Archive
Decision Notes
Decision Notes capture tradeoffs: what to automate, what to leave human, what to run locally, what to send to the cloud, and how to choose tools without creating more operational drag.
What Background Jobs Taught Me About Reliable AI Agents
Moving 18 scheduled AI jobs from cloud to local looks like a cost exercise. It is really a reliability audit:…
AI Agent Memory Needs the Same Discipline as Production Code
The failure mode nobody warns you about: AI agent memory that works for a week, then silently drifts. Memory architecture…
Local vs. Cloud AI: What Belongs Where in an Agentic System
The local vs. cloud question is not about cost. It is about which workloads you trust to run unattended, data…
What AI Agents Should Remember — and What They Should Never Store
Credential discipline and memory architecture for a self-hosted AI agent. The practical rules that prevent security nightmares and memory drift.