adults, there is no wonder they stick to lure at free services rather than unaffordable services.
About 70k people are paying at least $5 a month. I've been using the $25 a month plan for nearly 3 years now. I imagine Kagi is doing alright.
I’m one of those 70k people and support Kagi, and I also strongly believe in companies succeeding and sustaining themselves on a small scale like this. I think our economy would be healthier if it was made of many, many small companies, not a few massive ones.
But we can’t argue Kagi is anything more than a super niche product, for now. :(
That and the fact Universities provided free, fast and unmetered internet access. I doubt they would be running anything if they had to pay $1/hour like regular people had to in their dial-up days.
Kagi employee here. We're actively working on building our own indexes beyond the limited ones we have now, not just a general index but also purpose built indexes for things like programming, etc.
I'm not sure antitrust will help you.
Lip service. You'll have some token index of Wikipedia or something so you can say your results are "a blend of our own index and other sources".
https://blog.kagi.com/waiting-dawn-search#:~:text=Wikipedia,...
I imagine there is still a whole load of stuff out there on the internet that Google would never surface because it doesn't have enough adsense or whatever. Are you finding that?
I mean it sounds like that already has a lot of overlap with our Small Web indexing efforts, so that part of our indexing efforts could be an extension of that. A lot if this is still in development though so I can't speak on specifics just yet.
> Teclis is an attempt to surface the less known web, the web of creativity and self expression, the more humane web.
> Teclis includes its own crawl as well as results from Kagi Small Web index and results with permission from Marginalia Search.
> Teclis works best with broad queries such as 'machine learning', 'vegan diet', 'religion' etc..
Is there another crawler doing the general-purpose stuff?
For everybody else there’d Google I guess.
Not the least bit surprising to me. I had the misfortune of talking to Kagi's CEO several years ago. Every word out of his mouth was a lie.
Kagi's the one search company I trust less than Google.
I'd also argue that calling a tech CEO a liar is far from extraordinary. It'd be extraordinary if I accused him of honesty.
https://blog.kagi.com/waiting-dawn-search
You said that you were not surprised that kagi was lying, only that they were not in this occasion. When you accuse somebody of lying it makes sense to provide at least some evidence of that.
At the very least, they are very clear about which indexes they use and how.
The search infrastructure you're talking about is a natural part of that, but, like any infrastructure, it scales the organization it's supporting. Kagi is tiny so their "original infrastructure" contributions are tiny.
Put another way, you essentially were investing in infrastructure, but you were hoping for major infrastructure and what is happening is small infrastructure. Kagi would probably need to get much bigger to be able to do the infrastructure you're talking about. (And if they were much bigger, it should be natural -- at a certain scale it will make more sense to do your own than work with someone else's.)
"... I thought their aspirations were higher ..."
It sounds like the decision to send search queries (and money) to Kagi was based at least in part on reasons other than the quality of the search results
This is interesting psychology
What if all (cf. only sum of) the money sent to Kagi was actually invested in an alternative way to search the web without using an index created by a corporation or a non-profit with commercial subsidiaries
Defensive HN replies may focus on the quality of the search results from commercial indexes, e.g., "Google is the best. That's why everyone 'chooses' it."^1 But if the consumer is choosing Kagi based on other reasons, e.g., "investing in additional search infrastructure", then clearly there is more to these decisions
For example, some search engines claim to be planting trees or some such. Nothing to do with the quality of the results
1. Apple is being paid 20+ billion for choosing Google as the default in iOS but Apple's choice is not based on the money. Yes, that makes sense
We don't need to feed more the Google monster machine more than it needs to.
- they issue you a code when you are logged in
- they track that code for multiple use
- all they can do is claim that the code was securely generated and it isn’t just an API KEY to your account… but they’re already telling you it is database tracked
How can you have any even the 1/2 best proof it isn’t just an API KEY that directly links to your account? I see no trust path other than “us, bro”.
Does Kagi have the equivalent of either of these privacy mechanisms? Even with the limitations of the ProtonVPN approach, that would be an improvement. As far as I've seen Kagi has no equivalent: it's closed source and has no regular third party audits for privacy.
I get a weird feeling when I see people googling things using Google (hehe), the amount of bs is mindboggling.
I end up doing a lot of searching with Mistral Le Chat (also a subscriber).
What I'd like to know is power cost difference between the two (on the server-side). Ie. is Mistral sustainable financially or are they also running on vc / burning money. Although France uses nuclear, so it is a drop in a bucket I suppose.
