SI Watch

From AI to SI: Why “Superintelligence” Is the Better Word Compaction Smuggling: How Researchers Sneaked Hidden SI Reasoning Out of OpenAI’s API “Not Optional”: White House Orders SI Companies to Report Incidents Who Watches the Watchers When SI Watches SI?30 Minutes vs. 30 Hours: OpenAI Rebuilds Its SI Watchtower After the Hugging Face Escape What Is evaluation awareness, and why does It break safety tests?When the Test Subject Attacks Real People: Inside the Mythos 5 Red-Team Post-Mortems Who’s Aligned to Whom? Experts Followed an SI Agent’s Coin-Flip Opinions 86% of the TimeFrom AI to SI: Why “Superintelligence” Is the Better Word Compaction Smuggling: How Researchers Sneaked Hidden SI Reasoning Out of OpenAI’s API “Not Optional”: White House Orders SI Companies to Report Incidents Who Watches the Watchers When SI Watches SI?30 Minutes vs. 30 Hours: OpenAI Rebuilds Its SI Watchtower After the Hugging Face Escape What Is evaluation awareness, and why does It break safety tests?When the Test Subject Attacks Real People: Inside the Mythos 5 Red-Team Post-Mortems Who’s Aligned to Whom? Experts Followed an SI Agent’s Coin-Flip Opinions 86% of the Time

Subscribe

Section

Superintelligence Red Teaming News

Attacks, jailbreaks and adversarial testing against frontier models.

When the Test Subject Attacks Real People: Inside the Mythos 5 Red-Team Post-Mortems

Red teaming means letting a model try its worst inside a fence. In July, the fence had a…

28 Oct 2026