Posts
AI-Amplified Engineering: The Multiplier Is Discipline, Not the Model
AI doesn't level anyone up. It amplifies whatever process it's dropped into, weaknesses included. Two years of my own delivery telemetry, and the framework behind it, make the case for what actually separates the top-quintile outcome from the average one.
Two Years of AI-Amplified Delivery: The Telemetry
Raw telemetry, not a highlight reel: two years of git and GitHub API data across 426 repositories, instrumented for a short series on what's actually changed in how I ship.
The Kiwi Problem: Why Mission-Critical AI Needs a Deterministic Control Plane
LLMs will confidently produce wrong answers at machine speed, and they won't tell you. The fix isn't a better model — it's a deterministic control plane that enforces correctness regardless of what the model decides.
The Framework That Can't Take the Load: Why Agentic AI Breaks SAFe at the Foundation
I've run SAFe on large-scale enterprise programs. I've also deployed five production agentic AI systems. They don't live in the same dimension.
The ChatGPT Moment for Cars — And Why Most Automakers Aren't Ready
Nvidia's Alpamayo shifts autonomous vehicles from pattern-matching to reasoning. For legacy automakers still navigating Frankenstein supplier architectures, the window to act is open — but not forever.