Most active commenters

    ←back to thread

    212 points pseudolus | 37 comments | | HN request time: 0.642s | source | bottom
    1. NelsonMinar ◴[] No.42199494[source]
    I'm confused about how or why this is a new policy. My memory is inside Google we were discussing this risk back in 2003, probably earlier. Search quality was on it. I just assumed they'd lost the arms race, or that the parasites' ranking was justified for other reasons that were hard to tease apart. What are they doing new now?

    I think often about Mahalo, the sleazy shovel content that was spamming the web back in 2007. Google shut that down somewhat fast, although it did take several years. These days with AI and more aggressive spammers it's a losing battle. The real problem is the financial incentives that make this kind of spamming profitable in the first place.

    My tiny little blog gets about 3 requests a week for someone to "pay me to run a guest article". Going rate is $50-$200 and again, my blog is tiny.

    replies(9): >>42199551 #>>42199854 #>>42200207 #>>42200304 #>>42200373 #>>42200611 #>>42200832 #>>42200911 #>>42201266 #
    2. mjr00 ◴[] No.42199551[source]
    > I'm confused about how or why this is a new policy.

    My best guess is it's because they finally have a real competitor in ChatGPT.

    > The real problem is the financial incentives that make this kind of spamming profitable in the first place.

    Yeah, but the financial incentives exist on both ends. There's a gross symbiotic relationship between Google and SEO spammers, because Google also owns the ad network the spammers put on their page. If Google puts ad-laden SEO blogspam as the top result and a user clicks it, the user sees a bunch of ads from Google. Everyone wins: Google, the SEO spammers, and advertisers. Well, everyone except the user, but who cares about them?

    My guess/hope is that ChatGPT has made someone who actually cares about the quality of search results actually step in and say things have gone too far.

    replies(5): >>42199704 #>>42199751 #>>42199902 #>>42200157 #>>42200213 #
    3. chipsrafferty ◴[] No.42199704[source]
    Because ChatGPT is dependent on good search when it searches the web? Or because it completes with Google when it provides a good answer without searching? Or what do you mean specifically?
    replies(1): >>42199754 #
    4. NelsonMinar ◴[] No.42199751[source]
    You're totally right about that symbiotic relationship. We were aware of that risk in the early days when AdSense launched, we saw some very innovative and gross exploitation and created some policies to rein it in. But ultimately if Google makes a buck coming and going, they will do that.

    Wasn't there a big story last year in the wake of the DOJ antitrust investigation about Google manipulating search quality to boost ad revenue? I can't put my hands on a reference now, in part because Google is so bad at search these days I can't find anything more than a few months old.

    replies(1): >>42202026 #
    5. ryandvm ◴[] No.42199754{3}[source]
    I would say the latter. For software dev questions, my Google searches and Stack Overflow visits have fallen off a cliff since I started paying for ChatGPT.

    Ironically, I probably would have paid the same amount to Google for ad-free, old-style (accurate) Google searches, but no, they just wanted to keep cranking that ad dial up every year so that ship has sailed.

    At this point, I'm enjoying watching the old guard of search scrambling to find a life jacket.

    replies(1): >>42201102 #
    6. nurumaik ◴[] No.42199854[source]
    Just manually review top K websites and ban such garbage?

    Sometimes dumb, bruteforce and biased solution can work way better than any automation you can come up with

    replies(1): >>42201694 #
    7. nurumaik ◴[] No.42199902[source]
    > Google wins

    Define "wins". From what happening right now, it seems that google may lose much more than it earned by aligning with seo spammers

    Maybe they need to start locking employee stock options for 100 years to prevent them optimizing short-term gains?

    replies(1): >>42199950 #
    8. johnnyanmac ◴[] No.42199950{3}[source]
    Google's only taking the greedy approach. Spam sites on top, spam sites use google adsense, people click spam sites, they click google ads.

    It works great, until it doesn't. But that's a problem for the next CEO.

    replies(1): >>42200753 #
    9. klabb3 ◴[] No.42200157[source]
    > My best guess is it's because they finally have a real competitor in ChatGPT.

    Bingo. I always chuckle when people here say Google has lost it, and become incompetent. Well, they all make the mistake of assuming that they’re trying but failing, rather than that it’s deliberate simply due to boring economics.

