methodology
4 posts tagged methodology.
Does My Bug-Risk Model Still Work Once an AI Is Committing Code?
Churn, ownership concentration, and code coupling are old proxies, built from decades of human development patterns. An LLM doesn't get tired and doesn't lose context between commits the way a person does. I found Next.js — a repo with real, large-scale, ground-truth AI-assisted commit history — and checked whether my existing signals still know what they're looking at.
I Tested a New Risk Formula Against Real CVEs. The First Result Was Too Good to Be True.
A five-signal formula beat my production risk score on 20 of 21 repos when I checked it against real CVEs. The real result, once I found the bug in my own comparison, was 9 of 10 — still a win, for a completely different reason.
Why I let AI train on this
I just flipped ai-train=yes in robots.txt. Here's the reasoning, not just the setting.
How hotspots measures itself
A ranking tool that can't measure its own accuracy is just vibes with syntax highlighting. Here's how hotspots keeps itself honest.