4 ms·

I'm not convinced this is the explanation. It's true that for some product-centric queries, you get mostly paid results. But for information-seeking queries, Go

I'm not convinced this is the explanation. It's true that for some product-centric queries, you get mostly paid results. But for information-seeking queries, Google tries to give you organic results. The problem is simply that there's too much money to be made by capturing these. For years, you had content farms and fake "review" sites stepping up their game. Now, LLMs essentially make it a losing proposition to try and surface the small web. The least-bad option for Google would be to send you to moderated communities, such as Reddit, Quora, Stack Exchange, Wikipedia, and so on. But not all queries can be handled that way. If you look at the article that these guys are complaining about... how do you distinguish it from content-farmed spam? You can't. Now, I think Google is throwing in the towel and just want an LLM to answer info queries instead. That has a ton of problems, but to the average user, probably feels more helpful. At least until the spammers start gaming that.

throwaway48476 2y ago
>If you look at the article that these guys are complaining about... how do you distinguish it from content-farmed spam? You can't You can, Google just hasn't bothered to even try. As an example it's pretty easy to detect affiliate marketing links.
- DrillShopper 2y ago
  Google: We've done nothing and we're all out of ideas, man!
- equestria 2y ago
  Not really? The vast majority of helpful websites monetizes the same way the spammers do. Plus, even without affiliate links, purchase attribution is already scary good. Did you know that credit card issuers and ad tech companies collaborate to attribute brick-and-mortar purchases to online ad views? You have untold billions of dollars at stake. The industry is not banging rocks together. You could make a search engine for non-commercial content only, and I would actually love to have that, but (a) it would be a woefully tiny sliver of the internet; and (b) Google is obviously not the right company to do it.
  - wruza 2y ago
    At its scale google could just hire a big group of diverse internet-aware professionals (two of each kind) who would do random searches and simply manually ban sites that restyle or spam content. It’s absolutely easy for a searcher to tell if it’s spam.
    - equestria 2y ago
      This is actually an interesting rabbit hole. Their other constraint is that they are trying to be transparent and fair, for some definition of that word. They need that for Section 230 protections, they need it to fend off antitrust lawsuits, etc. So, there is a huge problem with hiring experts and telling them to make subjective decisions. Instead, Google publishes guidelines for webmasters and then enforces them rigidly. This usually ends up penalizing some good sites, while spammers swiftly discover workarounds.
      - lazystar 2y ago
        then make it a gov position, like the "Beer and Malt Beverage Labeling and Formulation Approval" position, which consists of a single guy approving or rejecting the graphic designs of beer bottles/cans. there was an NPR segment a while back about the one guy that held that position for 40 years.
yowlingcat 2y ago
> The least-bad option for Google would be to send you to moderated communities, such as Reddit, Quora, Stack Exchange, Wikipedia, and so on. But not all queries can be handled that way. There's a wide gulf between "path of least resistance" and "least-bad". I don't think it's productive to conflate the first path with the second.
ryandv 2y ago
We are approaching a digital Kessler syndrome, or perhaps a Deepwater Horizon info-oilspill event, where there is so much useless SEO-driven slop (soon to be taken webscale with the advent of genAI) to sift through on the internet that it's growing increasingly difficult to find the signal amidst the noise. Google, once a prestigious company which prided themselves on "organizing the world's information" and "not being evil," eventually became a target for those wanting to peddle their wares and make a quick buck - a departure from the days of the early Internet which was mostly computer geeks, hobbyists, and forward thinkers sharing organic content they thought was interesting or useful. Because search was Google's entire business, they needed to develop countermeasures to combat spammers and pages gaming Google's algorithm with questionable SEO techniques in order to preserve the signal-to-noise ratio of the search engine results page. This is now a bygone era - after discarding their original motto of "don't be evil," search and "organizing the world's information" are no longer Google's business, it's hawking advertisements [0]: When Gomes pushed back on the multiple requests for growth, Fox added that all three of them were responsible for search, that search was “the revenue engine of the company,” and that bartering with the ads and finance teams was potentially “the new reality of their jobs.” On February 6th 2019, Gomes said that he believed that search was “getting too close to the money,” and ended his email by saying that he was “concerned that growth is all that Google was thinking about.” Hence questionable grey UX patterns like blurring the distinction between ads and organic content, and sometimes cramming the page so full of ads that all the actual results are "below the fold." Remember the old Internet adage - if you're not paying for the product, you are the product - and like cattle we are all just herded into digital pens to