Comment on Cloudflare blocking AI crawlers

lambalicious@lemmy.sdf.org ⁨2⁩ ⁨weeks⁩ ago

• grue@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

  Yeah, it’s only anecdotal but I feel like hobbyists like us, who do slightly unusual things without nefarious intent, who are the ones who get hit with these sorts of issues the most. For example, I’ve noticed that some websites start throwing captchas at me or even just straight-up refuse to load with 403: unauthorized errors because I have my router set up to load-balance across two Internet connections. (At least, that’s my guess as to why it’s happening.)

  source

  • Buelldozer@lemmy.today ⁨2⁩ ⁨weeks⁩ ago

    For example, I’ve noticed that some websites start throwing captchas at me or even just straight-up refuse to load with 403: unauthorized errors because I have my router set up to load-balance across two Internet connections. (At least, that’s my guess as to why it’s happening.)

    I maintain several multi-wan commercial setups and they don’t have this problem. I obviously don’t know what your setup is but I’d guess something is wrong with how its handling flows / connections. Once a connection is established between your edge and an internet resource that flow should remain “stuck” to whatever wan port it started with and it sounds like that isn’t happening.

    source

    • grue@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      Could very well be. I’m using OpenWRT and basically did the bare minimum to get it to work.

      source
  • Scrollone@feddit.it ⁨2⁩ ⁨weeks⁩ ago

    Ahh yes. Imgur simply don’t work anymore at my place, it always errors out with 403.

    source
• irmadlad@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

  I’ve seen captchas for years before the recent influx of AI. It’s the way I go about obfuscating network activities that the site security cannot determine if I am a bot on not. There is a Captcha Buster extension for Firefox. If the captcha is ‘Pick the three busses from these blurry, pixelated set of pictures’ then I can solve those easily. It’s when the captcha is a full page of a motorcycle and you have to check all the relevant pieces, then on to the next full picture, that chap me. So you click Captcha Buddy and it ‘listens’ to the audio portion of the captcha, then solves it. It’s not 100% on all types of captchas, but it 90% of the time it works every time. It’s interesting to me that after a while, you start to notice patterns in the captcha images. For instance if the directions are ‘Pick the fire hydrants’, there will be at least 5 you have to pick. Crosswalks are the same way too.

  I’d much rather have to do captchas than have my jimmy out in the ether traffic. Anecdotal, but Stack Overflow doesn’t trigger a captcha for me. All I get is the cookie popup.

  source

  • Andres4NY@social.ridetrans.it ⁨2⁩ ⁨weeks⁩ ago

    @irmadlad @lambalicious I just manually do the audio captcha. Every time. Because the picture captchas often don't work correctly for me.

    It does bug me a little that I don't know what the audio captcha is being used for - am I helping an amazon echo transcribe whatever it is surreptitiously listening to?

    source

    • irmadlad@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      am I helping an amazon echo transcribe whatever it is surreptitiously listening to?

      I’ve always wondered where the hell they scrape all that audio from. I mean, it’s random shit.

      source

      • lambalicious@lemmy.sdf.org ⁨2⁩ ⁨weeks⁩ ago

        Gotta be physicists or fanfic writers. I can not imagine other better options.

        source

        • -> View More Comments