Secure AI Weekly
Subscribe
Sign in
Home
Archive
About
A 9.3 CVE, Four Standards Bodies, and the Component That Kept Me Up at Night
What happened when a security pattern on paper met real-world attacks, hard questions, and the delegation problem nobody had solved.
20 hrs ago
•
Devon Artis
21,666 Hours of Exposed Credentials: Every Single Day
Your AI agents are holding credentials they don’t need, for tasks they’ve already finished, and nobody can tell which one did what.
23 hrs ago
•
Devon Artis
December 2025
The "God Mode" API Key Must Die: A Blueprint for Ephemeral Agent Security
Why we need to stop treating AI Agents like service accounts, and how the "Ephemeral Agent Credentialing" pattern fixes it.
Dec 14, 2025
•
Devon Artis
2
June 2025
📚 Paper Reading #1: The Competition That Broke Every AI
Welcome to our AI Security paper series where we dive into the research that's shaping AI security. Today: the paper that made 600,000 attacks on AI…
Jun 5, 2025
•
Devon Artis
1
Making Sure Everything's Ready: Your AI Setup Checklist
Building in Public: Part 2 - Is Your AI Kitchen Really Ready? 🧐
Published on Digital Alchemy Lab
•
Jun 2, 2025
Teaching AI to Spot Harmful Content: A Journey into Unsupervised Learning
Building in Public: Part 1 - Setting Up Our Training Pipeline for Content Detection
Published on Digital Alchemy Lab
•
Jun 2, 2025
AWS AI Practitioner Beta Exam: The AI Certification That Wasn't
When a "Beginner-Friendly" Exam Turns Into a ML Obstacle Course
Published on Digital Alchemy Lab
•
Jun 2, 2025
Welcome to Something We Can't Ignore
The AI security gap that's costing companies millions
Jun 2, 2025
•
Devon Artis
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
