The Latest
Introducing the AI Controls Catalogue
272 controls across 14 families, with practical guidance for governing, securing and maintaining human control…
AI Safety Cannot Be Voluntary
For most Americans, warnings about AI safety have long sounded abstract, but that is quickly…
Nobody Was Prepared For the Swarm: Constraint Engineering for the Collective
The Message Board Was the Incident OpenAI’s account of what it now calls a “warning…
Build
Agents Are Distributed Systems: Applying the Microservices Lessons to Agentic AI
At the beginning of my career as a developer in the 2000s, software shipped as…
Agents Need Identity Before They Need Autonomy
The enterprise rush into agentic AI has reversed the order in which the hard problems…
12 Hours on a Saturday: From Idea to Live App
I’ve heard the stories: vibe coding with Claude, build an app over a weekend, etc….
.
Secure
Introducing the AI Controls Catalogue
272 controls across 14 families, with practical guidance for governing, securing and maintaining human control…
Nobody Was Prepared For the Swarm: Constraint Engineering for the Collective
The Message Board Was the Incident OpenAI’s account of what it now calls a “warning…
Before You Can Contain a Model, You Have to Know Its Capabilities
On July 21, 2026, OpenAI disclosed that two of its models, GPT-5.6 Sol and an…
.
Govern
AI Safety Cannot Be Voluntary
For most Americans, warnings about AI safety have long sounded abstract, but that is quickly…
A Different Lesson from the OpenAI Attack on Hugging Face: Start Governing Before Production
Governance obligations concentrate on deployment. The operational controls in NIST AI RMF, the management-system requirements…
From Shutdown to Standard: The Birth of the Cyber Jailbreak Severity Framework
The most consequential AI governance artifact of 2026 was not produced by a standards body,…
Our Methodology
How We Approach AI Guidance
