Writing
Where we show the work — the build logs behind the systems we teach, and the evidence behind every claim we make.
Cruse Control is built by practitioners, not theorists. This is where we publish what that actually looks like: build logs from the substrate we’re standing up, field notes from the calibration calls we run, and the citations underneath the numbers on this site.
Longer essays are publishing soon. We’d rather ship them late than fake a backlog. Until they’re ready, the most useful thing here is the evidence below — the real sources behind the L-frame and the numbers on this site.
- 01 · METR · 2025
Measuring the impact of early-2025 AI on experienced developers
A randomized controlled study found experienced open-source developers were 19% slower when using AI tools — despite believing they were faster. The gap between felt speed and measured speed is the reason we calibrate cold before we deploy.
Read the METR study → - 02 · The Web Strategist (Jeremiah Owyang) · 2025
AI-native startups are dominating traditional software on one metric
Leading AI-native startups report roughly $3.5M in revenue per employee, against a ~$611K per-employee baseline for leading SaaS firms. The spread is the prize — and it only shows up for teams whose people can actually operate the tools.
Read Owyang's analysis → - 03 · Epoch AI
Epoch AI — trends in machine learning, compute, and capability
Independent research tracking the trajectory of AI: training compute, model scale, and the pace of capability gains. We lean on Epoch's data when we talk about where the curve is heading rather than where it feels like it is.
Visit epoch.ai →
The honest reading usually surprises people. The only way to know where your team actually operates is to measure it cold.
Email us →