    Now look at how quickly decades-long problems, so big they have an entire cottage industry built around it, suddenly be cleaned up. Incompetence? Nah.

    Of course, this does nothing to convince regulators and not even average HN user that innovation is harmed by these dominant players. Someone’s gotta think of the poor mega-corps.

    10. lern_too_spel ◴[] No.42200207[source]
    Because it became an embarrassing news story (https://larslofgren.com/forbes-marketplace/, also mentioned in this article). They would have lazily left it unfixed if everybody weren't laughing at them.
    replies(1): >>42201239 #
    11. resoluteteeth ◴[] No.42200213[source]
    > My best guess is it's because they finally have a real competitor in ChatGPT.

    My guess is it's because a bunch of articles about it have been posted to hn recently.

    12. dawnerd ◴[] No.42200304[source]
    Seriously, they tackled this years ago with the panda update to kill off all the how to and similar seo spam. It's like after around that time they just stopped caring at all and let the best X sites take over.
    13. lumost ◴[] No.42200373[source]
    Searching for python documentation was the worst, geeks for geeks and others would get the top slot for reskinning the pypi docs with ads.

    The entire thing was so blatant and obvious that I assumed Google did not care due to ad revenue.

    When ChatGPT launched search, you could immediately skip over all the crap. It made search nice again.

    14. stackghost ◴[] No.42200611[source]
    >My memory is inside Google we were discussing this risk back in 2003, probably earlier.

    Yeah but that was before they hired the incompetent grifter Prabhakar Raghavan and eventually made him head of Search.

    replies(2): >>42201132 #>>42202438 #
    15. Scoundreller ◴[] No.42200753{4}[source]
    iunno, I used to rank pretty well for things in my country like "$company's tech support number". Unlike every $company, my page had a nice clean URL like whatever.tld/Tech-Support-Number-for-$company and I'd just list of their phone numbers with a few paragraphs about how $company is shit. Maybe 50kb total.

    Meanwhile $company's page was company.tld/234897234982-029823749823742-2340823492 and 3 pages down was a phone number if your browser didn't choke on the javascript.

    For ISP ones, I recommended people print a copy so they can call if they can't get on the internet, which kinda backfired when a major ISP changed their tech support number (!) and it redirected to a toll-free squatter's sex chat line.

    Turns out $company really hates it when you call their call (cost) centres.

    I had maybe 50 pages for our different oligopolies and averaged $500/month revenue on adsense, so GOOG's cut was $250/month.

    Today, for one $company, the first 9 results are different pages from $company.tld, each unhelpful with a phone number in their own way, and they don't run adsense!

    16. xivzgrev ◴[] No.42200832[source]
    I miss Google of 2003

    What would it take for someone to make it today? No AI, only 1 on mobile, and sites with heavy ad loads are punished

    replies(5): >>42201143 #>>42201222 #>>42201281 #>>42202333 #>>42202571 #
    17. masto ◴[] No.42200911[source]
    The Google you remember from 2003 was murdered and its corpse is being worn by an entirely unrecognizable company.
    replies(1): >>42201408 #
    18. thephyber ◴[] No.42201102{4}[source]
    Stackoverflow visits fell[1] off a cliff since GPT became popular.

    Google is getting destroyed by the chatbot workflow because it is no longer the start of a browser session and clickthrus (the things that earn the high sponsored link rates) are falling as more users get their queries answered faster with less effort.

    [1] https://x.com/altimor/status/1853893158368928124?s=46

    replies(2): >>42202064 #>>42202075 #
    19. thephyber ◴[] No.42201132[source]
    Seems like post hoc fallacy.

    But people were complaining about the sAme issues under Matt Cutts. Also, there has been A Ton more money and work chasing the SEO farm game. Now big private equity companies have focused on buying a stable of big brands to do the same that used to be garage startups.

    20. theendisney ◴[] No.42201143[source]
    Just traffic.
    21. Spivak ◴[] No.42201222[source]
    You're mostly describing Kagi. They do have AI results but you have to explicitly ask for them. They have an "No AI" image search option as well.

    I also like my "Before AI" lens I can click on to search the internet pre-2021. And you can downrank or fully block those garbage spam sites. They even have a "leaderboard" for most blocked/pinned sites you can use to get started.

