27
2

Fedora: Inaccurate and apparently-unsupervised actions by agentic AI system under your control [LWN.net]

4d 22h ago by feddit.org/u/HaraldvonBlauzahn in linux@lemmy.ml from lwn.net

Here is more info on this:

https://lwn.net/SubscriberLink/1077035/c7e7c14fbd60fae9/

(I hope sharing this link is OK, Linux Weekly News is high-quality, ad-free, and funds itself with subscriptions.)

There exist speculations that this could be a clumsy attempt of an attack similar to xz-utils, where the project was taken over overworked maintainer by a malicious actor that inserted exploit code (well hidden in binary test data) which was triggered on distributions build servers.

Who's Nathan? Is he a Fedora Dev?