A practitioner journal

Sanity Check is what I am thinking out loud about, before the consensus catches up.

Notes from the edge of agentic adoption, written by someone still figuring it out. Plain prose. Specific examples. The same questions I am asking my own AI agent every week.

Hop on this train of thought Read the archive

A pencil sketch about sanity-checking your data work

The latest three

Most recent issues

Apr 23, 2026 · 5 min read

The protocols nobody calls the protocols

Three independent specs landed in the last 12 months. They share a shape — and the shape is what's worth betting on, not the specs.

protocolsagentic-ai

Feb 24, 2026 · 3 min read

I stopped writing about data. Here's what brought me back.

I published my last Sanity Check in October 2023. Then I went quiet.

Oct 21, 2023 · 2 min read

SC 021 — The Advantages of Going Deep on Architecture

Three reasons deep analytics architecture knowledge pays off: (1) faster problem resolution during outages, (2) identifying optimization opportunities in mature systems without full rebuilds, and (…

All issues, chronological →