Chatbots are replacing Google's search, devastating traffic for some publishers

> People will pay for it.

I would love it if it were true, but sadly, the data doesn't support this. A lot of local newspapers did real journalism relevant to their communities. However, the local newspapers were the hardest hit by the social media wave and few remain today. Fast forward to now, you cannot get any real local news easily.

The avg person never really valued real journalism to begin with and the hyper targeting/polarization of social media and closed echo chambers has made it worse.

>People will pay for it.

I'm willing to pay, but not by individual subscriptions per news organization. I'm more interested in following journalists than news organizations.

Active webrings still exist surprisingly:

https://wiki.archiveteam.org/index.php/Webring

Nah, Webrings were an extension of the link page ... but not the same thing.

The Link page was curated by the site operator and usually a linear list. IT's main goal was to say "Hey, this is cool, too".

A webring was more like a collective, whereas individual webring members did not necessarily know or agree with every other site in the ring. And it usually was not a list either, but more of a mini topical directory, often with a token-ring-style "Visit the next / random / prev site" navigation you could add to your own page. Webrings were already geared to increasing visitor numbers to your own page ("Others will link to me").

Oh, those were easier times.

Early Google PageRank was notorious for how much additional trust a given page had based on many links back to it existed. It was why certain bloggers had massive ranks early on, because they would be in big webs of conversations with lots of high quality links out and back in.

Early SEO did weaponize that and broke it for everyone.

You can't blindly trust sources, either. Or, sometimes, you ability to understand the sources correctly.

I think of LLMs as bookworm friends who know a little bit about everything and are a little too overconfident about the depth of their understanding. They tend to repeat what they have heard uncritically, just like so many other people do.

If you don't expect them to be the ultimate arbitrer of truth, they can be pretty useful.

Seems to my untrained eyes like Google's AI search is actually the best on the market, no? Seems like a lot of HN users have trained themselves not to type queries into the search prompt anymore and then complain about the quality of a product they don't use.

That’s a cheap argument. Even with high quality results (which I still think Google is the best at), LLMs are infinitely easier to use.

I'm surprised the links work for you at all. 90+% of citations for non trivial information (i.e. not in a text book but definitely in the literature) I've gotten from LLMs have been very convincing hallucinations. The year and journal volume will match, the author will be someone who plausibly would have written on the topic, but the articles don't exist and never did. It's a tremendous waste of time and energy compared to old fashioned library search tools.

And what happens when you follow them?

In my experience, the answers tend to be sourced from fringe little blogs that I would never trust in a Google search.

Google at least attempts to rank them by quality, while LLM web search seems to click on the closest match regardless of the (lack of) quality.

One thing I did once with great success was asking chatgpt something like "I'm trying to find information about X, but when I Google it I just get results about the app named after X. Can you suggest a better query?"

X was some tehnical thing I didn't know a lot about so it gave me some more words to narrow down the query that I would not have known about myself. And that really helped me find the information I needed.

It also threw in some book tips for good measure.

So yeah I can highly recommend this workflow.

> I think this will probably just create a strong incentive to cover important things and move away from clickbait.

But clickbait is how they make money...

That's like saying, "Oh, Apple will just have to move away from selling the iPhone and start selling hamburgers instead."

I mean, sure, but they're not going to like it, and it's going to come with a lot of lost revenue and profits.

I find myself regularly copying URLs, sending it to Gemini, and asking it to answer what I want to get out of the article.

I'm not wasting my time scrolling through a mile of website and 88,000 ads to find the answer to the headline.

That's impossible unless the web reverted back to 2010, when walled gardens weren't prevalent, making your own blog was common, doable and often done by those without programming experience, forums were alive and well, and people wanted to share things on the web rather than group chats.

> No one wants these "AI" summaries

Not true, I use them all the time. They have links available for when I want further information, which is not very often.

Then what explains people doing millions of web searches on perplexity/chatgpt/claude?

Simply untrue. I don't want it back. I use ChatGPT's voice transcribing to do 99% of my searches today.

Google does need to adapt or die

they are losing more and more search to “AI.” my 12-year old never uses Google and couple of times I asked her to “Google it” she literally rolled on the floor laughing and called me a “boomer” :)

The first part of that statement is valid but the second one isn't.

