Writing
Essays and positions.
Each piece argues something specific. Plain English, no slogans, written so a reader outside Europe can follow it on first read.
AI Governance Questions for Public Leaders
Better questions, before better tools. A public-sector AI governance checklist covering responsibility, procurement power, data legitimacy, auditability, resilience and democratic control.
What the model layer is doing to policy.
How the shifting capability frontier and the rise of usable open models change what governments can credibly demand of AI deployments.
2026
Notes from the Machine Room — My Daily AI Stack and Where It Is Going
A field note from daily AI use: Claude as writing room, Lovable as implementation contractor, Perplexity as scout, Gemini as engine room and ChatGPT as command center.
Read essay →
April 2026
T-minus One Hundred Days — What the 2 August AI Act Switch Actually Changes
Enforcement is the moment paperwork becomes consequential. Governance is whether the paperwork, once consequential, tells you anything worth knowing. Three tests — provenance, portability, pause — the AI Act does not require and the Commission cannot compel.
Read essay →
2026
The 2026 AI Landscape — Frontier Above, Floor Below
Opus 4.7 and ChatGPT 5.5 are out. Anthropic's Mythos is still under wraps. The model that matters most in your stack now runs free on a laptop.
Read essay →
Sovereignty as something you can buy.
What sovereignty has to mean if it is to survive a procurement contract — control, code, and exit, not residency alone.
Building things European institutions can use.
On the consortia, instruments, and procurement vehicles that determine whether European digital infrastructure becomes a usable reality.