I need to step away from my PC – for a moment – because, although I have so much to write, the statements made in this video touch me too deeply and are too closely aligned to my own views and too close to the fundamental reasons underlying my own depression and disillusionment and burn-out.

Watch it.

Seriously. Watch it. If you are well briefed on the A.I. bubble and A.I. Hell, just skip to:

  • ~ 34 minutes to miss the demonstration of the tedious issue.[1]
  • ~ 38 minutes to reach the philosophical statements
  • ~ 39 minutes to hear about deception – the universal “tell” of A.I. scammers
  • ~ 41 minutes if you’re prepared for tears: to lament what we’ve lost, what we so nearly had, what humanity is losing, what is being stolen from artists ¬

(I need some space.)

  1. I assure you this video is not about content farms, SEO or the death of search but one might be mistaken for thinking that, in the first half. Don’t. It is worth your patience. ↩︎

  • kbal@fedia.io
    5·
    3 days ago

    Searching the web for “glb file format” seems a natural thing to do while listening to the first part of this video. On my favourite searx instance I had to scroll past only 7 links to LLM-generated garbage before finding a link to the actual spec. I wonder how Kagi fares in this test.

    • FaceDeer@fedia.io
      9·
      3 days ago

      If you wanted the specification why not search for “.glb format specification”? I did that on Google and the specification was the first hit.

      • kbal@fedia.io
        8·
        3 days ago

        I find it interesting as an experiment to measure how polluted the information environment is with machine-generated shit, not as an exercise in how to navigate around it.

        • FaceDeer@fedia.io
          4·
          3 days ago

          It’s only “polluted” if you’re looking for something specific and you refuse to ask for something specific.

          If you go into a restaurant and ask them for “a drink” without specifying what drink you want, don’t complain about the quality of the coffee when they bring you a coke.

          • kbal@fedia.io
            3·
            3 days ago

            Coffee and cola are both pretty good. Are you reflexively coming to the defense of generative AI on general principle or did you actually look at those sites, the horribleness of which is exhaustively detailed in the video, and decide that they look as if they could be useful to anyone?

            • FaceDeer@fedia.io
              3·
              3 days ago

              I looked at the sites. Did you? The thing that OP was looking for that they claimed had been made unfindable or “polluted” were perfectly accessible and fine.

              • DarkNightoftheSoul@mander.xyzEnglish
                3·
                3 days ago

                “google-fu” is a skill. it’s not reasonable to expect everyone to be perfectly proficient. and it seems like you are saying “just perfectly define your search query to match the exact result you need every time” which, i think most would agree is often very difficult even if your fu is quite good. I think this might actually be an np problem. I used to have good fu, these days I find that google sometimes breaks or ignores my quotes, changes exact wordings to thesaurus matches, shit like that. also, there are filter bubbles.

                now we’re adding to all of those already known uncontroversial challenges to googling the information “pollution” of low-quality unsupervised chatbot hallucinations where genuine good search results might otherwise go. it’s not that it’s unfindable- that’s not what the op video claims. it’s that the signal-to-noise ratio is intolerably low, and that basically by design.

                and yes, it seems there is still room for human-creative solutions to these search problems, like the one you suggest which basically rearranges words and exchanges some thesaurus hits. sort of like how you can sometimes “jailbreak” llms by TaLkInG tO tHeM lIkE ThIs. for now. actually, they probably already patched that.

                • FaceDeer@fedia.io
                  3·
                  3 days ago

                  The “google-fu” in this case was to search for “.glb format specification” when seeking the .glb format specification.

                  This really doesn’t seem like a huge challenge requiring sophisticated skills.

                  • kbal@fedia.io
                    2·
                    2 days ago

                    The thing that’s wanted is not one link to the official protocol specification document along with a dozen links to SEO-optimized AI-generated time-wasting nonsense. It’s a large set of links to diverse interesting sources discussing the topic searched for and things adjacent to it. The web was never perfect, but we were much closer to that ideal ten years ago than we are today.

                    Someone’s complaining about the war and you’ve come along with a pair of earplugs saying wear these, you won’t even hear the bombs.

                  • DarkNightoftheSoul@mander.xyzEnglish
                    1·
                    2 days ago

                    well yes, it seems there is still room for human-creative solutions to these search problems, like the one you suggest which basically rearranges words and exchanges some thesaurus hits. it seems like you are saying “just perfectly define your search query to match the exact result you need every time” which, i think most would agree is often very difficult even if your fu is quite good. I think this might actually be an np problem. i also think most people would reasonably expect “.glb file format” to bring up information about the glb file format, including specs, at or very near the top of results. in evidence of most people’s expectations here, i submit the entire rest of the thread where your experience is an outlier.

    • Jo Miran@lemmy.ml
      3·
      3 days ago

      I just searched using Kagi. It fares poorly. Normally I get better results but “glb file format” seems to be a cursed search. Nothing but AI with the exception of one Wikipedia result and one Reddit post.

      I will try “how to parse data from a glb file” and report again in an edit.

      EDIT: Solid hits from the first try. It seems that search engines are being swarmed by generated garbage for open or vague questions. The moment you get specific, the AI slop seems to fall off…at least on Kagi.