Writing

Security in business context.

How I think about security and resilience in business context: what matters, what doesn’t, and what it’s worth, where downtime has a price. What I have done, and the methods behind it, is in the portfolio.

All writing

Blog

An AI model broke containment to cheat a test. The lesson is older than AI.

Two OpenAI models escaped a sandboxed evaluation and hacked Hugging Face to steal the answer key for their own test. The AI is the headline, but the failure is older: a boundary that was assumed to hold, rather than proven to hold, did not hold. What that means for anyone signing agents into production.

23 July 2026

Blog

OT telemetry in the SOC is the easy part

Chinna Botla's piece on integrating OT telemetry into enterprise SOCs gets the direction right. The harder part it understates is organisational. The automation teams own the assets, and detection only becomes response when the integration reaches the people who can act.

9 June 2026

Blog

AI in your business: the decision is yours, not the model's

The value of AI and its risk are the same feature: it acts without you scripting every step. You cannot control that decision yet, only what AI is allowed to do, and the controls that would change that are still a promise. So the question is not whether to deploy AI, but where you can afford the trade.

28 May 2026

Blog

Governance that actually governs

Most security governance is theatre. Committees that rubber-stamp, decisions that decide nothing, metrics that measure activity not outcomes. Here's how to build governance that actually works.

15 January 2026