    22. walterbell ◴[] No.42201239[source]
    If so, the author deserves a Web Signal Award for succeeding where customers, regulators and execs failed despite years of trying.
    23. Electricniko ◴[] No.42201266[source]
    The air purifier review site Housefresh dug into why sites like theirs were seeing less traffic back in the spring, and it amounts to a handful of companies buying up popular magazine/blog brands and using them as affiliate farms that cross-post to sites within their networks of brands to boost visibility:

    https://housefresh.com/how-google-decimated-housefresh/

    On HN here:

    https://news.ycombinator.com/item?id=40239811

    And their previous article mentioned in that post generated a lot of discussion on HN:

    https://news.ycombinator.com/item?id=39433451

    24. vikingerik ◴[] No.42201281[source]
    It would take a benefactor who wants to pay for running it for its own sake and not for profit. As soon as there's a profit motive, enshittification sets in since you're serving whoever pays rather than your users.
    replies(1): >>42201466 #
    25. giancarlostoro ◴[] No.42201408[source]
    “Do all evil” its new unofficial slogan.
    26. nielsbot ◴[] No.42201466{3}[source]
    or maybe a government utility
    replies(1): >>42201816 #
    27. dotancohen ◴[] No.42201694[source]
    Yes, but that doesn't scale.
    replies(1): >>42201758 #
    28. danielheath ◴[] No.42201758{3}[source]
    Scale to what? You could manually look at the top million searches each year with four average college graduates.
    replies(1): >>42201867 #
    29. openrisk ◴[] No.42201816{4}[source]
    Governments have been in recent decades completely hands off from anything tech related. For the longest time they followed the usual neoliberal economist trope that markets solve everything optimally. As a result they have created in the so-called "big tech" conglomerates a monster fit for the history books, the most broken and cornered market there ever was.

    But a wholly government run search engine is not a solution. There are inherent biases in both constructing and presenting indices. You dont want to further stoke the anti-commons mistrust of polarized societies.

    What the public sector could do is fund all the background techonologies to make it easy to have much larger numbers of search engines. Some of those assembled services could be completely open source, others could be value adding with various added services and customizations.

    In any case the status quo is a disaster that has no future. Its effectively a forced dumbification of society as it artificially suppresses the flow of high-quality public information. Incidentally it also doesnt solve the problem that much of the world's information is private. Desktop search should become a thing again, in line with local AI etc.

    30. carstenhag ◴[] No.42201867{4}[source]
    In all languages of the world? Is that realistic? How could you know whether something in Vietnamese is spam or not?

    In my opinion it could work, but you'd definitely need 40-80 and not 4 people.

    replies(1): >>42201896 #
    31. adgjlsfhk1 ◴[] No.42201896{5}[source]
    Google has ~180k employees. that seems like a pretty small price to pay to make search work...
    32. Eisenstein ◴[] No.42202026{3}[source]
    > Wasn't there a big story last year in the wake of the DOJ antitrust investigation about Google manipulating search quality to boost ad revenue?

    This is the email chain you are looking for:

    * https://www.justice.gov/d9/2023-11/417581.pdf

    33. rockskon ◴[] No.42202064{5}[source]
    Are they, though? Inaccurate info is pretty common from LLMs.
    34. throwaway2037 ◴[] No.42202075{5}[source]
    StackOverflow has been dying a slow death since longer than before ChatGPT. Sure, ChatGPT is helping to accelerate it. The real data (leave aside social/community for a moment) issue with SO.com: Many answers don't age well. So, you have an answer from 8 years ago with 65 upvotes, but now the lang/lib was updated in 2023. A newer, more relevant answer is waaaaaaay down and only has one upvote. Personal note: I still pine for the old days when Joel Spolsky and Jeff Atwood were at the helm. They really knew how to build and sustain a vibrant community.
    35. oneshtein ◴[] No.42202333[source]
    Install a spam filter for search engines, like uBlackList.

    Use bunch of different search engines. In Firefox, enable search entry, then visit search engines and click green + in the entry, to add search engine.

    36. robk ◴[] No.42202438[source]
    I think it's nick fox now and he's old school and as competent as they come
    37. Shorel ◴[] No.42202571[source]
    I think it would be easy to make, with two decades of hardware improvements.

    The problem is that the web of 2003 doesn't exist any longer.

    Google existence changed the websites for better or worse. The Google of 2003 is no longer capable of dealing with today's web SEO dirty tricks.