Posts
All the articles I've posted.
The Cost of Context: Measure the Work, Not Just the Window
Updated: at 06:00 AMCompare long context and retrieval using actual token categories, latency and task quality. A larger window is capacity; its value depends on the workload.
The Model Updated. Now What?
Published: at 10:00 AMOpenAI pushed an update. Your prompts stopped working. Your evals are failing. Your users are complaining. Here's the playbook for surviving model updates—because this is your new normal.
The Invisible Technical Debt of AI Systems
Published: at 10:00 AMYour AI system works today. It worked yesterday. It will probably work tomorrow. But you're accumulating debt you can't see, can't measure, and won't notice until it's too late. Here's the taxonomy of invisible AI debt.