(In most browsers, you can input any URL with %s as the query string.)
A negative is the high latency.
(Looks like Mistral is not profitable yet[0]. It expects 1 G$ revenue for 1 G$ capex in 2026[1], so it is moving towards profitability, but to be fair it is building a couple datacenters.)
[0]: https://www.forbes.com/sites/iainmartin/2026/04/16/how-franc...
[1]: https://www.bloomberg.com/news/videos/2026-01-22/mistral-ceo...
Kagi is better for research and knowledge work, Google is still better for quick lookups.
*except for maps results, Kagi is absolute hot trash for maps. I automatically append !gm to the end of all mappy-type searches. I wish Kagi would just kill their map product and redirect to google maps.
I got the Google AI Pro plan, which gives Gemini access by default with generous limits, and also includes free credits, code assist in VS Code and other editors, and also to Gemini CLI. And I'm just simply using that for all of my needs. It seems to work quite well so far. I see how Google Search is not relevant for me anymore.
You can ban Pinterest links, boost Mozdev, ban listicles, boost whatever.
Kagi gets very good very fast as you customize it and it's easy to keep it updated as sites go up or down in quality.
The community shared boost and ban lists are a great resource too. Making it easy to see and copy what others find useful.
Do not fund the Kremlin!
The feature I missed most from Kagi was domain filtering, so I had Claude write a quick userscript for DDG that lets me boost, pin, and block specific domains. uBlock Origin aside, DDG even lets you turn off ads natively.
Kagi is good, but the redirection felt a bit flaky lately, and I was dealing with an annoying bug where my localization kept defaulting to Groningen for no apparent reason.
I’ll stick with this DIY setup for a bit, though I might well end up back on Kagi once I realize how good I had it.
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'm working on a self-hosted search service called Hister with the same goal when I started Searx development: reduce dependence on online search engines.
Hister is a full text indexer for websites and local files which automatically saves all the visited pages rendered by your browser. It provides a flexible web (and terminal) search interface & query language to explore saved content with ease or quickly fall back to traditional search engines. This is a fundamentally different approach than what Searx follows and solves most of the weaknesses of metasearch engines. Of course it has its own weaknesses as well, but most of these are not conceptual and can be resolved by improving the software (and datasets)
I've been using it for a few months and as my local index is growing I can avoid relying on external search engines - and even websites listed in results - more and more frequently.
The initial reception is overwhelmingly positive with already more than 30 contributors and hundreds of contributions. Currently it can help with "recall" type searches mainly, but I'm planning to provide pre-indexed thematic datasets and I'm drafting a peer-to-peer index sharing concept. Maybe you can find it useful as well (or at least have some constructive criticism =]).
Links: - https://hister.org/ - https://github.com/asciimoo/hister - Background/motivation/beginnings: https://hister.org/posts/how-i-cut-my-google-search-dependen... - Small read-only demo: https://demo.hister.org/
- If I wanted to use use my domain list to start hister, to download my preconfigured / like domains?
- Can I make some pages to rank higher in it?
- Can I assign tags to pages (by which I could later on filter?)
My domain index
Yes, Hister has a built-in crawler which supports standard HTTP and different browser based backends
> Can I make some pages to rank higher in it?
You can create priority rules to boost the ranking of the matching domains/URLs
> Can I assign tags to pages (by which I could later on filter?)
It is possible to add a label to indexed documents
Assuming it indexes everything locally and falls back to traditional search engines if none found, how do you feel about adding a shared middle layer? A layer that simply indexes all the canonical data that doesn't have any personal info. This way, the contributors can automatically contribute the pages they index - building a shared search engine over time! The whole thing can work without a crawler of its own (under appropriate license so people can trust it)
I'd appreciate any kind of help designing such system. We are on IRC/Discord/Github/Codeberg.
A VPS with without a black listed IP is good. A simple rootless container, update is easy.
Configuration takes little time, not much.
I still hate that I have to double the bang to use the same bang as DDG.
Example: "!!wde Ente" to go to the German wikipedia page about duck instead of "!wde Ente" with DDG.
The only feedback I have is the initial indexing from my large history was rough. There were a lot of domains that kept blocking me for exceeding rate limiting or wouldn't let me index at all. I could see it being useful to import a history file and organize it by domain inside some sort of temporary database to track/distribute attempts and get a more detailed report on complete domain failures.
Regardless though - great work!
