Blog
Notes on identity, security and the move into AI engineering.RSS feed
- 10 Nov 2026
Is AI governance keeping pace with innovation, or are we kidding ourselves?
Most AI governance frameworks quietly assume mature identity and data governance already exist underneath them. In most organisations I see, that foundation isn't there yet — and that's the actual gap.
- 19 Oct 2026
AI agents don't skip ALM — they just make you watch it happen in real time
Retiring one tool and rebuilding another on this site surfaced the same lesson twice: build, test and review each catch things the one before it missed, whether the developer is human or an AI agent.
- 22 Sept 2026
Revisiting "AI and Our People": what's actually changed
A year and a half ago I wrote about bringing people along with AI. Rereading it now, the cultural argument holds up — but what "people-first" actually requires has changed underneath it.
- 15 Sept 2026
Building a Homelab SOC on a Raspberry Pi with MCP and Microsoft Graph
How I turned a £70 Raspberry Pi 4 into a real-time security operations centre — complete with MCP servers, a live dashboard, Microsoft Graph alerting, and Tailscale as the backbone.
- 06 Sept 2026
I rebuilt my website with AI agents in six days. Here's what that actually looked like.
A first-person account of rebuilding cohenholmes.co.uk with Claude Code and GitHub Copilot — what got built, what went wrong, and why running two orchestrating agents on one repo was a mistake.
- 03 Sept 2026
Conditional Access policies: what exactly are they, and where do you start?
A practical, ordered baseline for rolling out Microsoft Entra Conditional Access in a normal commercial tenant — the reasoning and the traps, not a 40-policy checklist nobody deploys.