• daniskarma@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    2
    arrow-down
    1
    ·
    edit-2
    8 days ago

    I tried, and not really.

    I had to scrape a site that have some captcha and no AI was able to consistently solve it.

    In order to be able to “crack it” I had to replicate the captcha generation algorithm best I could and train a custom model to solve it. Only then I could crack it open. And I was lucky the captcha generation algorithm wasn’t to complex and it was easy to replicate.

    This amount of work is a far greater load than Anubis crypto challenges.

    Take into account that AI drive ocr drinks from existing examples, if your captcha is novel enough they are going to have a hard time solving it.

    It also would drain power, which is the only point of anubis.

    • mholiv@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      ·
      8 days ago

      There is a difference between you (or me) sitting at home working on this and a team of highly motivated people with unlimited money.

      • daniskarma@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        1
        ·
        edit-2
        8 days ago

        The thing is not that it cannot be done, the thing is that the cost is most likely higher than Anubis.