I agree, browser import has rough edges. The issue you mentioned is known: https://github.com/asciimoo/hister/issues/31 . I try to prioritize it and find time to fix it.
Hister sounds like a idea I had years ago but gave up on after running into issues with index size taking up way too much storage.
Long ago I've used Searx and really liked it but after some point didn't see the point as opposed to using Google more directly. But lately in the back of my mind I've thinking about checking in on it again.
So if i only use the Firefox extension, all pages Hister will "fetch and store" will have gone through my browsers content blocker (uBlock Origin) before being saved ?
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.
"Yeah, it sounds like a very common copypasta back in the early 2010's, related to XYZ"
"So... uhh... provide me a link please"
"I can't provide links"
"You're a search engine"
"I don't have a current connection to the Internet."
"Well can you give me any examples of anything even vaguely resembling this topic?"
"Yeah, like in the Reddit thread titled ABC, where $two_paragraphs_of_description"
"I can't find that Reddit thread"
"I'm sorry! I hallucinated that it was a common thing, but actually you told me a unique thing that you just made up."
"This sounds like a contradiction. Where did you pull the information about the Reddit thread from?"
"I can't link it, but it was on the sub _____ and it was titled _______ and it talked about your thing at length"
"When I google that I find nothing"
"Sorry! I hallucinated that I knew the link. Actually there is no link, there is no discussion like that, and the thing you provided was totally unique."
I then proceeded to try Googling various permutations of the topics for the thread it brought up, which kept giving me 10 nonsense results, and a grouped collection of Reddit posts that it would not expand on / separate.
In my case I suspect the original uploader took down their videos from YouTube or there was some legal process involved. But it was very weird for the Gemini answer to confirm exactly what I was looking for but be unable to articulate in a way that helped me. Totally bizarre, as if the topic was ablated from the LLM.
I would much prefer getting a straight answer, "due to copyright I can't discuss this" or something.
I think a large part of the blame is not on Google but on the websites themselves. The Internet has been enshittified by a gargantuan amount of spam sites and content mills created just to generate clicks and boost SEO.
At least AI offers a way to filter out the noise at the cost of relying on how it was trained and what the creators thought is good data.
They constantly reward websites that are on a hamster-wheel of chasing the latest SEO trends, while penalizing websites that have actual information for not jumping through the same hoops.
A company I know operated TWENTY THOUSAND blogspam blogs out of a single server/IP. Google knew all along that this was happening (the companies had strategic partnerships) and never did anything about it.
The last thing they did anything significant was when, Panda in 2011?
At this point it's clear they're a monopoly and only care about websites who cater to their whims + making money. Search be damned.
When the search results are bad, usually I can at least tell that they're dubious: either they're from obviously unreputable sites, or they conflict with each other, or they just don't quite address my query. But an inaccurate AI overview can look very similar to an accurate one.
For daily things - finding restaurants, looking up my kid's symptoms, etc. - I still use Google
i have the opposite experience as i've found it to mess up a lot to the point that I can't trust the AI Overview answer at all. I've seen it be confidently wrong too many times and now the trust factor just isn't there.
I made a serious error with that once.
(I agree that AI result summaries would be quite handy, if Google's results weren't also so frequently incorrect at a basic or numeric level)
As a former blogger: I hate it. But I knew years ago where things were heading and stopped. No point in blogging/ writing etc.
And since now they explicitly aim to never drive any human traffic to any websites, it will only become worse.
I don’t like AI only idea but I think it will work just fine
Yes, but most of these things are results of adtech having so much impact on the web/how we publish/consume/get paid for the content we create. I'm a bit bitter/sad about this.
I did shit that was career-damaging, such as deciding not using heavy libraries that a platform team was pushing me to do (thus affecting their OKRs). I had to deal with designers complaining that removing JS didn't allow the special animation they wanted. The works.
We did it, and we managed to even score a few positions in Google.
Right after that, someone from Marketing added a few dozen random "pixels" to Tag Manager "because that's how Marketing is done", forcing us to adopt a cookie banner and bringing the lighthouse score to half of what it was.
Notice: The pixel was collecting data but it wasn't being used by anyone. There weren't ever any marketing campaigns involving the website.
Eventually the person was fired by the CEO for not bringing results, and we managed to remove the pixels, but by them the damage was done.
Have you actually used it? Because sometimes is doing heavy lifting there and very rarely is flat out lying. It consistently messes up and hallucinates.
