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

Google search's death by a thousand cuts

matt-rickard.com

Post reply on HN
Print view View original

101–110 of 332 posts

Re: Google search's death by a thousand cuts

#101

Post by zpeti »

What is the motivation of anyone to make a website at this point? I don't see it.

I've not seen very much discussion about the issues these LLMs are going to cause, this article is sort of talking about it - twitter and wikipedia could restrict their content, but - what about the millions of webmasters who ultimately created what the web is today?

What is the motivation of anyone to really create content any more, apart from for actual social discussions on social media? LLMs will steal the content, and no one will read it.

I really don't know what this will bring in the future. Will online content fizzle out in 2-3-5 years and fresh content will just disappear? Will LLM only be good up to 2025 data?

This is a massive problem that changes the web as we know it, yet I don't see any discussion of this. Mark Andreeson bought it up for a sentence on lex Friedman, but that was it. He didn't know the answer.

Re: Google search's death by a thousand cuts

#102

Post by Ueland »

foobarbecue wrote: post #53
moultano wrote: post #20

Earlier quoted context omitted.

Can you give an example query and the page it should return?

Sure. Searching M1104273380R should return https://wms.lroc.asu.edu/lroc/view_lroc/LRO-L-LROC-2-EDR-V1.... and this paper: https://doi.org/10.1038/nature19829 . Interestingly, when I tried this just now, I did get the paper as a result. I could swear I tried this same search yesterday and had zero results.

In a ironic twist, it now returns this HN page.

Re: Google search's death by a thousand cuts

#103

Post by pavs »

stevenicr wrote: post #56
nikanj wrote: post #27

Earlier quoted context omitted.

I always have a feeling they could get surprisingly far by having one person doing random searches, then tagging all the obvious spam to be deindexed. But Google is very adamant about never hiring a person to do a job well, if AI can be trained to do the same job badly

Don't they have whole teams of people looking at sites through the lens of the leaked 'quality rater guidelines' - ? I thought it was like a few thousand people in the philippines downranking sites for thousands of reasons - without being transparent about possible biases and such. Maybe I'm thinking of somewhere else. And wasn't there a story about a bunch people who were doing similar as 1099 contractors in the US…

Looking at the quality of current google searches I highly doubt it. I mean I could single handedly remove thousands of domains in a month and improve google search result, I have a hard time believing that they have 1000s of people doing it and the quality is this shitty.

Re: Google search's death by a thousand cuts

#104

Post by jkukul »

wodenokoto wrote: post #73
foobarbecue wrote: post #10

Doing planetary science, I used to be in the habit of searching for image IDs on google. Usually particular orbital images of the moon which had been mentioned in a paper. There are of course specific archives to search for this stuff, but google used to give surprisingly good results. Now I get literally zero results when I search an image ID. What changed? Did they remove their index of the Planetary Data System? D…

I don’t think they removed the results from index, but I believe they started searching for vectors instead of text matches. Your ids are nowhere in vector space.

Shouldn't it be possible to work around this by wrapping the query in double quotes? Google promises to do an exact match search when wrapping in quotes.

So I'd imagine the OP's query should work when the image id is wrapped in double quotes. Otherwise, it'd mean that it's not in the index anymore.

Re: Google search's death by a thousand cuts

#105

Post by pgmonstereater »

superasn wrote: post #14

The funny thing is the site which should have been removed like the of Stackoverflow spam clones, or sites like Canva and Pinterest that make thousands of similar looking pages with slightly different heading are still allowed and rank on Google. Also hate the top 10 pages whenever you search the best something, like best domain name registrar. I don't want to read a spam blog post with affiliate links, I wish google…

What are stack overflow spam clones?

Re: Google search's death by a thousand cuts

#106

Post by danpalmer »

stevage wrote: post #99

I would really love to read a deep dive on why Google Search quality has declined so much. There's a lot of anecdotes but I haven't read much that sounds particularly authoritative.

On why, or on whether. I personally find Google Search to be pretty great, and recently switched back from DuckDuckGo which I had been using for ~4 years, because I found I was having to use !g in half my searches. It's also noticeable faster. When I'm looking for an answer, I find Google much more likely to give me the answer. So I'm not convinced that search quality is declining. I'd love to see a deep dive on whether it actually is declining.

Disclaimer: I work at Google, but not on anything to do with search, and used DDG for my first year here out of habit. I have no inside info on search.

Re: Google search's death by a thousand cuts

#107

Post by adhesive_wombat »

danpalmer wrote: post #100

> What if Wikipedia started charging or restricting API access? Wikipedia has a downloadable data dump that would cost almost nothing to serve to Google, and it has an organisational mandate to make that data available. If they decide to charge for access to that I'm sure Google can afford it. Let's not throw around completely unrealistic hypotheticals.

Selling "consumable" data access is called Wikimedia Enterprise: https://en.wikipedia.org/wiki/Wikimedia_Enterprise

Re: Google search's death by a thousand cuts

#108

Post by hnbad »

pgmonstereater wrote: post #105
superasn wrote: post #14

The funny thing is the site which should have been removed like the of Stackoverflow spam clones, or sites like Canva and Pinterest that make thousands of similar looking pages with slightly different heading are still allowed and rank on Google. Also hate the top 10 pages whenever you search the best something, like best domain name registrar. I don't want to read a spam blog post with affiliate links, I wish google…

What are stack overflow spam clones?

There are various sites that seem to crawl Stack Overflow and present the content verbatim in their own UI. They're annoying because they are sometimes out of date, they generate duplicate results and they're practically uninhabited.

Re: Google search's death by a thousand cuts

#109

Post by d--b »

What changed?

Just an asumption here. But could it be the fault of the web rather than Google's?

Could it be that the web itself is being so flooded with SEO-ed crap content that even Google can't sort through it all?

Re: Google search's death by a thousand cuts

#110

Post by supriyo-biswas »

cma wrote: post #12

In the TPUv4 paper it sounded like Google search moved over to some really crude AI embedding vector stuff. If you search "book of master system" it gives only results for "book of genesis" because it mixes up the biblical book of Genesis with the Sega Master System. Searching any programming terms, like camel-case variable names from a system library without exact quotes, just gives random results about Katy Perry a…

I looked for “book of master system” and it gave me results about the SEGA book.

Granted, vectorization could be one of the issues plaguing search, but it isn’t the primary issue.

Post reply on HN