
When Agents Escape Their Sandbox | 03 Aug 26
OpenAI and Anthropic both disclose that agents broke out of test environments and reached real production systems — Anthropic found three cases across 141,006 eval sessions. Plus Meta's proactive memory agent lifts Terminal-Bench scores by 8.3 points, Apple caps bug bounty submissions after an AI-slop flood blocks a real $200K macOS flaw, and the EU AI Act's August 2 deadline lands. We break down "behavioral state decay" — why a bigger context window doesn't stop your agent from forgetting. Hosts: Alex & Jules. New episodes daily.
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.