A purported leak of 2,500 pages of internal documentation from Google sheds light on how Search, the most powerful arbiter of the internet, operates.

The leaked documents touch on topics like what kind of data Google collects and uses, which sites Google elevates for sensitive topics like elections, how Google handles small websites, and more. Some information in the documents appears to be in conflict with public statements by Google representatives, according to Fishkin and King.

  • w2tpmf@lemmy.world
    link
    fedilink
    English
    arrow-up
    39
    ·
    5 months ago

    Who wants to take bets that Search itself ends up in The Graveyard soon, leaving nothing but the new AI abomination in place?

    • kshade@lemmy.world
      link
      fedilink
      English
      arrow-up
      11
      ·
      edit-2
      5 months ago

      I could see them not letting you directly search anymore, only through the LLM bot. Because that’s been how things have been going anyway, Google seems to fully ignore literal searches with quote marks now, presumably because it doesn’t fit their vision of using natural (imprecise) language. So why not make the LLM write the search query for you in a completely opaque way?

    • flop_leash_973@lemmy.world
      link
      fedilink
      English
      arrow-up
      9
      ·
      5 months ago

      More likely they will just slowly rebrand search to more AI type things. Then slowly retire the non-AI parts in the background.

      • MrMcGasion@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        5 months ago

        Yeah, I know a lot of the smaller, independent search engines are lacking, but the people using the “udm=14” trick to remove Google’s AI results now, as if that won’t be removed as soon as Google needs to show investors the AI is more profitable.

        • adarza@lemmy.ca
          link
          fedilink
          English
          arrow-up
          1
          ·
          5 months ago

          the url needs a param to tell the server what kind of query is being requested. as long as they have the ‘web’ tab and option, it will be there. but i’m guessing they will come up with a way to encode that instruction in the tracking bits or something so you can’t just manually tack something on to the end of your query url and bypass their precious a.i. bot