If anything, most of big tech has shown exceptional humility against new threats

Instagram incorporating stories (Snapchat)

YouTube incorporating Shorts (tiktok)

Google search incorporating AI Mode (perplexity et al)

This is in stark contrast to Kodak and the likes who scoffed at digital camera and phone cameras as distraction. They were sure that their ways were superior, ultimately leading to their demise.

Yahoo had 90%+ of the search market and they lost it in a few years to google because they were unable to innovate. I don't think anyone saw that coming. Everyone was building "portals" (remember those? AOL.com? I think verizon.com was one at some point with news and weather) to try and compete with Yahoo's dominance in search. It can happen again. LLM Chat is certainly an existential threat to the googs. Part of the OpenAI lore is that google originally viewed it as a threat to their search/advertising revenue model and defunded it on that basis.

The fact that people are willing to pay for LLM and use it over search seems to indicate that Google's free product isn't as good, and llm chat is better "Enough" that people are willing to pay for it.

You think we wont start seeing ads or paid for refs/links in those AI responses? Not defending Google here, when they turned that feature on I posted to some friends "another nail in the coffin for the web as we know it" or something to that effect.

What is infobox?

They have a few interfaces that you can trigger an AI answer through

The first, and most useful to me, is when you just append a ? to your query. This pops up the answer in an info box at the top of a search, and then shows related results below

The second is the assistant mode, which pops up either when you use the continuing link at the bottom of an info box, or trigger it directly via the assistant URL. This is the standard conversational interface

Kagi assistant is hybrid search / LLM

110% this.

Real web search is disappearing locked behind large datasets unavailable for normal users. The AI screen ensures you’re fed exactly what they intended while siloing off the web more and more to block competition. All the while signing exclusivity deals which should realistically be illegal (try finding current Reddit results anywhere but Google).

AI based interaction makes it much easier to manipulate users into buying your items as they add a layer of human-like trust on top of the machine. It won’t be long before prices are hidden behind LLMs generating prices based on who you are. I’m already noticing ChatGPT becoming more and more enthusiastic about any product it thinks I have the potential of buying. Try asking it if something is a good deal, 9 times out of 10 it will say, “Yes, go for it.”

We don’t want to live in a proprietary world where LLMs exist. This technology needs to be open. The search data needs to be open and not walled off to only monopolies. This is an inflection point.

Do you trust an ai run by a shady company but not an attributable human editor.

I wasn't aware of dia browser, thank you for sharing. I've been doing browser automation for a while and have become convinced the way to accomplish human in the middle is to fork Chromium and create a custom browser. However, there is one problem, there are 3.5B Chrome users. Getting someone to install a browser extension is hard enough, a whole new browser much, much more difficult.

I went with experimenting with stock trading as a demo because people need huge incentive and value to make the critical jump to install an app. There is potential for niche curated business intelligence in trading or real estate, for example, where not only providing the chat bot but also time series data embeddings ect.

I'm a little sad because I'm late to this party.

Nevertheless, there needs to incentive also for people to continue to publish the data, ideas, and information so the chatbots are going to have help the content creators help them curate and provide the data by getting users to navigate to the webpages.

I remember reading Google's Search Engine Optimization guide back in 2009 when I built a news publishing website for an industry newpaper. The tone was here is how to optimize your website for google crawlers to help us help you get traffic to your website. Google is nothing without people creating.

As an experiment, because p47's social media posts move the markets $60 in a day, and the last thing on Earth I want to be doing is reading them so the system makes an API request for any new ones, then checks for links, video, and images. It uses OpenAI whisper running with transformers.js on the local machine using webgpu for the inference to transcribe the video and audio and image to text for ocr. I tried to do the text generation locally but any decent model although will run caused my Macbook M3 to get so hot I could cook a steak on it while freezing the rendering for the whole computer.

image-to-text and video, audio-to-text works fine, there are lot of uses for text generation that work but to get high quality analysis to see if a social media post might cause the stock market to crash requires sending the data out to an api. If the side panel requires searching for links to navigate to it requires a third party api.

Working with it, I think the next hardware race will be getting these models to run on personal computers in next 2 - 5 years and I have a suspicion Microsoft is ahead of Apple.