be served marketing slop to serve the real customer - the advertisers. If you want to be treated as a customer instead of cattle, you ought then to pay for your services, including search, to align the financial incentives with your own. Advertising needs to die, for it is a root cause of most of the ills of the modern internet [1]. If you can pay for streaming services or music, you can certainly pay for access to high quality organic information that actually aligns with your interests - not that of the advertisers. I've been using Kagi for a few years now and it really does hearken back to Google SERP quality maybe not at its peak, but rounding near to it. At the risk of sounding elitist (and so what), this is just another consequence of the recurring Eternal September phenomenon - highly focused communities with a strong concentration of geeks, hobbyists, and experts were the norm back then, when computers were still new and arcane devices that were difficult to operate. The bar to entry was much higher, and one had to do a little bit of "reading the fucking manual" simply to get online and understand how to navigate the net effectively. Now that all the balls have been poured into the Galton board we have regressed to the mediocrity of content that exists on the contemporary Web, absent those pressures that once selected for high quality content online. [0] https://www.wheresyoured.at/the-men-who-killed-google/ https://www.wheresyoured.at/the-men-who-killed-google/ [1] https://news.ycombinator.com/item?id=41940718 https://news.ycombinator.com/item?id=41940718
wruza 2y ago
Sometimes I lose track of a bookmark and google a phrase from it verbatim and in quotes. Guess what I get most of the times? Not even a verbatim search spam, just nonsense with almost no original words. It doesn’t even try and I doubt it even has a database where a direct path exists from a phrase to it with a link. Google is not “I found this for you” anymore, it is “I think you meant this”.
- ghxst 2y ago
  I thought this was just me! Even if I use double quotes to search for a 1:1 phrase I often get bad results and nothing linking back to the source I was looking for. If you happen to remember the site something was on, adding "site:example.com" still produces decent results fortunately.
  - ipaddr 2y ago
    You have to go into search options and select verbatim. But since they stopped indexing many sites it might not even be there anyways.
smcin 2y ago
> "But for information-seeking queries, Google tries to give you organic results." No, often it absolutely doesn't, and I posted two epic fails here previously. Worse still, Google organic results no longer understand(/distinguish) the difference between information-seeking queries vs product searches (or else, SEO people have been gaming it for years, and Google search rankings have made this worse): ____________________________________________________________ 1) Google search relevance fail: result for “Africa longitude” https://news.ycombinator.com/item?id=30337563 https://news.ycombinator.com/item?id=30337563 Googling for "Africa longitude" should return a range of longitudes, like: "17.5°W - 51.5°E" or "["17°31′13″W - 51°27′52″E"]" [1], or for the A+ answer: *"Africa lies between 17°33'22" W, (Cape Verde, westernmost point) and 51°27'52" E, (Ras Hafun, Somalia, easternmost projection). But it doesn't. Moreover these coordinates haven't substantially changed in 10,000 years (other than political/territorial disputes about islands, but the coords for the mainland certainyl haven't). Googling returns the grossly misleading "Africa/Coordinates 8.7832° S, 34.5085° E". When you dig into why this so, it seems to be "optimized" for the SEO activities of a digital map storefront, MapsofWorld.com, acquired by MapSherpa Inc., based in Ottawa. And those mystery nonsense coordinates ("8.7832° S, 34.5085° E") bizarrely point not even to the geographical centre of Africa but to a random rural location 2400km ESE away, in southern Tanzania, which appears to have been deceptively mislabeled, in Cyrillic, as a Russian store (by Russian SEO?). For a pin dropped in rural Tanzania. No QA! https://news.ycombinator.com/item?id=30337563 https://news.ycombinator.com/item?id=30337563 ____________________________________________________________ Similarly: 2) Google search for "How many landings on dark side of moon" is grossly incorrect https://news.ycombinator.com/item?id=40862146 https://news.ycombinator.com/item?id=40862146 (Update: the AI overview factbox has at least since been corrected to give Two: "Chang'e 4 (January 3, 2019) and Chang'e 6 (2024), instead of repeating what jagranjosh.com says). However the #1 organic search hit is still the woefully inaccurate unauthoritative page: https://www.jagranjosh.com/general-knowledge/list-of-all-successful-moon-missions-1568103956-1 https://www.jagranjosh.com/general-knowledge/list-of-all-suc... which cheerfully claims "Aug 23, 2023 — There are over 21[!!] moon missions that have been launched successfully on the dark side of the moon by 4 countries." This is hopelessly wrong, even if we utterly misunderstood the key word "landing" and also count any mission which merely photographed the dark side (Luna 3, 1959) or human overflight over it (Apollo 8, 1968). But not "landing". And why on earth did Google decide jagranjosh.com was more authoritative than any reference website or wiki? ____________________________________________________________ [1] https://en.wikipedia.org/wiki/Africa#Geology,_geography,_ecology,_and_environment https://en.wikipedia.org/wiki/Africa#Geology,_geography,_eco...