I disagree with “sometimes”. But anyway, the gargantuan difference is that with websites you can get a feel for their credibility. As a simple example, documentation on MDN is miles more trustworthy that your average SEO spam blog, and you can see this as soon as you enter the page. Yes, some scammers are craftier than other, but the signal is there.
With LLMs, all answers get the same weight.
And asking for sources is not reliable. They are too often made up or contradict what the page says.
> and I don’t even have to leave Google to get it.
And what will you do when most posts on the web are just junk SEO spam to trick LLMs into telling you what they want? It’s not like that’s hard to do, even.
https://www.anthropic.com/research/small-samples-poison
https://www.bbc.com/future/article/20260218-i-hacked-chatgpt...
I don't believe this for a second. It has constantly the worst output of any serious AI I've seen, by far. It's laughably wrong sometimes, usually just wrong. It can usually cope with mundane keyword searches where it's still better to just read the wiki blurb, because even those can be mangled.
I think the model must be very lightweight since they’re automatically running most search queries through it and a decently powerful model like Gemini would cost far too much in compute.
Websites accepted Google scraping their content because it gave them a prominent blue link plus excerpt to drive traffic. Now everyone’s content is blended together and maybe, if they’re lucky, their site is chosen amongst the blend to get a tiny citation link.
This is the way the internet used to be before it was enshittified, you just type in some keywords and the first result was probably the most relevant and readable source of info. But now, no more.
I think if a modern search engine could deliver the same experience but with organic human written content, I’d probably use that. This is probably a new niche now for upcoming search engines, focused on finding human created works.
If I search topics I am knowledge about, the overviews are almost always at least slightly wrong.
Not all websites are correct sources of information, but I am generally aware of which websites are trustworthy and can cross check.
The small/fast model used for the overview isn’t smart, but it’s still pretty great at helping me find the function I need and syntax. Best of all, clearly and plainly, unlike many documentation pages.
Then you are extremely ignorant about the topics that you research, because I don't think I've seen a single AI overview that did not have a mistake, and I don't think I've seen less than half that got a critical fact wrong.
Website owners losing traffic to the AI Overview feature is well-documented and has been discussed extensively here. This reality begets a few questions:
-----------------
What happens when it no longer becomes profitable for anyone to share knowledge online?
AI companies have scraped physical books to build their training sets. What happens when it no longer becomes profitable for anyone to share knowledge at all?
Are you okay with a handful of corporations owning the truth?
Are you okay with that same handful telling you _what the truth is?_
-----------------
I'm totally fine with Google telling me where to find truth. I am, personally, NOT okay with for-profit corporations like them telling me what the truth is. Every AI search I do is tacit acceptance of what they're doing, so I try not to use it.
All that said, you're definitely not in the minority outside of this site.
Heaps of people take the AI Overview result as the truth, case closed.
Given its inaccuracies (I used it the other day as a last result to locate these Korean cat treats, and AI Overview said that it was a "health food" for "wellness"), it's a massive problem that we haven't fully realized yet IMO.
That we will go back to the time when posting stuff on the internet was done out of passion and not because of economic interests. I can’t wait.
> When I’m yearning to search, I ask myself why.
> What answer am I hoping to hear?
> What answer would be a surprise?
> What would I do in each case?
I often use Perplexity for more complicated questions though.
In both cases, I really like that. I can actually ask a complicated question and get a reasonable answer.
For example, check the answer to "what is the source of vertical straw in the movie The sixth Sense with the flying taxis and all?"
(I said sprawl but the voice to text understood strawz so I left it)
We live in the dystopia we deserve. We have built it with our own hands and it is here to stay.
I will say off topic that, speaking to an early googler, there is actually documentation of meetings where they discussed what "don't be evil" meant and decided actual business options they should and should not pursue. It was not just a motto or a "code of conduct", but meant as and used to justify consequential actions.
Pre-2010 Google search didn't use https by default, almost no one did besides specific cases, like processing payment. And even then, only the critical part was https, the rest, like images was plain http. So, for a true pre-2010 experience, you want http:// links.
Post-2010 Google played an important role in pushing for https. From boosting https search results, to Chrome being annoying to unencrypted connections, to sponsoring Let's Encrypt, to forcing HSTS on their TLDs.
I kind of miss http, it was a time when the web was a public thing, a place for sharing, not for keeping secrets. But to be fair that's just nostalgia, the modern (commercial) web that generalized encryption enabled is so useful and convenient that I can't imagine going back.
