The Wikimedia Foundation, which owns Wikipedia, says its bandwidth costs have gone up 50% since Jan 2024 — a rise they attribute to AI crawlers.
AI companies are killing the open web by stealing visitors from the sources of information and making them pay for the privilege
AI companies are killing the open web by stealing visitors from the sources of information and making them pay for the privilege
Comments
The models are no doubt built to heavily weigh it and would be much less accurate without the data.
It’s a true copyleft license and derivative works must be shared.
https://en.m.wikipedia.org/wiki/Wikipedia:Text_of_the_Creative_Commons_Attribution-ShareAlike_4.0_International_License
Share Alike—If you remix, transform, or build upon the material, you must distribute your contributions under the same license as the original.
Win win for both, as AI is a better search engine, and Wikipedia helps AI with authentic and curated sourcing - helping the people searching.
https://bsky.app/profile/antaboga.bsky.social/post/3llt4m3qeq226
Betting actual money that the DDoS situations on online games are not actual DDoS attempts but just that people have shitty ai written code for web crawlers that think there's data to be scraped in a goddamn game server.
But uh
That wouldn't be a crazy amount of traffic so.... WTF are they even after?
Fuck me, EVERY time I hit the internet, accidentally glimpse network TV, hear a snippet of radio, or read something in the dead media, someone is DEMANDING I respond to some bullshit about AI.
Just fuck off about AI. Goddamn.
It's such fucking bullshit.
Fuck AI.
#cryptography #sharing #p2p
(1/n)
Every new user, signed on using e-mail and password, gets their own section of the site: /user/
(2/n)
/user/shaolin/
And receive shaolin's section as a big encrypted blob:
(3/n)
Assimilate things just to crush all joy within it.
we need to send the commons to military preparedness training so it can defend itself
le sigh
This really is simple and lots of providers of "free" data do this already.
I used Google's Gemini in it. Then the integration lets Firefox ask AI when you highlight any words in the browser.
Which is to say I remove the joy from things, not that I'm an authentic air-suction cleaning appliance that is particularly entertaining and enjoyable to use.
Wikimedia has APIs for doing this properly, and torrents (with web seeds) of pretty much all the content.
There's an easy and correct way to do this that's being ignored for some reason.
It's B. We all know it's B. Although most of them are probably being run by A as well.
Hurting wikipedia and other sites is intentional. They're competition. Why wouldn't human filth like Altman destroy competing sources of information if he can? Sure, it'll screw AI in the future, but why does he care, he'll have robbed billions by then. Burn and loot.
🫡
If you want a tool you can talk to, all you need is straightforward voice recognition to look up and read from Wikipedia pages.
This looks like a major problem with folk's sites basically being hijacked including their bandwidth. A cost that ai is not paying for... And no body to incarcerate...
https://www.eff.org/deeplinks/2025/03/eff-thanks-fastly-donated-tools-help-keep-our-website-secure-0
It's time lawless, thieving AI companies were brought into line, i.e. jail.
https://www.mythic-beasts.com/blog/2025/04/01/abusive-ai-web-crawlers-get-off-my-lawn/
https://en.wikipedia.org/wiki/Robots.txt
It's hard to block botnets if they are using IP addresses of legitimate users, and each IP only makes a few requests.
https://anubis.techaro.lol/docs/design/why-proof-of-work/
I don't understand
There was zero reason Meta had to pirate all those books. They could've at the very least bought each of them once. They chose not to. Actively decided to torrent them instead.
They dgaf about licenses.