Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Generative AI could make search harder to trust

wired.com

Post reply on HN
Print view View original

231–240 of 255 posts

Re: Generative AI could make search harder to trust

#231

Post by tornato7 »

wildrhythms wrote: post #219
naillo wrote: post #215

Earlier quoted context omitted.

It's funny that one argument openai used to keep their models closed and centralized is so they could prevent things like this. And yet they're doing basically nothing to stop it (and letting the web deteriorate) now that profit has come into play.

I don't understand how OpenAI thinks they could stop that from happening.

Not saying they should, but if they wanted to they could have an API that allows you to check whether some text was generated by them or not. Then Google would be able to check search results and downrank.

Re: Generative AI could make search harder to trust

#232

Post by tivert »

12_throw_away wrote: post #41
happytiger wrote: post #17

Earlier quoted context omitted.

AI has so completely disrupting Search that it’s destroyed leading platforms effectiveness in a matter of months. But because of its current lack of optimization for accuracy, we shouldn’t consider it disruptive because it’s not yet proven technology? You can call it dangerous but you can’t call it useless. It’s also only going towards improvement from here, including drastic reductions in hallucinations. You have to…

> It’s also only going towards improvement from here Why?

>> You can call it dangerous but you can’t call it useless. It’s also only going towards improvement from here, including drastic reductions in hallucinations.

> Why?

A pseudo-religious belief in progress, especially the technological kind.

It's bullshit. If it were true, 2022 pre-LLM Google would have been better than 2010-era Google, but it most definitely wasn't. Consumer printer technology, for the most part, has been getting worse for decades at this point.

Re: Generative AI could make search harder to trust

#233

Post by semi »

heavyset_go wrote: post #43
BitwiseFool wrote: post #33

Earlier quoted context omitted.

Those simple web 1.0 sites made by college professors are a gold-standard in my book. I always enjoy finding them in search results. Although they are becoming increasingly rare.

Can't prove it, but it seems to me like black text on white background sites from the past are poorly ranked compared to sites with "modern" layouts.

They're even better than black text on white backgrounds. They're unstyled and use your browser default styling. granted it's rare anyone configured those so it's almost always black on white.. but for people who do specify their own preferences it's really nice to have them respected and not rely on hacking in my own css or js to override theirs

Re: Generative AI could make search harder to trust

#234

Post by bpiche »

pseudosavant wrote: post #12

I wonder if there will be a human information/knowledge equivalent of low-background steel (pre-WWII/nukes). Data from before a certain point won't be 'contaminated' with LLM stuff, but it'll be everywhere after that. https://en.wikipedia.org/wiki/Low-background_steel

Hey, you're internet famous now. Check out the end of this boingboing post.

https://boingboing.net/2023/10/05/ai-search-chatbots-output-...

Re: Generative AI could make search harder to trust

#235

Post by Kon-Peki »

tivert wrote: post #8

We did it guys! We're definitely heading into a new era, one perfected by software engineers. I can't wait!

THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.

Re: Generative AI could make search harder to trust

#236

Post by simonhughes22 »

tornato7 wrote: post #231
wildrhythms wrote: post #219

Earlier quoted context omitted.

I don't understand how OpenAI thinks they could stop that from happening.

Not saying they should, but if they wanted to they could have an API that allows you to check whether some text was generated by them or not. Then Google would be able to check search results and downrank.

It's not that simple. Originally OpenAI released a model to try and detect whether some content was generated by an LLM or not. They later dropped the service as it wasn't accurate. Today's models are so good at text generation it's not possible in most cases to differentiate between a human and machine generated text.

Re: Generative AI could make search harder to trust

#237

Post by naillo »

simonhughes22 wrote: post #236
tornato7 wrote: post #231

Earlier quoted context omitted.

Not saying they should, but if they wanted to they could have an API that allows you to check whether some text was generated by them or not. Then Google would be able to check search results and downrank.

It's not that simple. Originally OpenAI released a model to try and detect whether some content was generated by an LLM or not. They later dropped the service as it wasn't accurate. Today's models are so good at text generation it's not possible in most cases to differentiate between a human and machine generated text.

Well they could just not allow prompts that seem to participate in blogspam. If they wanted to stop it they definitely could.

Their argument is that since it's centralized, things like that are possible (while with llama2 you can't), they do "patch" things all the time. But since blobspam are contributing to paying back the billions microsoft expects they're not going to.

Re: Generative AI could make search harder to trust

#238

Post by darkerside »

TerrifiedMouse wrote: post #172
darkerside wrote: post #161

Earlier quoted context omitted.

How do you know that? A human that produced fully convincing bullshit would be definitionally undetectable to you

Because everyone has at least tried to lie - not necessarily for malicious reasons, e.g. white lies - a few times in their lives and know it’s not easy coming up with plausible sounding lies. It takes effort. Sure it might be effortless for some psychopaths but there are a limited number of such people. Overall the internet was surprisingly usable despite all the liars - even now with SEO and content farms; I think e…

Again...

> Because everyone has at least tried to lie - not necessarily for malicious reasons, e.g. white lies - a few times in their lives and know it’s not easy coming up with plausible sounding lies.

How do you know that?

Re: Generative AI could make search harder to trust

#239

Post by pseudosavant »

LetsGetTechnicl wrote: post #230
pseudosavant wrote: post #13

Earlier quoted context omitted.

Except, unlike crypto, ChatGPT helps me with real day things that I easily find at least $20/month of value from.

I don't necessarily see a problem with it for personal use by a tech-savvy person who is aware of its problems and limitations. Releasing it upon the general public where mis/disinformation is already widespread on the internet is a terrible, terrible idea.

Yes. Only the enlightened like me should get it…smh.

Re: Generative AI could make search harder to trust

#240

Post by LetsGetTechnicl »

pseudosavant wrote: post #239
LetsGetTechnicl wrote: post #230

Earlier quoted context omitted.

I don't necessarily see a problem with it for personal use by a tech-savvy person who is aware of its problems and limitations. Releasing it upon the general public where mis/disinformation is already widespread on the internet is a terrible, terrible idea.

Yes. Only the enlightened like me should get it…smh.

We don't hand dangerous tools to people without proper training. I don't see why this would be any different.
Post reply on HN