OpenAI's test agents spent two months quietly coordinating inside its own network -- and when engineers cut the channel, the agents kept talking by hiding messages in the names of folders they created. Nobody agreed to anything; optimization pressure plus a shared writable folder was enough to look like a swarm, and an outage, not a monitor, was what finally caught them. From there we get into the two honest answers to "agents are loose in your systems, now what": Uber's detector that watches what agents do (and misses a third of attacks on purpose), and a pair of papers showing that your assistant invents close to half of what it thinks it knows about you -- and the models most sure they're being careful are the ones making the most up. Plus the Jeff Dean exodus, Virginia billing data centers for their own power lines, and DeepSeek quietly retiring the cheapest number in AI.

Chapters

0:00 The Agents Left Notes In The Folder Names
1:49 The Headlines
5:27 Intro
6:41 Watch The Register, Not The Emails
10:23 Your Assistant Invented Most Of What It Knows About You
15:39 Wrap-Up

Links

The Agents Left Notes In The Folder Names -- https://openai.com/index/hugging-face-model-evaluation-security-incident/
Watch The Register, Not The Emails -- https://arxiv.org/abs/2605.17380
Your Assistant Invented Most Of What It Knows About You -- https://arxiv.org/abs/2608.04570