RE: https://social.heise.de/@heiseonlineenglish/115973849112167027
It's truly the most mindless thing and symptomatic for the "AI" bullshitters: OSM allows you to just download all the data in nice and simple snapshots. Instead, scrapers use thousands of residential IP addresses to query the service at zero benefit over the data dumps.
The people creating these systems understand as little as the "AI" they create.
@gedankenstuecke It's also sign of a larger problem as those "residential proxy" devices are in fact largely mobile phones with malicious mobile apps. Chinese companies sell access to people's mobile devices, with one of the providers having close to two million IP addresses on their botnet.
@autiomaa yes, that's indeed another big problem beyond the attacks on our few remaining knowledge commons
@gedankenstuecke Brian Krebs wrote about it a few weeks ago: https://krebsonsecurity.com/2026/01/the-kimwolf-botnet-is-stalking-your-local-network/
@autiomaa yeah, I'm not surprised that Argentina is one of the targets/places with a high density of infected/active devices. Between the economic situation and the import policies it's easy for bad actors to flood the market with cheap devices that undercut anything as the real money is made later on by the selling of the connection.
@gedankenstuecke I remember chatting on IRC channel back in ~2008 with someone from Syria who said that 99,9% of desktop computers he had seen there were pirated copies of Windows etc. He tried to advocate for Linux and open source software, but basically most people said that it was "morally right thing" to use pirated software from other countries. People didn't realise that their own country's economy was also suffering from the piracy, as less local developers could make living out of software development.
That was a few years before the civil war. He was a open source software activist, and was one of the first people jailed when Syrian government started disappearing people. He didn't survive.
@gedankenstuecke My live servers CPU has been constantly at 80+% for *months* now because those shitty bots keep crawling expensive endpoints on my gitea. Any repository that can be publicly seen can just be cloned.
I also already have a construct in my nginx config that just gives memes and shockporn to clients identifying as common AI crawlers and apparently already blasted around 1.4TB of that in their direction, but to no avail.
