Column Title Two

Bias and Prejudice in AI Search

When Claude Met DEJAN I was helping a developer debug a machine learning pipeline. Forty million training samples, weighted loss functions, checkpoint management — technical work. At some point, they asked me to generate test queries for their keyphrase volume classifier. I needed examples across the search volume spectrum, from high-volume head terms down to […]

Most People Don’t Read

This is a qualitative study on a small number of anonymized users while collecting a very large number of datapoints from each one. In December 2025, we published an article asking a simple question: Do you read or skim? We tracked 269 visitors using mouse movements, scroll patterns, and time-on-page data, then asked them to […]