• Baku@aussie.zone
    link
    fedilink
    English
    arrow-up
    83
    ·
    8 months ago

    A couple of years ago I found a black and white photo from the late 1800s and wanted to figure out what station it was from. Google was useless and only showed unrelated stations, but surprisingly, Bing found a page with the exact photo on it. It was on one of those shitty scraper pages that just lists thousands and thousands of random photos, but nonetheless I figured out what station it was

    • lars@lemmy.sdf.org
      link
      fedilink
      arrow-up
      28
      ·
      8 months ago

      Jesus Christ I guess I’m not misremembering.

      Bing’s reverse image search is essentially dead in 2024 unless you’re uploading the Mona Lisa. It’s really, really terrible and even worse than Google.

      My favorites right now are Tineye, Yandex, and Google, in that order.

      • net00@lemm.ee
        link
        fedilink
        English
        arrow-up
        22
        ·
        edit-2
        8 months ago

        Strange, for me Tineye has not a single time been able to identify ANY of the images I ever tried. Yandex has worked best for me

        • Microw@lemm.ee
          link
          fedilink
          arrow-up
          3
          ·
          8 months ago

          Same, tineye only ever worked for me if I uploaded a picture that was by Reuters or something and therefore on lots of reputable sites. In any other cases it found nothing.

      • Baku@aussie.zone
        link
        fedilink
        English
        arrow-up
        11
        ·
        edit-2
        8 months ago

        My last experience with bings reverse image search was in 2022 or so, so no vouches for its quality these days. I’ve had mixed results with tineye, but there was another one which I don’t even remember the name of that generated reverse search links for all the search engines, I think it even listed that Chinese one and a few others I’ve never heard of rather than being its own thing. I had decent luck with that, I found Bing still worked the best but I haven’t tried it since

        Google lens definitely wins for object search though. Not the point of the post, I know, but it’s kind of funny how their reverse image search is dogshit but their object recognition is flawless

          • BlueFootedPetey@sh.itjust.works
            link
            fedilink
            arrow-up
            2
            ·
            8 months ago

            Tineye? Yea it’s used in Branden Sanderson’s Mistbourne books. People have the ability to ingest different metals for different abilities.

            People who ingest tin, gain heightened senses. Vision, hearing, touch, etc. They are known as “tineyes”.

    • mindbleach@sh.itjust.works
      link
      fedilink
      arrow-up
      1
      ·
      8 months ago

      Neither ever worked for lineart. Photos? Used to be reliable, occasionally bordered psychic, now just dumb. Drawings? Yep, that’s a drawing. Did you need anything else?

  • beeng@discuss.tchncs.de
    link
    fedilink
    arrow-up
    79
    ·
    8 months ago

    It does seem like the ideology of those inside google went from “tech” , to “I know better than you do”. Not sure it’s fixable really…

    • ArtificialLink@lemy.lol
      link
      fedilink
      arrow-up
      18
      ·
      8 months ago

      That’s the problem with most tech these days. They assume they know the best way to do something or know better than you. Its infuriating

      • NoSpiritAnimal@lemmy.world
        link
        fedilink
        arrow-up
        14
        arrow-down
        2
        ·
        8 months ago

        Spotify is a prime example of this. There are so many “features” I hate and that no one has asked for, yet shuffle doesn’t even work.

        Everytime I start spotify in ny office after listening on my commute, it tries to start playing on my phone since that was playing in my car.

        Or when I was still there, Reddit search. Absolutely useless and so fucking smarmy with that stupid doge.

      • Sanctus@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        8 months ago

        The problem with tech is managers are wearing engineer coats and calling the shots with no true credentials.

      • lorty@lemmy.ml
        link
        fedilink
        arrow-up
        1
        ·
        8 months ago

        Yeah, tech people nowadays have this attitude with most people, they only show some restraint when they think it’s for other people like them.

      • joonazan@discuss.tchncs.de
        link
        fedilink
        arrow-up
        1
        arrow-down
        1
        ·
        5 months ago

        It’s about minimizing the annoyance for the majority of users who will misspell some popular thing.

        Also, I believe that showing actually interesting content is bad for the businesses because it might make the user stop to think and pursue something meaningful instead of continuing to use the product.

  • Klear@lemmy.world
    link
    fedilink
    arrow-up
    42
    ·
    8 months ago

    Unrealistic. I usually have to scroll way down in the results to find a link to wikipedia nowadays.

    • pachrist@lemmy.world
      link
      fedilink
      arrow-up
      10
      ·
      8 months ago

      So true. If I want to know how old a celebrity is, first result is something about their latest work that doesn’t mention age, and then the next 3-4 are usually some ranking articles, “top 10 ceberities you didn’t know were 50,” and then Wikipedia comes in with the answer.

