Blog launched with two essays — on agent memory, and on shipping LLM agents in clinical genomics

The blog is live with two essays: I measured memories people couldn’t report. Now I build memory for AI agents — what eye-tracking research on human episodic memory taught me about building and evaluating memory for agentic systems — and Lessons from shipping LLM agents where wrong answers matter — evals as the product, structure over eloquence, and designing the abstention path first.