Blog
Notes from the boundary.
Benchmarking LLM Structured Outputs
Three Providers, Three Symmetric Ways They Break
Read →
The Agentic Bottleneck
Why We Need to Rethink CI
Read →
Why Coding Agents Are Getting More Expensive
Prompt Caching, Idle Sessions, and the Real Cost of a Million-Token Window
Read →
The Multi-Repository TypeScript Problem
Navigating Type Safety Across Service Boundaries
Read →