    • Hellfire103@lemmy.ca
      link
      fedilink
      English
      arrow-up
      72
      arrow-down
      6
      ·
      8 months ago

      I don’t even care if the results are good. I’m not about to use any part of RUnet.

      • jeremyparker@programming.dev
        link
        fedilink
        arrow-up
        4
        arrow-down
        1
        ·
        8 months ago

        Honest question: why not? Facebook/Google/Microsoft are up to some disgusting shit, are their Russian counterparts significantly different?

        • Hellfire103@lemmy.ca
          link
          fedilink
          English
          arrow-up
          11
          arrow-down
          1
          ·
          8 months ago

          Privacy, for one thing. I don’t use Google, Bing, Windows, or any Meta software*, and Yandex aren’t much different.

          Security, though, is another thing. I live in a NATO country, and I would imagine the Russian government are monitoring Yandex (and other RUnet services). Frankly, I think contributing any data to such a government would be against my interests.

          There’s also a lot of censorship on RUnet. Yeah, Google has that too, but Mojeek and Brave Search do not.

          TL;DR: Google is data-hungry and supplies data to the NSA; Yandex is data-hungry and supplies data to the Kremlin; Mojeek and Brave Search are good; DDG and Startpage are the best for the average user.

        • CCF_100@sh.itjust.works
          link
          fedilink
          English
          arrow-up
          40
          arrow-down
          3
          ·
          8 months ago

          I’m going to make the assumption that a lot of people on Lemmy are FOSS enthusiasts and are therefore adverse to anything closed source, especially a Russian web service…

        • jeremyparker@programming.dev
          link
          fedilink
          arrow-up
          4
          arrow-down
          1
          ·
          8 months ago

          Apparently the brigade has found you, but i want you to know that i agree (mostly). Obviously it kind of sucks tohavve Russian as the default language on everything you get from there, and there’s some super-obscure music I’ve failed to find on there, but it’s basically my first stop these days, whether it’s Abbot Elementary or CompTIA training videos.

    • Lvxferre@mander.xyz
      link
      fedilink
      arrow-up
      21
      arrow-down
      2
      ·
      edit-2
      8 months ago

      Yandex reverse image search often works considerably better than Google’s.

      And their translator too. I use it all the time for Latin. Google Translate was made mostly for Romance and Germanic languages, so it sucks at assigning the right case to Latin, and the word order is often a mess. Yandex was however made with Russian in mind, that is syntactically EDIT: grammatically* closer to Latin in those two aspects.

      *case division is morphology in this case.

          • suction@lemmy.world
            link
            fedilink
            arrow-up
            1
            ·
            edit-2
            8 months ago

            I had 6 years of Latin in school and remember less than 1% of it. Talk about a waste of time. Bada bing bada boom shoulda learned Italian-English rather.

            • Lvxferre@mander.xyz
              link
              fedilink
              arrow-up
              1
              ·
              8 months ago

              Sure. I’m not saying that Latin is useful for most people. It isn’t; I don’t expect most people to fuck around with Plautus, Cicero, Catullus etc. I’m saying that Yandex Translate is useful for me because of decent support for Latin, while Google offers better support for a handful of Latin descendants aka Romance languages (like Italian).

              And it’s mostly due to a coincidence - because the platform was made by Russian speakers and Russian happens to still keep a similar case system as Latin does.

              (…anche parlo italiano, ma davvero per italiano non uso nessuno - se non so qualcosa uso dizionari. Dà meno lavoro.)

      • lars@lemmy.sdf.org
        link
        fedilink
        arrow-up
        1
        ·
        8 months ago

        Sounds a little suss. All the propaganda I see about Latin mentions its free word order.

        • Lvxferre@mander.xyz
          link
          fedilink
          arrow-up
          5
          ·
          edit-2
          8 months ago

          Word order in Latin is only syntactically free. As in, if you change the word order, you aren’t changing who did what. However, you’re still changing the topic (whatever we were talking about) and comment (the new info that I’m adding in).

          I’ll give you an example:

          • Puer puellam amat - boy loves girl; but more like “the boy loves a girl”.
          • Puellam puer amat - boy loves girl; but more like “a boy loves the girl”.
          • Amat puer puellam - boy loves girl; but more like “speaking on love, the boy loves a girl” (hard to convey in English).

          Note how I used articles to convey roughly the same meaning in English. That’s because what Latin is doing with the word order is not too unlike what English does with articles. Sure, you can use “the boy”, “a boy”, or simply “boy”, it won’t change the basic meaning, but it’s still not “random”.

          And guess which language happens to use a similar system? Russian. The only major difference is that by default (i.e. you aren’t focusing on any element), Latin would put the verb at the end and Russian in the middle; but it’s the same variability.