A practitioner's notebook.Written at midnight.When curiosity wins over sleep.

Welcome to my cabinet of curiosities. Take everything with a grain of salt — or maybe the whole shaker.

Latest Findings

I spent two weeks going deep on how Large Language Models actually work. This is what I found — compressed into something worth coming back to.

A comparative synthesis of McKinsey, BCG, Deloitte, Accenture and PwC — what actually works, and what doesn't.

Two types of prompts, eight practical reflexes, and one loop that makes your prompts better over time. A field guide for non-technical operators.