Articles
2026
- Agentic Engineering Requires One More Shift LeftSecurity shifted left into CI a decade ago. The authoring moved into the agent harness, and the gates are following.
- Contribution Graphs Don't Matter AnymoreThe contribution graph's level is dead as a signal. The tempting replacement, the bend, dies too. What survives is the conversation.
- Hire Where Failure Is ExpensiveAgents made feature work cheap. The scarce skill you can screen for is judgment in the domains where a wrong model output costs an outage, a breach, or a regulator.
- The Coding Interview InvertedA proposal on how engineering leaders can modernize the hiring experience for agentic-first development.
- The Agentic Maturity Ladder"Do you use AI?" stopped discriminating a year ago. "Show me your loop" separates engineers across a six-level scale.
- Taskmaxxing > TokenmaxxingThe per-token price is a proxy. The real KPI is cost per completed task.
- Agent Readiness Is Engineering Discipline with a DeadlineThe checklist for agent-ready codebases is the same one we've ignored for twenty years. What changed is the cost of ignoring it.
2025
- Minimum Viable Rigor: The Pre-PMF Verification BudgetStartups burn runway on two equal errors: premature platform engineering, or sloppy duct tape that pays a 27 percent rework tax on every pivot.
- The One-Year Law: Why AI Broke DORA's Ten-Year InvariantFor ten years, software delivery operated under a proven invariant: throughput and stability move together. Generative AI broke it in a single year.
- Evals > VibesAgent skills are code now. Almost nobody tests them, and the few honest measurements should worry you.
- Feeling Is the MoatWhen AI makes every feature cheap to copy, what stays defensible in consumer software is how it makes people feel.
- The Stability Paradox: Why Priority Whiplash Destroys Agentic CapitalWhy roadmap churn liquidates deterministic verification capital, and how engineering leaders must report stability-adjusted yield.
2023
- Postmortem the Near MissAn engineering organization that only reviews customer-facing outages is running on luck. Unexamined close calls train teams to become more reckless by mistaking fortunate timing for architectural resilience.
- The Wave Carried ThemWhy hiring from the first wave of a hot industry is harder than it looks, and how to tell surfers from swimmers.
2022
- Supply Chain Security Without CI Is Security TheaterWhy bolting SLSA compliance onto manual builds slows delivery and fuels burnout, and why automated CI is the strict prerequisite for real security.
2021
- Documentation Is the Unpaid Debt of ReliabilityWhy internal documentation is the highest capability multiplier of operational uptime, and how to govern it as executable infrastructure.
- Build Proprietary, Buy CommodityEvery line of code is an ongoing operational liability. Building custom commodity software is no longer an engineering flex but a capital allocation error.
- MTTR > CABWhy Change Advisory Boards fail to prevent outages, and why optimizing for recovery speed beats pre-release gatekeeping.
2020
- Disaster Recovery Testing Is Not Chaos EngineeringFault injection tests whether software fallbacks run. Surviving a real catastrophe tests whether engineers can coordinate under ambiguity.
2019
- Centers of Excellence Are Where Transformations Go to DieWhy centralized transformation offices become bureaucratic bottlenecks, and how elite teams scale practices through community networks.
No articles match that title.