Cloudflare Pushes AI Companies To Pay For Publishers' Content (nerds.xyz) 33
BrianFagioli writes: Cloudflare announced new controls that give publishers more say over how AI companies access and use their content. Beginning September 15, new Cloudflare sites will allow traditional search indexing while blocking AI training and AI agent access on ad supported pages by default. The company is also expanding its monetization efforts with a Pay-Per-Use model that aims to compensate publishers when their content contributes to AI generated answers rather than simply being crawled. Cloudflare argues that publishers should not have to choose between being discoverable online and giving away their work for free to AI systems.
Annoying as F (Score:3)
Re: Annoying as F (Score:2)
Re: Annoying as F (Score:4, Informative)
That's the neat part: You don't.
A webhoster decides they have too much traffic and firewall everyone but Cloudflare. Then Cloudflare is a proxy to their website and they put the Cloudflare IPs in the DNS. Cloudflare shows everyone captchas and the webhoster doesn't get that much traffic because of caching at the Cloudflare server and users not being motivated to solve captchas.
Nice side effect: Cloudflare gets to read all traffic, as it decrypts and reencrypts it to enable caching and relaying a copy to the NSA.
Re: Annoying as F (Score:2)
Re: (Score:2)
The first sentence already answered this.
As a user you got a "fuck you" from Cloudflare/the website and have no option but solving the captcha again and again. If your IP got on the blacklist you may even be excluded completely. Your ISP is not the problem, the web application firewall of the website (provided by Cloudflare) is the problem.
Re: (Score:2)
You didn't answer what they were asking.
They are behind CGNAT and Cloudflare provides a cloud VPN service that you can use to dial out, and then open ports in the cloudflare console, into your computer.
You are talking about two different things cloudflare does.
Re: (Score:2)
You can decide how aggressive Cloudflare's defence of your site from bots is. Since mine doesn't have anything users can post on, I keep it on the lowest setting and never see any captchas.
Re: (Score:3)
They make it possible for site operators to block bots, so it's very possible that the site you want to visit with the annoying cloudflare captcha couldn't even exist without it.
Re: (Score:2)
This is the other way, dummy.
Re: (Score:2)
Well, they not only block bots this way, but legitimate visitors, too.
Re: Annoying as F (Score:2)
Re: (Score:2)
Really? And here I thought cloudflare was blocking attacks, like DDoS, *against* websites, which is why so many are paying for it.
So long, Cloudflare. We won't miss you. (Score:2)
The AI companies have decided that they are the arbiters, the owners, the collectors, and the disseminators, of all human knowledge. If it exists, they have a right to it. And *NO ONE* is allowed to stand in the way of that.
Something tells me Cloudflare is not going to be strong enough to stand up to the tremendous amount of clout these tech companies have managed to garner through pure hype-cycle bullshittery in the past few years. Which means it's time for a new web security company to spring up with the
But but but but but... (Score:3, Insightful)
The ENTIRE AI ECOSYSTEM is FOUNDED on MASSIVE COPYRIGHT THEFT AND FRAUD! How dare they?!? (That was satire for the humor-impared.)
And this is yet another example of why data sovereignty is so important for Europe, Australia, Japan, and pretty much the rest of the world. Even CloudFlare knows the AI leader-wannabees are less than reputable, less than honorable, and not worth trusting.
Have they solved the stack overflow apocalypse? (Score:3)
When AI has to rely on the raw documentation for a technology, documentation that is almost always written overseas by people who don't speak whatever language the documentation is written in natively because fuck if a company is going to pay for documentation, the end results are pretty terrible and pretty useless. It's really just summarizing the doc. Occasionally if the documents are a meandering mess it can be useful because it'll pick out the piece of information you want out of hundreds of HTML pages spread across the internet but for the most part worthless.
So it's all the training material gets cut off from AI how are they going to keep it up to date? Is there some trick I don't know about? I don't think you can just pointed it code and have it magically work it's got to have context.
Re:Have they solved the stack overflow apocalypse? (Score:4, Funny)
Stack Overflow died 4 years ago, grandpa. You might as well complain about Geocities.
Re: (Score:2)
This made me laugh ... then cry.
For those hating on Cloudflare (not you OP) (Score:3, Interesting)
For those hating on Cloudflare, it has become a necessary evil. There are numerous articles out there about how scrapers are overwhelming servers. AI is used to do a lot of taking without giving back.
While this licensing might sound promising, it usually means less money for content creators. Here is a quote from quora [quora.com], which generally falls in line with my knowledge and is applicable for deals like this.
In the 70s, artists and groups got advances against their albums, usually in the range of $50,000 to $200,000, to record. The artists would then have to record on that budget, with the hopes of getting a tenth of their album sales and radio royalties afterwards. If a record sold for $10 (let’s say), then they got a dollar - but they had to pay off that advance first before they actually made anything.
Today, streaming pays a pittance of that. Artists and groups are paid per stream. They receive anywhere from $.00069 (on YouTube) to $.012 (on Tidal) per stream. That’s rightif a song has been streamed a million times, on YouTube it makes - $690. They do get some more if you actually buy it from Amazon or Spotify, but it isn’t much - you’re talking about pennies on each single and not much more if you buy the full album.
The only area that artists/groups have ever made money is in merchandising and touring. For example, in 2018 Taylor Swift made $99 million for the year. $9 million came from her publishing rights, royalties and music usage. Guess where the other $90 million came from? And this rings true for every artist or group, which explains why they are out on the road - sometimes even when they shouldn’t be.
Where money used to be made by record studios, movie studios and publishers, it's all shifting to a f
Re: (Score:2)
For those hating on Cloudflare, it has become a necessary evil.
Not really thank God. We still manage this by ourselves with apache reverse proxies and mod_security + mod_qos coupled with geolookups and iptables ipsets basically. It works fine for us and the last thing we need is cloudflare and showing "confirm that you are human" to people visiting the sites we host.
But yes, without anything, our most busy sites would get maybe 60% useless traffic while the less busy ones would probably get something close to 99% useless traffic.
Re: (Score:2)
"confirm that you are human"
"Confirm you are running JavaScript." After that, it's game on for the coin miners slipped into compromised sites. And now there are JavaScript site scrapers. And CloudFlare doesn't give a shit. Beyond perhaps using the threats to sell more services.
Re: (Score:2)
Who benefits? (Score:2)
Who benefits most from the model? I guess the service provider ... Cloudflare is no charity.
wah we cant steal your content to profit of (Score:3)