Tools & Insights for Building
with AI
Practical guides, open-source utilities, and experiments in AI & software engineering - created and tested by me, Stephen Collins.
Featured Posts
A Code-Risk Score That Failed Calibration — and the Narrower Claim That Held Up 6-7x
Stephen Collins
•
Nov 12, 2026
7 mins
A risk score that doesn't calibrate as a probability can still be a real signal in a narrower, more honest form — here's how ...
Read More
Everyone Agrees Code Review Is the Bottleneck. Almost No One Shows Their Math.
Stephen Collins
•
Nov 12, 2026
7 mins
The claim that AI broke code review isn't new or contested anymore — it's the consensus take of 2026. What's still missing is...
Read More
3 Hotspots Analyses Worth Reading This Week
Stephen Collins
•
Oct 21, 2026
4 mins
A weekly pick of the most interesting Hotspots analyses — the ones where the code, not just the score, is worth a look....
Read More