Topics
Posts grouped by topic, oldest first so each topic reads as a series. A post can sit in more than one topic.
AI platform engineering
8 postsShared platforms for teams building with models: what to standardise, what to expose, and how to tell if it's working.
Latest: Good abstractions hide complexity, not evidence, 15 Oct 2026
Identity and security
3 postsWho is calling, what they may do, and where that decision should be made.
Latest: Patterns for agent tool calls you can trust in production, 19 Oct 2026
Observability
2 postsThe evidence a team needs to locate a failure without filing a ticket.
Latest: Good abstractions hide complexity, not evidence, 15 Oct 2026
Developer experience
5 postsHow hard it is for another team to get from zero to a safe production workload.
Latest: Measure the journey to the first production call, 8 Oct 2026
AI agents
5 postsRunning agents in production: tools, delegation and the boundaries around them.
Latest: Patterns for agent tool calls you can trust in production, 19 Oct 2026
Reliability
2 postsTimeouts, retries and partial failures, and how to make them safe.
Latest: Patterns for agent tool calls you can trust in production, 19 Oct 2026