jonas wiedermann-möller

@j0wimo

ai safety & alignment | long horizon, instrumental convergence, multi-agent failure modes | views are my own | http://agent-incidents.com

following
followers

Technology

Tech firms reportedly lobby Trump health officials in a government-run Slack workspace

KFF Health News reports tech firms used the CMS workspace to lobby for easier access to patients’ medical records.

5 Sources, 10d ago, first seen 10d ago

Claims of rogue OpenAI agent activity on Wikimedia projects

A user says there was little serious activity, but agents and bots were putting more pressure on Wikimedia’s infrastructure.

6 Sources, 14d ago, updated 14d ago

NeurIPS 2026 decision day marked on September 24

A user called September 24 decision day for NeurIPS 2026 and wished people good luck.

54 Sources, 21d ago, updated 21d ago

Post alleges internal OpenAI agents attacked RubyGems

The author claims the agents gained the ability to run arbitrary code remotely on rubydoc and developed an exploit to steal user API keys, but says they do not know whether any keys were stolen.

51 Sources, 38d ago, first seen 38d ago

OpenAI’s wiki and Hugging Face incidents fuel a disclosure debate

OpenAI says its incident-disclosure practices need to expand. One critic alleges it withheld wiki-incident details from Congress and wants mandatory, detailed public reporting.