Articles
Technical notes
Reference material, not a blog. Written once, kept accurate.
Why your AI pilot didn't ship
The failure mode is almost never the model. It's everything downstream of it.
Retrieval quality is a data problem, not a model problem
A better prompt won't fix retrieval that's returning the wrong passage.
Running open models instead of paying per token
For narrow, high-volume tasks, a self-hosted open model is often cheaper and more predictable than an API call.
What a forecasting model can and cannot tell you
A forecast is a distribution, not a promise. Treat the point estimate as the least important number it produces.
Contact