Writing
Essays on applied AI, analytics, and what actually ships.
-
My Wife and I Pay the Same for AI ...but Should We?
She opens ChatGPT twice a day. I burn tokens through agents all day. We pay roughly the same, and I keep wondering who is covering the bill, and what it will feel like when they stop.
-
We Are All Just Predicting Tokens
The sage-engineer dismissal of LLMs as 'just autocomplete' looks like a technical claim, but it's a definitional and social one that fails against humans too.
-
The Three Pillars That Enable My Long-Running Agents
Long-horizon agents fail at three quiet disciplines underneath them that you must master: blueprint, context hygiene, and back-pressure.