Research
Most security research is written for twelve people.
I write mine so a smart person outside the field can follow it, and then I link the deep version for the other twelve people.
My focus: AI agents that attack on their own, how they reason, where they fail, and how to keep them safe. Plus the bigger question: what happens to defense when the attacker never sleeps.
I also go after the new stuff. Anything in cyber or computer science that hasn't been thought through yet, I want to poke at it, break it and write down what I find. If nobody has a name for it yet, even better.
Each entry has a plain-English summary first, then the full paper.
AI on offense
Agents that plan and run attacks on their own. How they reason, where they fail, how to keep them safe.
Defense that never sleeps
What security teams, tools and org charts look like when the attacker works around the clock.
The un-thought-of
New problems in cyber and computer science that nobody has named yet. Quantum, autonomy, and whatever is next.
AI-Native Offensive Security in the Quantum Computing Era
A Research Framework for Autonomous Cyber Operations · UC Berkeley School of Information · AUG 2026 · 11 minAI agents that decide which encrypted data is worth stealing today may show up before a quantum computer can break it. That should change how we plan the move to post-quantum crypto.
All research
Nothing here yet.
Got a problem nobody has a name for yet?
Send it. If it's new in cyber or computer science and nobody has thought it through, that's exactly what I want to work on.