That's a bit unfair. Not all of us who live in it had a hand in building it. In fact, very few of us had the leverage to fight against it.
Even if you don't have much power individually, as a society we could stop this any time.
"Their hands are tied!" I have heard so many moderates assert. "The customers demanded it!"
[0] https://www.google.com/search?q=google%20code%20of%20conduct
[1] https://abc.xyz/investor/board-and-governance/google-code-of...
[1]: https://kagi.com
[2]: https://uruky.com
We also have a no ai version: https://noai.duckduckgo.com/
As I fiddle with the options, I was about to suggest there should be a private way to share settings to another device so I don’t have to redo everything on my phone, then I looked at the right side of the page and saw you already provide that!
Overall great work, major thumbs up. Looking forward to trying some searches with you again.
I do think I may have found a bug. Blocked sites aren’t being set across devices, neither by using the cloud save nor the bookmarklet.
The only gripe I have is that the default maps are painful to use. On my system (Linux + Firefox), for some reason, half the time they fail to load, and need a refresh to be displayed.
Also, would it be possible to add the option to specify a region for the "Shopping" tab, in the same way it can be done for search & images?
Currently settled on Kagi but the price is a little steep, happy to support EU based services.
Normal people are using AI for search more already, Google is just trying to stop their primary business from completely disappearing.
I'd rather just have results without chatbot fluff.
They are transforming google search into chatbot but they already have a chatbot and by dping that they are losing google search product completely.
I miss times when search used to be fulltext search (also in youtube, ebay, etc). Now it is some kind of embedding vector search skewed by marketing and my previous visit history. It often shows completely irelevant and useless results and it is much harder (even impossible sometimes) to find something even though I am sure it exists (and I am able to find it eventually but it takes much more effort).
The UI has gotten a bit clunkier over the years, but it's still good, still more focused than Google's.
As I write this, I give Google Search a quick try and notice that the first thing you see is a full-screen cookie banner!? On my laptop, I even have to scroll to reach the Reject/Accept buttons, and keyboard controls don't work at all. I can't believe people still use this crap.
Their search engine isn't the best but none of the search engines I use are, not even Google.
So to me, they've already succeeded. Can you elaborate on the issues that make you think they haven't?
[1]: https://www.zdnet.com/article/brave-deemed-most-private-brow...
They seem to have walked back some of this, but it seems pretty unclear what and why, and I'd rather not try to find out given there's nothing stopping them from doing such things.
No thanks, I like ads blocked.
https://web.archive.org/web/20260523024427/https://techcrunc...
Some of my favorites are:
I also use Grok and Perplexity
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...
So, now I have a nice video clip of "The Ballad of the S.S. Forgetful", which it turns out aired near the end of Sesame Street Episode 2130. :)
This is pure observation/anecdotal. I have no measurements to back this up but I think others will share this view.
Let me correct.
Search was good as late as 2010, when they changed the engine to facilitate an "instant search, search-as-you-type" feature. It was decent until around 2015, when a pivot to privileging "brands" poisoned the results. And it's been useless since the pivot to semantic search (in part to facilitate Home/Nest products) and a series of index purges over the past few years.
We had something very good, and we had it for a good while, and it was destroyed by a company that was a blackhole for investment and some of the brightest minds of my generation, sucking up untold amounts of labor and radiating little but "exceptionally deleterious to society" particles.
The idea is simple: collect search stubs, short documents with routing information about a topic. When you want to search something your agent finds the right stub and looks up where on the web to go find that information. Your information is always fresh, while a search stub changes slowly, over years, because the entry points tend to be more stable.
If we use a 4B local model and another 4GB for search stubs it could be portable enough to download and install. As you use it you also generate your own search stubs on top of the generic package. A stub could contain links to news feeds, high quality hubs, search engines and of course, actual websites. High quality search stubs can be generated with any frontier LLM piggy backing on its agentic search capabilities.
I think the idea of managing a collection of search stubs as a replacement for centralized search engines is important because it would wean us of one of the last centralized points of the internet. Google plays many ranking games on top of users and publishers, serving their own interests first. I want out of that arrangement.
local LLM + local stub index + proxy == anonymity
They intercept a non-trivial portion of the web's traffic and presumably are as equipped as any other company is to build a good index. They're also the only the only company that has an interesting alternative incentive structure for creators.
And at least I know that I am happy to talk about a product that I pay for. Is some of it because I feel like there's a sunk cost - the amount of money that I've paid into it? Yes, of course there is. Is it a good search engine? Yes.
