Hacker News Reader

Comment by achierius | Hacker News Reader

It's fundamentally self-destructive though. In time, the sites which rely on search clicks for revenue will essentially cease to be paid for their work, and in many cases will therefore stop publishing the high-quality material that you're looking for.

View on HN

I assumed that, after having using LLMs myself increasingly, that LLM's killing search was inevitable anyway. Further I assume that Google recognizes it as well and would rather at least remain somewhat relevant?

Google search, as others have mentioned in this thread, increasingly fails to give me high-quality material anyway. Mostly it's just pages of SEO spam. I prefer that the LLM eat that instead of me (just spit back up the relevant stuff, thankyouverymuch).

Honestly though, increasingly the internet for me is 1) a distraction from doing real work 2) YouTube (see 1) and 3) a wonderful library called archive.org (which, if I could grab a local snapshot would make leaving the internet altogether much, much easier).

By high quality material do you mean the 15 ads per page, ultra SEO optimized, content length extended posts that will give you a mediocre carbonara recipe?

Ad web has never incentivized quality and never will.

Sites rely on ad impressions for revenue. I block ads anyway so either way they aren’t getting money from me.

And if ad supported content ceases to exist, nothing of value will have been lost. I’m not morally opposed to advertising, I find ad supported content not worth reading especially on mobile.

>And if ad supported content ceases to exist, nothing of value will have been lost.

"Absolutely nothing on any website that relies on ads for revenue is of any value" is a wild take.

To be more precise, it might be air value but not worth the trade off of ads - they take away the value.

My sense is (call me naive if you like) but if the thing was of value, it would survive (or be reborn in another form) without ads. Donation supported sites (looking at you Wikipedia) or just passion projects would continue.

So you believe in free lunch.

If the ad sites die, the AI crawlers will rise the hosting costs of all others sites buy putting more traffic on them.

So either they start adding ads too or the sites die also

No, I pay for Apple News, Slate, Stratechery, Downstream+ (podcast), Accidental Tech podcast, etc.

I donate to electoral-vote.com via Patreon. Most of my news comes from podcasts and those ads don’t bother me as much as web ads. I said I wasn’t morally opposed to them.

I don’t use any app that is ad supported without the option to pay for an ad free experience the same with streaming services.

That's not my point.

Thanks to AI crawlers the traffic costs will rise and all the sites who were ad free either need additional income, most likely ads, or will shut down.

Services like CloudFlare are offering better services all of the time to block web crawlers.

>I don’t use any app that is ad supported without the option to pay for an ad free experience

That is, obviously, still ad-supported content...

Before web advertising the entire web was a "free lunch" — if by free lunch you mean individuals putting their sites up; their only compensation being some degree of joy at spreading and sharing their passion.

The early web was eccentric and kind of … "shambly", but I would not be upset if we returned to that.

Before AI crawler web traffic for sites was less.

So now individuals putting up their sites face the risk of higher traffic fees.

The “web” has been ad infested since shortly after Netscape was introduced. Are people forgetting about the X11 pop under ads and the “punch the monkey” banner ads.

ChatGPT lists clickable sources in a lot of nontrivial queries. Those sites don’t even need to pay OpenAI for the traffic (yet). If you ask „what’s happening in the world today“, you might get 20 links. How is this worse, exactly?

How many people click the links? What happens to LLMs if people don’t provide training data anymore because nobody visits their sites?

They will either pay for it to be generated or get good enough at producing synthetic data that actually improves LLM quality.

So either even higher costs and hope that a bug problem of LLMs get solved somehow.

Given how much data they need that will be pretty expensive, I mean really really expensive. How many people can write good training data and how much per day?

Doesn’t sound sustainable.

Cloudflare publishes a "crawl-to-refer" ratio, which can be used to estimate the traffic from LLMs:

https://radar.cloudflare.com/ai-insights#crawl-to-refer-rati...

Most of the time when I find a good answer from search it's one of a few things:

- Hobbyist site

- Forum or UGC

- Academic/gov

- Quality news which is often paywalled

Most of that stuff doesn't depend on ad clicks. The things that do depend on ad clicks are usually infuriating slop. I refuse to scroll through three pages of BS to get to the information I want and I really don't care if the slop farmers die off.

Reply on news.ycombinator.com