Every app has to have fucking AI now for some reason.

  • guycls@lemmy.worldOP
    link
    fedilink
    arrow-up
    6
    arrow-down
    1
    ·
    6 months ago

    Thanks for the suggestion. What differences have you noticed compared to Google. Do you always find what you’re looking for. Have you tried searching niche stuff like some really old painting or an obscure compiler error etc.

    Weird that it gave me a link to franchise a taco joint when I searched for tacos near me.

    You want tacos? Open a taco shop.

    • Tlaloc_Temporal@lemmy.ca
      link
      fedilink
      arrow-up
      6
      ·
      6 months ago

      I wouldn’t consider “tacos near me” a representative search; google specifically optimizes searching for products and especially local food.

      I just searched for “first speedrun” and the first few results are decent but wrong, and the videos, shorts, and related searches after the first 2 entries are complete garbage.

      Being served 70% links to products sucks when searching anything related to a product isn’t fun either.

      • guycls@lemmy.worldOP
        link
        fedilink
        arrow-up
        2
        ·
        6 months ago

        Fair point. I don’t think it’s ready to become a default search engine as of now.

        • Tlaloc_Temporal@lemmy.ca
          link
          fedilink
          arrow-up
          2
          ·
          6 months ago

          Oh, whoops, I wasn’t comparing Stract, I was comparing Google. Those are the reasons I don’t use Google search, I hadn’t tried Stract yet.

          After trying it, it seems cool. Not the best at broad meanings though. “Ram” returns an Indian politician as the “answer”, a site in Japanese for the first link, and then mostly results for Random Access Memory after. No reference to the Dodge Ram (thank Odin), but also no reference to male sheep.

          It also feels very anti-store, which is a nice change, but might ve an artifact of the seemingly anti-SEO stance, with random results from anywhere. Maybe that’s just the European focus?

          It also has issues with getting context from multiple keywords, and doesn’t prioritize say “street car” over pages that happen to contain both “street” and “car”. Excluding keywords with “-” works though, very nice. Quotes can help with phrases to, so " “street car” " finds exactly things called “street car” with the space. Both still miss streetcars though. Misspelling corrections are offered but not assumed, which is very nice.

          Definitely the biggest issue is the seemingly random results. This might be good if you’re searching for an exact string that is only present in a few places, but anything common and it’s a crapshoot. It’s nearly unable to find anything to do with shamrocks, prefering to find business’ named Shamrock.