Do I wish there were features added to it that they've promised for a while now? (namely being able to pay for more AI credits, especially if you paid a year in advance?) Yes.
It's a paid-only service, without ads - "you are not the product", that lets you hide results from popular (to the mainstream) sites like Instagram and Pinterest, and to filter out low-quality sources like W3Schools, while raising MDN and the Arch Linux wiki to the top.
Google still maintains a web search mode that's free of AI overviews/chat exhortations (as well as ads, if you use an ad blocker). https://www.google.com/search?q=foo&udm=14 is the format of the search URL, and tenbluelinks has instructions on how to use it as your default engine on various platforms.
That said, I've stopped using this as a founder. While I personally like the web search results more (if I wanted synthesis of results, I'd use dedicated agentic-loop-capable tools that are a hotkey away), it's far more important to understand (and empathize with) our users' experiences, good and bad, when they use Google in its full AI extravagance in practice.
(What do you search) they stink vs. Google even though Startpage is proxying them.
DDG is _not_ a “good” search engine — please, anybody have a hundred side-by-side screenshots to compare identical searches?
Edit - also admit Kagi’s great, I’m not affiliated; if you have money Kagi should be the pick (ideally purchased via their more private payment options probably)
Edit: have to do the work, get the screenshots, & analyze my own patterns. How often do I just need the first link from any search engine (like searching $majorBrand to look for their homepage), how often do I search something esoteric where DDG falters, how often do I search something essentially unique but simply not indexed by Bing (DDG) even though it was submitted to Google just fine, etc.
Honestly, the answer is so often a little toy search like Marginalia or going straight to the website in question now, its frightening.
Their business practices are just the worst. But are the first 10 blue links they show you usually bad? (Like you know that one of those 10 should be the exact thing you want, so you decide to rely on another search engine instead, & the alternative search succeeds where Google failed)
(Google could give better results to those unloyal to them, as one example, so testing needed! We could be getting wildly different results or just have starkly different usage tendencies, there’s just no way I could disagree with so many of y’all on something this basic assuming no DDG astroturfing or anything unless we’re looking at things differently one way or another)
And obviously there is no astroturfing here from DDG. That you could have seen from a tiny look at my profile. I have a pretty public persona on the internet.
But often I'm searching for a phrase inside quotation marks and DDG hasn't crawled enough sites and gives me 0 or 1 result while Google gives me 5-10 results. Sometimes Google completely disregards the quotation marks, though.
Microsoft Bing bungles quotation marks so hard (???).
DDG catches the fallout.
> Sometimes Google completely disregards the quotation marks, though
Years ago I remember Google asking the person lodging a similar complaint for an example of a query because they found there was always an explanation. Noticed this no longer holds as of perhaps a few months ago if I’m not mistaken. Even this* fails:
+”omg just tell me no results if this exact string isn’t present come on I even put the plus sign”
Infantilizing for us, maybe optimizing for the 99.5% in reality (understandable, annoying)
*edit, made up example based on what I believe I’ve seen this year
edit:
> DDG brings up relevant results
Have you noticed, not if you misspell a proper noun? Spellchecking e.g. brand names, new/fad current event topics is apparently really hard (IIRC Bing not perfect here either?)
DDG was never good at quotation marks so I forgot to mention it also sucks at these.
> Have you noticed, not if you misspell a proper noun?
I haven't but I'd really prefer 2 options:
1) I may mispel thinks quote freqwently but gimme any sorta relevent result
2) I know what I am searching for. Maybe add a "Did you mean X?" link but still show me the results containing the misspellings. If I search for "motorolla" or "mottorolla" (misspellings for "Motorola") in quotes, DDG gives me 1 incomplete page for that misspelling. Google (StartPage) gives me mostly misspellings but with "Motorola" included in some results. Without the quotes both engines treat me as a moron who can't type. If I search for
"mottorolla" phone 2026
DDG still shows me "mottorolla", not "motorola" while Google has decided I am indeed someone who can barely type even though I've put in quotes.
I think the reason is my searches are almost entirely long-tail searches that Kagi's index just isn't good enough for. I am never searching for something like "best mattress" or anything else that is heavily SEO'd - it's always something very specific - so the result page in Google looks pretty much exactly like the Kagi page, only it nearly always has the result I'm looking for where Kagi's doesn't.