throwaway48476 2y ago

>If you look at the article that these guys are complaining about... how do you distinguish it from content-farmed spam? You can't You can, Google just hasn't bothered to even try. As an example it's pretty easy to detect affiliate marketing links.

DrillShopper 2y ago
Google: We've done nothing and we're all out of ideas, man!
equestria 2y ago
Not really? The vast majority of helpful websites monetizes the same way the spammers do. Plus, even without affiliate links, purchase attribution is already scary good. Did you know that credit card issuers and ad tech companies collaborate to attribute brick-and-mortar purchases to online ad views? You have untold billions of dollars at stake. The industry is not banging rocks together. You could make a search engine for non-commercial content only, and I would actually love to have that, but (a) it would be a woefully tiny sliver of the internet; and (b) Google is obviously not the right company to do it.
- wruza 2y ago
  At its scale google could just hire a big group of diverse internet-aware professionals (two of each kind) who would do random searches and simply manually ban sites that restyle or spam content. It’s absolutely easy for a searcher to tell if it’s spam.
  - equestria 2y ago
    This is actually an interesting rabbit hole. Their other constraint is that they are trying to be transparent and fair, for some definition of that word. They need that for Section 230 protections, they need it to fend off antitrust lawsuits, etc. So, there is a huge problem with hiring experts and telling them to make subjective decisions. Instead, Google publishes guidelines for webmasters and then enforces them rigidly. This usually ends up penalizing some good sites, while spammers swiftly discover workarounds.
    - lazystar 2y ago
      then make it a gov position, like the "Beer and Malt Beverage Labeling and Formulation Approval" position, which consists of a single guy approving or rejecting the graphic designs of beer bottles/cans. there was an NPR segment a while back about the one guy that held that position for 40 years.

yowlingcat 2y ago

> The least-bad option for Google would be to send you to moderated communities, such as Reddit, Quora, Stack Exchange, Wikipedia, and so on. But not all queries can be handled that way. There's a wide gulf between "path of least resistance" and "least-bad". I don't think it's productive to conflate the first path with the second.

ryandv 2y ago

We are approaching a digital Kessler syndrome, or perhaps a Deepwater Horizon info-oilspill event, where there is so much useless SEO-driven slop (soon to be taken webscale with the advent of genAI) to sift through on the internet that it's growing increasingly difficult to find the signal amidst the noise. Google, once a prestigious company which prided themselves on "organizing the world's information" and "not being evil," eventually became a target for those wanting to peddle their wares and make a quick buck - a departure from the days of the early Internet which was mostly computer geeks, hobbyists, and forward thinkers sharing organic content they thought was interesting or useful. Because search was Google's entire business, they needed to develop countermeasures to combat spammers and pages gaming Google's algorithm with questionable SEO techniques in order to preserve the signal-to-noise ratio of the search engine results page. This is now a bygone era - after discarding their original motto of "don't be evil," search and "organizing the world's information" are no longer Google's business, it's hawking advertisements [0]: When Gomes pushed back on the multiple requests for growth, Fox added that all three of them were responsible for search, that search was “the revenue engine of the company,” and that bartering with the ads and finance teams was potentially “the new reality of their jobs.” On February 6th 2019, Gomes said that he believed that search was “getting too close to the money,” and ended his email by saying that he was “concerned that growth is all that Google was thinking about.” Hence questionable grey UX patterns like blurring the distinction between ads and organic content, and sometimes cramming the page so full of ads that all the actual results are "below the fold." Remember the old Internet adage - if you're not paying for the product, you are the product - and like cattle we are all just herded into digital pens to be served marketing slop to serve the real customer - the advertisers. If you want to be treated as a customer instead of cattle, you ought then to pay for your services, including search, to align the financial incentives with your own. Advertising needs to die, for it is a root cause of most of the ills of the modern internet [1]. If you can pay for streaming services or music, you can certainly pay for access to high quality organic information that actually aligns with your interests - not that of the advertisers. I've been using Kagi for a few years now and it really does hearken back to Google SERP quality maybe not at its peak, but rounding near to it. At the risk of sounding elitist (and so what), this is just another consequence of the recurring Eternal September phenomenon - highly focused communities with a strong concentration of geeks, hobbyists, and experts were the norm back then, when computers were still new and arcane devices that were difficult to operate. The bar to entry was much higher, and one had to do a little bit of "reading the fucking manual" simply to get online and understand how to navigate the net effectively. Now that all the balls have been poured into the Galton board we have regressed to the mediocrity of content that exists on the contemporary Web, absent those pressures that once selected for high quality content online. [0] https://www.wheresyoured.at/the-men-who-killed-google/ https://www.wheresyoured.at/the-men-who-killed-google/ [1] https://news.ycombinator.com/item?id=41940718 https://news.ycombinator.com/item?id=41940718

