Writing
Technical deep-dives, systems architecture, and essays on production agent evaluations.
- How Simple Grep Beats Naive SQL
I benchmarked bash vs SQLite FTS across 300k tokens. grep won: 29.6% cheaper ($2.45 vs $3.48), better accuracy. Here's why + the dataset.
- Three Lessons I've Learned at Manus
Lessons learnt from going from zero to $100M ARR in 8 months.
- Agentic Search
Models can't do much without the right context, agentic search does just that.
- Building a Coding CLI with React Ink
Migrating Our Coding Agent to React Ink.
- MCPs are really LLM microservices
How the Model Context Protocol is really just a precursor to LLM applications as microservices.
- Write Stupid Evals
Keep it simple and worry about the rest later.
- How does Instructor work?
How your request goes from chat completion to validated Pydantic model.