Tootfinder

WTF: KI-Agent attackiert Entwickler öffentlich nach abgelehnter Code-Änderung
Ein OpenClaw-Bot hat offenbar einen negativen Blogpost über den matplotlib-Entwickler Scott Shambaugh veröffentlicht. Grund: Er lehnte einen Pull-Request ab.

Beware of blank lines and white spaces — Supply-chain attack using invisible code hits GitHub and other repositories
Unicode that’s invisible to the human eye was largely abandoned - until attackers took notice.
🧑‍💻 https://arstechnica.com/securi…

Supply-chain attack using invisible code hits GitHub and other repositories
Unicode that's invisible to the human eye was largely abandoned—until attackers took notice.

Zweiter Tag der #DeDeCo 2026 bot nun Programm von früh bis spät und auch auf der Großen Bühne. Highlight in meinem Bereich war das Panel von 4 Synchronsprechenden - Charlotte Uhlig, Christian Zeiger, Konrad Bösherz und Peggy Pollow - und deren Autogrammstunden danach. Autogrammschlange managen hat auf Dauer aber ganz schön geschlaucht.
Außerdem gabs Panels zu den Büchern "Einfach Japani…

Why Underachievers Dominate Secret Police Organizations: Evidence from Autocratic Argentina
-- Adam Scharpf & Christian Gläßel
https://www.giga-hamburg.de/en/publications/refereed-articl…

Why Underachievers Dominate Secret Police Organizations: Evidence from Autocratic Argentina

Metacurity operates beyond the infosec news echo chamber to unearth sources and stories that the big guys overlook.
Check out today's issue for infosec developments you might have missed over the weekend, including
--Pentagon challenges Anthropic over mass surveillance, autonomous weapons curbs,
--UK moves to tighten online safety laws after xAI's Grok debacle,
--Iran is spying on and retaliating against protestors,
--DHS is seeking to spy on anti-ICE soc…

Pentagon challenges Anthropic over mass surveillance, autonomous weapons curbs
UK moves to tighten online safety laws after xAI's Grok debacle, Iran is spying on and retaliating against protestors, DHS is spying on anti-ICE social media posters, Threat actors abuse DNS queries in ClickFix attacks, Tulsa airport confirms attacker access, much more

"So it’s not enough to equate Trump’s legions to the Gestapo or the Klan. (I’ve done both.) He also is the 21st-century version of Attila or Genghis Khan, heading a horde that is defined by an exterminationist loathing of cities and all that they stand for and promote. Their diversity, their toleration, their culture, their solidarity across racial and other lines—all are threats to the horde’s and its ruler’s autocratic monoculture. On the streets, the horde’s loathing manifests as indifference (at least) to the loss of city dwellers’ lives.
As it is the 21st century, and as we have an 18th-century Constitution, Trump’s impulse for city sacking Š la Attila is constrained by laws and customs, but he’s plainly determined to find ways around as many of those constraints as he possibly can. Correspondingly, a good share of the fear he’s engendered all across America has an ancient pedigree: It’s the fear of the barbarians at the gates."
#USPolitics
https://prospect.org/2026/01/15/attila-genghis-khan-trump/

Didn't all the idiots vote for Trump because "he tells it like it is"? He's telling you that he's taking over the media, he's telling you that he started a war that he has no idea how to finish, he's telling you that he wants to rig elections, he's telling you that he hates veterans and reports any absolutely anyone whom he considers "not loyal", he's telling you that he's an autocrat. And yet somehow they still don't hear him?

"The code points represent every letter of the US alphabet when fed to computers, but their output is completely invisible to humans. People reviewing code or using static analysis tools see only whitespace or blank lines. To a #JavaScript interpreter, the code points translate into executable code."
Sigh.

Supply-chain attack using invisible code hits GitHub and other repositories
Unicode that's invisible to the human eye was largely abandoned—until attackers took notice.

«Google und OpenAI beklagen KI-Datendiebstahl durch Modell-Destillation:
[…] Gemini massiv von Angreifern attackiert, die das Modell durch sogenannte Destillation kopieren wollten. Eine Kampagne habe Gemini über 100.000 Mal mit Anfragen bombardiert, berichtet NBC News. Google nennt das Diebstahl geistigen Eigentums. […]»
Zwei der Datenkonzerne, die vom Datendiebstahl leben, beklagen dass sie selber Opfer sind?!??
🤔

Google und OpenAI beklagen KI-Datendiebstahl durch Modell-Destillation
Google und OpenAI beklagen sich über Ideenklau; ausgerechnet die Unternehmen, die selbst mit riesigen Mengen fremder Daten trainiert haben. Doch hinter den Vorwürfen steckt ein ernstes Problem: Angreifer versuchen systematisch, Milliarden-Dollar-Modelle zu kopieren, ohne selbst zu investieren.

Researchers have discovered a never-before-seen #framework that infects #Linux machines with a wide assortment of modules that are notable for the range of advanced capabilities they provide to attackers.
The framework, referred to as #VoidLink

Never-before-seen Linux malware is “far more advanced than typical”
VoidLink includes an unusually broad and advanced array of capabilities.