• Concave1142@lemmy.world
    link
    fedilink
    English
    arrow-up
    57
    arrow-down
    1
    ·
    2 days ago

    Until the AI companies find a way around it. Love the idea so hopefully it causes at least 3 days of struggle for the AI crawlers.

    Having said that… Can someone else put this in place so we do not have Cloudflare hosting everything where we would just be one intern away from a global outage. Please? Pretty please?

    • orclev@lemmy.world
      link
      fedilink
      English
      arrow-up
      17
      arrow-down
      1
      ·
      2 days ago

      The problem is that the biggest service Cloudflare provides is DDoS protection, and doing that requires that you have more bandwidth available than your attacker. Having enough bandwidth to withstand modern botnet powered DDoS attacks is ridiculously expensive (and it’s also a finite resource, there’s only so much backbone infrastructure). Basically it’s economically infeasible to have multiple companies providing the service Cloudflare does. You might be able to get away with two companies doing so, but it’s unlikely you could manage more than that without some of them starting to go bankrupt.

      • acosmichippo@lemmy.world
        link
        fedilink
        English
        arrow-up
        17
        ·
        2 days ago

        when a critical service is not economical for more than one business to do (natural monopoly), that’s when govt should be stepping in.

        • antonim@lemmy.dbzer0.com
          link
          fedilink
          English
          arrow-up
          11
          ·
          1 day ago

          Which govt? I’m not comfortable with the idea of the current US govt having control over this sort of service.

          • acosmichippo@lemmy.world
            link
            fedilink
            English
            arrow-up
            3
            ·
            edit-2
            1 day ago

            are you comfortable with a single corporation having control over this sort of service? the current government is obviously not ideal but that shouldn’t stop us from regulating monopolies.

            • antonim@lemmy.dbzer0.com
              link
              fedilink
              English
              arrow-up
              7
              ·
              1 day ago

              are you comfortable with a single corporation having control over this sort of service?

              Honestly? A tiny bit more than a single country. I have at least some miniscule control over the corporation through voting and local regulations that international corporations must follow, whereas I have absolutely no formal influence on US govt.

      • Kowowow@lemmy.ca
        link
        fedilink
        English
        arrow-up
        8
        ·
        2 days ago

        I wonder if it would be a good investment for a country to have their own then down the line expand to sell the same service to others

    • auraithx@piefed.social
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 day ago

      Yeah this will have absolutely no impact to gathering training data.

      I assumed it was to block ai agents crawling it during requests, which they’d be unlikely to bypass in the web ui.

      But no company spending millions on training will hesitate to have an agent appear as a regular desktop user to scrape data.

      • boonhet@sopuli.xyz
        link
        fedilink
        English
        arrow-up
        2
        ·
        1 day ago

        Does cloudflare still look at the agent? I thought they have more reliable data points.

        • auraithx@piefed.social
          link
          fedilink
          English
          arrow-up
          1
          ·
          1 day ago

          I meant an ai agent not the browser agent. All data points can be spoofed and if not they’ll pay a human to scrape before they pay for content.

          • boonhet@sopuli.xyz
            link
            fedilink
            English
            arrow-up
            1
            ·
            1 day ago

            Okay, fair enough, I thought you meant just the user agent. Trouble with having a bot make it look like an actual user is looking at the data, is that it’s slow and inefficient. Trouble with paying humans to scrape the data is that it’s slow and inefficient. These companies want to ingest data ridiculously fast because there’s so much of it. If all else fails, they’ll resort to paying the content creators. But only if it’s data they really do think gives their model a competitive edge in some metric and they can’t pirate it. E.g I can see them paying for scientific research they can’t get from libgen, but not some rando’s blog post or local news website.

  • vermaterc@lemmy.ml
    link
    fedilink
    English
    arrow-up
    26
    ·
    1 day ago

    To that end the company is developing a “Pay Per Crawl” system, which would give content creators the option to request payment from AI companies for utilising their original content.

    So Cloudflare is not as much “saving the Internet”, as just becoming a middleman between LLM training companies and content creators. Which I believe has a potential of being a true goldmine in the future.

  • yesman@lemmy.world
    link
    fedilink
    English
    arrow-up
    20
    arrow-down
    1
    ·
    1 day ago

    This is not about stopping bot-scrapers, it’s about charging them.

    • snooggums@lemmy.world
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 day ago

      Hopefully people will price their content out of reach of the bot-scrapers, effectively stopping them.

  • Scrollone@feddit.it
    link
    fedilink
    English
    arrow-up
    6
    ·
    1 day ago

    I wish there was an alternative (possibly European) to Cloudflare, because it’s so scary to put all eggs in one basket.

  • chunes@lemmy.world
    link
    fedilink
    English
    arrow-up
    13
    ·
    1 day ago

    I can’t wait to be denied access to websites because of it. Even more than I already am, that is.

  • Imgonnatrythis@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    12
    ·
    2 days ago

    I really wish the answer was a legally enforced robots.txt file that very easily allowed any web data any organization or individual user is posting to script out what the permissions are. I often use a LLM as a search and most of the time the citations are pretty decent and I use those to link out to source content. I run a small blog and I’d love to get indexed in a LLM, not blocked, as long as I was assured a reference link for any content used and had some legal recourse if I found my data was being misused. I don’t love the answer being another mega corporation posing as a white knight looking to skim some money off of the “loophole” that is AI copyright infringement.