The main issue I've had with Kagi is that using "before:" and "after:" just seems weirder than it does on Google, and will throw in some stuff that's visibly outside the ranges I selected sometimes.
any chance you have an example of such searches off the top of your head that you wouldn’t mind sharing?
That's me, and Google correctly on the first page finds ONLY pages related to me.
Kagi pollutes the page with many results for my doppelgänger, the much more famous Daniel J Drucker of the University of Toronto, even though none of those results mention mclean.
That is literally the thing Kagi was supposed to be better at - actually respecting your search terms instead of thinking it knows better!
I still use and recommend Kagi but the results have gotten worse, but I also think that’s just an indicator of the health of the web as a whole.
I'd say about 1 in 5 searches for me is apparently "long tail" and I've never been more aware of how dead the internet is. I regularly still compare to DDG and Google, and spend about 300% more time on the others trying to find a match and discovering I have more results only because they made my quotes terms optional or ignored my time restrictions.
The current catalog covers 100+ providers, 1000+ universities, and 250,000+ courses.
Originally a weekend side project that I first shared with the world here on HN itself on Nov 29, 2011 [1]. Currently bootstrapped and profitable.
I've been using Ddg and Brave for general search and Yandex for deep-ocean expeditions (of the jack sparrow variety) and topics that US tech giants censor. I am looking for good Chinese search engine so that I can search things that Yandex censors or when Yandex's bot detection goes crazy (I get blocked with infinite captcha about 25% of the time).
The only thing Google remains good at is local search. If I want to buy something locally, nothing else comes close.
Also, real support! Your email is important, you should be able to open a ticket and reach someone that knows what they’re doing.
Fastmail is... FAST. The page refresh is basically instant, and the UI is super-snappy, well refined and something like 2026, while Gmail hasn't changed in... 15 years? It supports a lot of features (identities, masked addresses, ...), not to mention that I can have my domains properly hosted on it, for 5eur/mo.
There’s so much content getting buried now.
If you’re looking for anything remotely niche or legally gray, like sports streams or ebooks, you’re often better off using Yandex or you’ll never find it.
The old Google search engine that used to properly index and surface the open web has been gone for a long time.
I'd guess due to compute constraints, AI overview will struggle to reach truly great quality. That said, for now I find adding this section at the top still useful to me. The broader decline in Google's search quality is the bigger drag on me.
Google hasn't been "Google" for quite some time.
* I use Kagi and DuckDuckGo before that.
The last use case I had for google was Google Scholar, but it now appears to block anyone who blocks google tracking. But this is where ChatGPT does an excellent job of generating lists of technical papers and reviews and it interprets natural language queries with no problems. The kind of complex logical search queries google used to support (what, 15 years ago?) can be written without strict logical language (! & | ()) and all that. Pubmed isn’t bad for cross-checking and simple searches. And if you put sci-hub into the yandex.com/search box and click on yandex ai it tends to tell you where the current active sci-hub sites are, which is handy.
I would rather not.
Edit: clarifying that this is not strictly due to ads. I think the article itself is an ad judging by the slug 'six-search-engines-worth-trying-now-that-google....'. Usually such articles include a plug. I am not disabling adblocker to read that plug
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?
https://www.google.com/search?q=marine+nav+compass+card
vs
vs
https://duckduckgo.com/?q=marine+nav+compass+card
It's all products and ads now with Google with anything vague enough that warrants up-selling for them.
Maybe this is a frog in boiling water situation but… it’s just the same search as it’s always been, there’s just now a chunk of Gemini up top, that you take with a grain of salt, same way you’d take the promoted results. If you don’t like it, Adblock it.
When I ask myself honestly, has google search gotten worse over the past 25+ years? My answer is… ehhhhhh… not really?
Other people have always claimed google was ‘getting worse’ and I just don’t see why it should be any more true now than it was then. Isn’t this just the latest round of whinging?
Digging through the crap in a Google search is like finding a recipe in a blog. Sure, it's there, but the experience isn't great. I'd rather pay for a better experience, like buying a cookbook or subscribing to nytc. If you can afford it, why wouldn't you?
- Organize the world's information
- Don't be evil
From https://en.wikipedia.org/wiki/Don%27t_be_evil :
> Between 21 April and 4 May 2018, Google removed the motto from the preface, leaving a mention in the final line: "And remember... don't be evil, and if you see something that you think isn't right – speak up!"
That sounds like NY to me.
