The gall of Cloudflare to think they can be the arbitrator of scraping… this is what happen when all y’all give a singular publicly-traded, US-based major control of the general internet infrastructure by purposefully letting them man-in-the-middle your production sites. Now they get to sell access to what was once an open internet. Instead every time Clouflare or Fastly go down, half of the internet goes down with them.
“We can scrape open internet” is such a CEO take. I’m no fan of Cloudflare but what they’re doing here is good for open web and bad for AI bros
This assumes that just since abuse could happen means we should block access for everyone. Folks might make illegal photocopies of books so we should ban libraries. I & others have done general scraping for our own uses that isn’t done in some abusive manner. But to assume a company beholden to US shareholder is going to “to the right thing” would be to go against the history of US corporations.
And you know who is going to be able to afford to do the scraping? Big US-based “AI Bros” that can do it with venture capital preventing the average user or researcher from grepping the net.
Good points, you made clear this is no black/white case, it depends on how much we can trust cloudflare. It’s a US based for-profit company so the answer is “not much”.
However, if you check the link, you’ll see that they’re able to distinguish AI bot scrapping from other forms of scrapping. They also give the website owners a choice, so if it’s about principles, owners can choose to boycott, as most are already doing with robots.txt, which AI bros have no respect for.
Cloudflare does not benefit from a handful of websites getting all the traffic, and their track record is good so far. They also don’t profit from people visiting websites with a browser, they don’t show/own ads. To me, they have enough credit for me to believe they can protect open web.
We’ll see how this take ages though. I still won’t put all my eggs in a single basket.
They don’t think they’re the arbitrer of scraping dumbass, this is a service they’re selling to websites who don’t want AI scrapers taking adnatage of them.
They are precisely deciding who can scrape what by sitting in the middle of like 30% of internet traffic & denying access. There is no way to tell if this ‘scraping’ is for research, hobby, commercial, or “AI” purposes; conveniently if it can make Cloudflare money, they’ll let you charge a toll. If Cloudflare cared about AI issues, they wouldn’t be having unpaid users solve/train their hCAPTCHA models just for visiting a site from Tor, a VPN, or even just a non-‘Western’ IP address. The fact that folks/businesses bought into this centralization is frightening—with little open access to information or allowing folks to stay anonymous (whatever their motivation).
Also don’t dare call someone “dumbass” if you can’t be bothered to turn on spell check or understand how commas work.
Them being able to offer this service, and them proxying 30% of the internet are completely unrelated. Any other company could offer this scrape protection if they wanted, with roughly the same cost of entry.
You can hate cloudflare all you like, but only a certified dumbass would try to pretend this feature is somehow enabled by their market dominance…
A US corporation cornering the market then charging their own fees to use the service (surfing & scraping are basically the same action)? Never seen this before. Sounds uber evil—I couldn’t airbnbelieve it.
CF needs some good PR after the gambling website stuff.
The what
I guess if you can’t beat them it’s the next best thing.
Another wall to try and make it so only the rich corporations can make AI, what a shock.
This is what all the useful idiots bleating against ai have been working towards.
There are models that have only used data with permissive licenses.
