AI

My first instinct, whenever I’m using an AI coding assistant, is to follow along with every step. Guide it, redirect it when it drifts, catch mistakes before they compound. It feels responsible. But it also puts the human squarely inside the AI’s feedback loop — which defeats a lot of the point. Recently, a few colleagues and I started an experiment to find out what happens when you deliberately remove that tight human-in-the-loop. The result is drellabot, a lights-out software factory. ...