Notes & Deep Dives
Build Skills, Not Agents: Why Banks Should Ration Agentic Autonomy
In regulated banking, the maturity signal is not how many agents you have. It is how deliberately you ration autonomy — and how well-governed the skills beneath it are.
Read article →How AI Agents Improve Over Time in Global Financial Services
A practical guide to governed agent learning in regulated financial services — the flywheel, the five improvement levers, and why the model should not silently retrain on customer …
Read article →Locking Down Claude Code: A Practical Guide with Simple Examples
A layered security guide for Claude Code — permissions, OS-level sandboxing, runtime hooks, and enterprise managed settings explained with simple examples.
Read article →LLM Token Economics
How LLMs process tokens, why prompt caching cuts input costs by 90%, and why output tokens are always the biggest line item.
Read article →Skills vs MCP Servers: The Hidden Token Cost of Claude Code Extensions
MCP servers consume up to 50x more context than skills. Here's how each loads into memory, what it costs, and when to use which.
Read article →