Insights on agentic AI, governance, and building production-grade systems for regulated industries.
The EU AI Act is an operational maturity test. For high-risk systems, most obligations are engineering obligations. Compliance is architecture.
An open-source, AI-assisted operating system for senior delivery managers: ten agents, fifty-two workflows, local-first, audit-by-default.
If a Tuesday conformity assessment says little about Wednesday, the static model breaks down for orchestrated AI. A continuous-conformity proposal.
A .npmignore mistake shipped Claude Code's source to npm. The leaked codebase reveals practices that should inform AI toolchain due diligence.
The fractional AI leadership market has two models: brokers deploying generalists, and practitioners making architectural decisions. Six tells.
Generic AI leadership optimises for the wrong thing in regulated industries. Three scenarios show why defensibility, not accuracy, is the goal.
The EU AI Act attaches obligations to architectural decisions that used to be purely technical. A Head of AI needs to know where the lines are.
Open-source static analysis that scans your codebase for AI framework usage and validates risk classifications against the EU AI Act. Snyk for AI.
The barrier to building bespoke legal AI has collapsed. The EU AI Act's obligations have not. Every vibe-coded tool is potentially high-risk.
The Omnibus proposal could delay high-risk AI Act obligations by 18 months. Or not. Which deadline should engineering teams plan for?