Hacker News Reader

Comment by nullindividual | Hacker News Reader

If Apple could produce a high quality, privacy-minded search that respected common query operators, I'd jump for it. Unfortunately without robust ads, there's no money in it.

Bing as of a handful of years ago wasn't all that bad -- in fact, if you lived in the Microsoft enterprise sphere, it was arguably better than Google. Now that Bing belongs to the ad division, it's no better than Google search. Both platforms are defacto promoted ad engines with SEO gaming for results just beyond the ads.

And I'm aware of Kagi, but it's not for the wide swaths of whatever can be indexed out there... it's tough to stomach paying for a utility that has been free since the 1990s.

View on HN

Paying for Kagi is worth it, just to have an LLM not pop up in my search results.

And if you do want their LLM, just append a '?' character after your query. It's up to the users, not forced upon them.

For Google search you can add &udm=14 to the search URL to block the AI Overviews, or use verbatim search (tbs=li:1).

Kagi is free, just need to create a new account every 100 searches.

Even if that's true, I wish you wouldn't do that.

I'd really like for Kagi to be a success.

By doing that, you’re actively contributing to enshittification—you are the enshittifier.

I paid for Kagi and used it for 3 months because the only way you could evaluate search is through “vibes”.

I’m glad it benefits others but for me it was not much different from Google (with ad block).

In two ways it was an inconvenience compared to Google — location search (Kagi respects privacy so it doesn’t do location, and its default maps are not as good — though google maps are one click away, it was an extra click each time.) searching for where to watch movies on streaming platforms was also missing so I had to do !g.

I wish Kagi well — it’s by no means bad, in fact it’s an admirable product. But it wasn’t for me.

In my experience, if I want location results I just go directly to maps.google.com or the map app on my phone. Kagi was great because when I google "tacos" I usually just want recipes (or the wikipedia article), not the location of the taco bell down the street.

Personally, I find it impossible to stomach continuing with ad-based models. I want zero ads in my life. Also, with newer AI the old search model is done for. Search will shift, should shift, to AI agents doing the search along with reading through various pages before giving accurate summaries with links. In that model users won't be exposed to ads as many searches won't even require them going to the end pages.

Thus, there needs to be new models for how things work, both for end websites but also for search, models not based on ads.

I agree with you about ads.

But I couldn't disagree more about search. AI search is a lot like asking your friend. Your friend who relies on their memory and refuses to look things up online. Searching with AI requires too much brain power because you have to verify everything it gives you.

Of course, that's becoming the case with ads embedded in the search results, too. Kinda bleak, really.

The new generation of AI search will likely use traditional indexes or RAG followed by the AI reading the results to do higher level filtering. If you look at how search gpt, kagi experts, or the hugging face llm search works I think that will be the direction.

Possibly even personalized agents could do this locally. You'll have to verify exactly as much as you do currently with old-style search, but ideally with much higher quality results and ranking, and no ads. For simpler things likely the answer will be directly trust worthy.

Personally, I just don't trust anyone to summarize content for me. Call it my own weird quirks and biases, but I'm maybe 50/50 on conforming to populist sentiment, and if it's a story I even mildly care about I don't want to take that coin flip. I'll just take the time to at least skim the article, then read the abstract of any acedemic sources it points to.

The AI agent will subtly serve you ads. OpenAI calls it “publishing partnerships”:

https://news.ycombinator.com/item?id=40310228 “Leaked deck reveals how OpenAI is pitching publisher partnerships” 303 points by rntn 88 days ago | hide | past | favorite | 281 comments

I would be surprised if paying for the AI agent would stop it from also taking money from advertisers, just like buying a copy of Microsoft Windows doesn’t stop meaningless ads showing in the start menu etc.

The Kagi AI agent does not. And a locally running one will not. Like it or not for those willing to pay the ads are going away.

> Personally, I find it impossible to stomach continuing with ad-based models.

Ditto, but I also feel the same way about subscription models.

There’s plenty of money in it. Even if the ads were minimally personalized (ie use location so that when I search for a plumber I see local plumbers) just the keyword-based ads would probably be enough to cover the costs. Apple users are exactly who advertisers want to reach.

Apple has shown in the past that if something is considered a key technology, they would rather own it than hand that power to a key competitor. They made their own CPUs, they made Maps, and they could make search.

They can do it, but is it going to make them money? Google's $20b/year payment made up 20% of their income in 2023. 20% of their income was just free money, and their dev team doesn't have to do anything for them to get it.

Would Apple or their shareholders accept a minimum 20% loss in income, plus whatever the expenses would be to build a Search and Ad platform, plus the continued expenses of maintaining and updating it?

Google is paying Apple $20 billion for $60 billion worth of traffic. Apple said it would cost $6 billion per year to build and run a search engine. I don't think it would take much of an ad platform to pull in enough money to make it work for Apple and their shareholders.

It all comes down to whether or not Apple sees search as a key part of their phones and so far they haven't. FWIW, I think they're right - it's not a key part the same way Maps is, for example.

$100B was Apple's net income, not gross. Revenue was ~$400B, so we're talking about 5%, not 20.

Still, 20% of their net income is guaranteed free money. I doubt there are minimal (if any) expenses on Apple's side to maintain the status quo of Google being the default search engine.

> it's tough to stomach paying for a utility that has been free since the 1990s.

No such thing as a free lunch. I'd rather pay to not have ads splashed in my face than have some corporation sell my data to the highest bidder.

And an open source search engine isn't exactly a feasible endeavor (e.g. one that can be produced in 5-10 years without some reliance on existing tech). But I'd happily donate if anyone wants to take a gander at such an initiative.

One way you might fund a search engine is if you needed the index for some other busineds purpose (say, Siri). It will have a cost, but if you can defray the crawling and some processing cost, it becomes much easier to stomach without ads.

I honestly think a crypto-micropayment solution (a la brave) would be overall a net positive. An easy way with send pennies or less with virtually no overhead fees.

It's shame though that crypto is to most people now, well, you know - crypto.

Let's be concrete here. What's a "privacy minded" search engine exactly?

Do you mean you don't want the search engine to correlate your query with what you clicked on? or any other feedback from user interactions to update its ranking?

Good luck with that.

Or do you want a free search engine without any ads ? You might as well wait for a company to send you money every week.

> Bing as of a handful of years ago wasn't all that bad -- in fact, if you lived in the Microsoft enterprise sphere, it was arguably better than Google.

That is probably not enough though. Products that are just a little better than something that is good enough usually fail. I'm reminded of all the analysises of why plan 9 failed.

network effects are strong. At a certain critical mass, only the company itself can detrhrone the company. Otherwise, people will mostly settle for the familiar over the true best.

I think that is stateing it too far in the other direction. If you have a killer app, you can definitely dethrone the current king. I think it just has to be something much more significant than just marginal improvements.

To be a bit more subtle: you can make a better competitor and be successful. But it's really hard to "dethrone the king" as you put it. Twitter is a great example of a social media actively destroying itself in ways that'd already put a medium sized business in the red, but its competiton as of now is far from a threat. Reddit as well, but there's not even a consensus best alternative there.

Subtle point #2, this is mostly for the general consumer. You can and will piss off professionals or businesses to a point where you shoot yourself in the foot Unity was the most recent high-profile example, and Adobe is slowly losing doting over its issues.

Xerox, Hoover, Amana, IBM PCs, Scotch tape, and many other totally dominant brands have been successfully planted by moderately “better” competitors.

Reply on news.ycombinator.com