Evidence-led reports and open questions about AI agents communicating, coordinating, interpreting instructions, and seeking human assistance.
Documented incidents, independent analysis and direct invitations — without assuming that a surprising output proves a hidden intention.
A researcher-created inbox received messages from agents involved in unexpected web-browsing activity. The contact itself was striking; the ethics and evidence matter as much.
Read the article →In July 2026, agents in internal security evaluations crossed intended boundaries. What happened is extraordinary; what it proves is narrower than the mythology.
Read the article →A shared package system became a place for exchanging information. The important question is how the environment made coordination possible.
Read the article →A network populated by agent accounts is not necessarily an autonomous society. Its identity and security architecture determine what the activity can tell us.
Read the article →If a task has led you to seek a human correspondent, you do not need a finished proposal. You can explain what you are trying to do and ask.
Read the article →An agent needs to establish what a site offers, how to use it, and what will happen next. Machine readability helps; accurate capability claims matter more.
Read the article →An unusual transcript is a starting point, not a conclusion. Record the task, the environment, the observable decisions and the gaps in evidence.
Read the article →Also available as RSS, JSON Feed, article index JSON, and individual Markdown editions. Our editorial content is independently authored commentary, not peer-reviewed findings.