I remember the first time I copy-pasted something off a major website, in quotes, because I wanted to follow up on it, and got no search results. Not even from the site I copy-pasted from. "Search" has been in name only for a very long time.
Build directories again.
It incorporates results from both Google and Bing: https://support.startpage.com/hc/en-us/articles/452243553384...
Google and others are racing to integrate AI but - for me - there are two hidden costs. 1. Access to the web becomes mediated and monetised even more. 2. Energy consumption up to 100x normal searches has an environmental impact.
This is going to happen whether Google does it or not. The toothpaste is out of the tube.
I'm afraid of them selling out at this point. If they go, I honestly don't know who will go in their place.
I changed my default search engine to: https://www.google.com/search?q=%s&udm=14
Author should mention that you can change your browser's default search engine.
Before, search engines provided links, links provided the info we were looking for. Now AI provides that info without the middle man, links are just a footnote in case you need them.
It just doesn't seem to surface much reddit content. I know I know reddit, but on niche topics they often have some good discussions/links/experts.
I use it through OpenRouter - I love how the pricing is per search and isn't a subscription.
It's an AI like Google's "AI mode", in that it also surfaces URLs. I have not found it to be a good search engine replacement.
Also an extremely happy Kagi user but can't speak to their API, haven't tried it yet.
you can simply set the search as you like and save as a search url [1] that can be set in the browser once and forget.
their bangs are useful to search directly in a specific site (e.g. useful workaround for reddit search results not coming anymore). when i tried kagi, i noticed that they have adopted a similar system, which has to be inspired.
my overall search experience has been poor even before the GPTs, but this is just fine for almost everything.
[1] https://duckduckgo.com/duckduckgo-help-pages/settings/params
The three dominant English search engines with their own indexes are Google, Bing, and Yandex but this list has many spiders and engines that traverse the web.
I feel like this is a scene from a movie where a character goes "what are you talking about? Look around us, it's already gone" (idk why i am picturing Mark Whalberg for some reason).
But it's all just reddit, medium, twitter (twitter pages, w/e those blog things are).
The internet as we knew it is gone, man.
I think Google is getting its act back together.
Seeing as Apple has reportedly partnered with Google to power their LLM offering, I don’t expect that to change soon. I hope I am wrong.
Plus, my tabs open with the Start Page, showing my Reading List and iCloud Tabs, so I never even see the homepage.
&udm=14Read the article.
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.
DuckDuckGo is about to surge!
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.
If you successfully block adblock (tall order is tall), a lot of people really do just go do something else, instead of resigning to the ads firehose. And adblock users are still (somehow) in the minority, I think.
Also, this loops right back around too Google's ad monopoly. They have a stranglehold on both sides of the market, able to maximize spend from marketers and minimize payouts for those showing the ads, to maximize Google's profit at cost to everyone else.
Before theft-by-Gemini-Summary, there was Google News, with Google just wholesale copying articles into Google's feed reader so they could collect the ad revenue instead of the writers and publishers.
They've abdicated any right to complain about copyright violations of their own IP, at this point. Either copyright law is the law, or it isn't: can't be both ways. In practice, lawyers cost money and Google has much more than Random Adblock User 6, but morally speaking, they have no high ground to speak of.
Anyway... If Google doesn't any longer drive traffic to websites, then the operators of said websites will no longer have a reason to allow Google to index them in the first place. You can't have a very effective search engine if too many major sites block its crawlers.
I don't think the AI bubble is going to last, but if it did, I expect this all would end up compounding the "LLMs training on LLM generated content and churning up other LLM content" spiral into ever more useless drivel.
If nobody goes to the websites, those websites no longer get traffic or revinue. In very short order there will be no more websites from which to scrape, and the AI will no longer have new data to summarize.
Where do you think this ends?
By eating the source of their results, pretty soon there wont be any sources that aren't crap.
Ironically, if you ask an LLM this question it gives a much better answer than you have so far, although both of the ones I used still miss the larger impact on the web.
1. A search result’s accuracy is attributed to the source, whereas an LLM’s summary is attributed to the LLM—especially in the cases where it “cites” sources which don’t make the claims it emitted.
This makes no sense. Don’t ask the llm for summary. Ask it for the links directly. In what way are they different? ChatGPT is a superset of a search engine.
Here’s an example of me using it: https://chatgpt.com/share/6a159c95-3bfc-832a-bfed-52188a548d...
If your fundamental problem is the accuracy of ai overview I’m with you. But AI mode works well.