blog
Release with trust, one idea at a time.
Essays and research notes on measuring how teams really ship, with humans and AI agents on the same ruler.
Two Ways to Measure AI Adoption, Both Broken
Peer stigma and token surveillance both break AI adoption numbers. What a trustworthy, team-level way to measure AI at work looks like.
Measure the Team, Not the Model
Coding agents excel when a human steers them. Discover why measuring the human plus agent pair, not the model, is how teams build trust in AI.
Fast to Generate, Slow to Trust: The 1:24 Ratio
One AI rewrite took 7 hours to generate and a week to verify. What the 1:24 ratio reveals about trust and where engineering teams should invest.
Code Review Is Not Dead. It Is Becoming Governance.
Why review layers survive AI speed. What data from 10,000 developers and the 75% evolvability finding reveal about code review in the age of AI agents.
The Trust Gap Is the Governance Gap
84% of developers use AI coding tools. Only 33% trust the output. The trust gap now has a measurable number, and measuring it is how teams close it.
The AI Adoption Spectrum and the 6x Gap
OpenAI data from 9,000 workers shows a 6x productivity gap. It is a spectrum to climb, not a binary to flip. Here is what that means for your team.
The Benchmark Paradox in AI Code Review
Vendor benchmarks report 60% F1. Academic tests show 19%. What the 3x gap means for teams measuring AI code review effectiveness.
Measuring AI in Software Development
What a randomized trial and data from 600+ organizations reveal about measuring the real impact of AI on software teams.
See what trust built with data looks like on your team.
A 30-minute demo on your own repositories. No guesswork, just your team on one ruler.
Schedule a demo