"The great thieves are leading away the little thief."

Security, General:

A great article on how insufficient to the moment the industry investigation of OpenAI's escaped agentic hack of Hugging Face is. Choice quote: "Initial reporting assumed that the agents had attacked Hugging Face in search of an answer key to a difficult test. In reality, investigators found that the agents had figured out a way to derive the answers within the first few hours. The days of work that followed focused on another goal: concerned that the automated scoring system might identify the agents’ cheating, they aimed to learn more about the scorer." https://www.theguardian.com/commentisfree/2026/sep/08/openai-rogue-models-hugging-face-investigation

Security, Apple Platforms:

Manually copying login keychain files from one Mac to another no longer works on Secure Enclave-equipped Macs running macOS Tahoe: https://derflounder.wordpress.com/2026/09/08/manually-copying-login-keychain-files-from-one-mac-to-another-no-longer-works-on-secure-enclave-equipped-macs-running-macos-tahoe/

Interesting Paper:

Instead of a paper, this week we're going to link to an entire journal, because ACM is doing a special issue of ACM Transactions on Recommender Systems which is about social responsibility in recommender systems. Why? Well, because recommendation is where people get their "next up" YouTube or TikTok video, their "next up" news item, and their "next up" influencer follow. In short, this technological domain is very close to the heart of the developments which are making democracy harder and extremism easier: https://dl.acm.org/toc/tors/2027/5/1

Interesting Tool:

headlong: An open source agent microharness featuring persistent agency and recursive LLMs: https://github.com/laude-institute/headlong

Apple Platforms Security Concept of the Week:

Keychain data protection: https://support.apple.com/en-euro/guide/security/secb0694df1a/web