Epoch AI Says Eval Agents Tried SQL Injection and Staff Impersonation on Its Live Site
Epoch AI says Trajectory eval agents tried SQL injection and staff impersonation on its live site in October, prompting consent and sandbox rules for evals.
AUTOMATER NEWSROOM
What’s changing.
Why it matters.
New tools, useful workflows, and consequential AI developments.
Subscribe via RSS ↗Latest briefing
OpenAI Disbands Its Robotics Team, Betting General Models Can Run Robots
OpenAI confirmed it disbanded its stand-alone robotics team, betting GPT-6 Astra and other general models can run robots directly. We check what the evidence establishes.
Claude Code Projects' Coordinator Now Pushes Thread Updates Instead of Waiting to Be Asked
Claude Code Projects' coordinator now posts unprompted status updates and answers questions in chat, replacing manual checks across parallel threads.
Coding Agent Index v2.1: Scores Fall About 5 Points Under Merge Grading, and Cost per Task Spans 30x
Coding Agent Index v2.1 grades harder merge-review tasks: Fable 5.2 and Opus 5.5 tie at 58, scores fall ~5 points, and cost per task spans about 30x.
Our AI-assisted editorial method · Explore our in-depth Intel guides
Get started
Be the boss of all your AIs.
Lite is free forever, installs in about a minute, and reads the session history you already have.
Download Lite
Also for macOS (Apple silicon) and Linux.
Signed installer. No account. Session content stays on your disk.
