How to Evaluate AI Output at Scale Without Reading 10,000 Responses
Reading more outputs isn't evaluation. How to build an LLM judge that routes human attention instead of trying to replace it.
Read →Writing
On building AI products that actually ship.
AI architecture, startup technical decisions, and what it takes to go from demo to 2am reliability.
Why Your AI Agent Passes All Tests and Fails at 2am
The gap between demo reliability and production reality — and what to actually do about it.
Read →