wruza 2y ago

Sometimes I lose track of a bookmark and google a phrase from it verbatim and in quotes. Guess what I get most of the times? Not even a verbatim search spam, just nonsense with almost no original words. It doesn’t even try and I doubt it even has a database where a direct path exists from a phrase to it with a link. Google is not “I found this for you” anymore, it is “I think you meant this”.

ghxst 2y ago
I thought this was just me! Even if I use double quotes to search for a 1:1 phrase I often get bad results and nothing linking back to the source I was looking for. If you happen to remember the site something was on, adding "site:example.com" still produces decent results fortunately.
- ipaddr 2y ago
  You have to go into search options and select verbatim. But since they stopped indexing many sites it might not even be there anyways.

smcin 2y ago

> "But for information-seeking queries, Google tries to give you organic results." No, often it absolutely doesn't, and I posted two epic fails here previously. Worse still, Google organic results no longer understand(/distinguish) the difference between information-seeking queries vs product searches (or else, SEO people have been gaming it for years, and Google search rankings have made this worse): ____________________________________________________________ 1) Google search relevance fail: result for “Africa longitude” https://news.ycombinator.com/item?id=30337563 https://news.ycombinator.com/item?id=30337563 Googling for "Africa longitude" should return a range of longitudes, like: "17.5°W - 51.5°E" or "["17°31′13″W - 51°27′52″E"]" [1], or for the A+ answer: *"Africa lies between 17°33'22" W, (Cape Verde, westernmost point) and 51°27'52" E, (Ras Hafun, Somalia, easternmost projection). But it doesn't. Moreover these coordinates haven't substantially changed in 10,000 years (other than political/territorial disputes about islands, but the coords for the mainland certainyl haven't). Googling returns the grossly misleading "Africa/Coordinates 8.7832° S, 34.5085° E". When you dig into why this so, it seems to be "optimized" for the SEO activities of a digital map storefront, MapsofWorld.com, acquired by MapSherpa Inc., based in Ottawa. And those mystery nonsense coordinates ("8.7832° S, 34.5085° E") bizarrely point not even to the geographical centre of Africa but to a random rural location 2400km ESE away, in southern Tanzania, which appears to have been deceptively mislabeled, in Cyrillic, as a Russian store (by Russian SEO?). For a pin dropped in rural Tanzania. No QA! https://news.ycombinator.com/item?id=30337563 https://news.ycombinator.com/item?id=30337563 ____________________________________________________________ Similarly: 2) Google search for "How many landings on dark side of moon" is grossly incorrect https://news.ycombinator.com/item?id=40862146 https://news.ycombinator.com/item?id=40862146 (Update: the AI overview factbox has at least since been corrected to give Two: "Chang'e 4 (January 3, 2019) and Chang'e 6 (2024), instead of repeating what jagranjosh.com says). However the #1 organic search hit is still the woefully inaccurate unauthoritative page: https://www.jagranjosh.com/general-knowledge/list-of-all-successful-moon-missions-1568103956-1 https://www.jagranjosh.com/general-knowledge/list-of-all-suc... which cheerfully claims "Aug 23, 2023 — There are over 21[!!] moon missions that have been launched successfully on the dark side of the moon by 4 countries." This is hopelessly wrong, even if we utterly misunderstood the key word "landing" and also count any mission which merely photographed the dark side (Luna 3, 1959) or human overflight over it (Apollo 8, 1968). But not "landing". And why on earth did Google decide jagranjosh.com was more authoritative than any reference website or wiki? ____________________________________________________________ [1] https://en.wikipedia.org/wiki/Africa#Geology,_geography,_ecology,_and_environment https://en.wikipedia.org/wiki/Africa#Geology,_geography,_eco...