Technical Articles & Guides
In-depth thinking on software delivery, team design, and the engineering decisions behind AI-assisted work.
Clash of Giants? Why the US–China AI Race Is a Split, Not a Collision
Chinese models are close on benchmarks, but enterprise adoption, safety rules, talent flows, and chips tell a different story: two ecosystems winning on different ground. The 2026 evidence area by area, and a decision rule for choosing a model ecosystem.
Stepping Off the Dashboard: The Quiet Push Against Metric-Driven Engineering
Big tech is tightening monitoring and ranking, and AI speeds it up. A counter-movement is forming: companies that refuse to rank individuals, and engineers who leave to be judged by customers. Three pressures push people out. What the data shows, what it doesn't, and why you still have to measure.
Why Agentic AI Isn't Paying Off in Business Processes: The Accountability Gap
Coding agents and personal assistants work because a human watches and accepts the risk. Business processes need someone accountable for every action, and making an agent reproducible, auditable, and safe often costs more than it saves. The four walls, and the design that flips the ROI.
Will Science Become Wizardry? How AI Is Speeding Up Discovery, and Now Verification Too
AI has already made scientific discovery dramatically faster. Now AI agents and lab robots are speeding up verification too. Examples industry by industry, what still won't compress, and who stands to gain.
AI Layoffs Aren't Paying Off: What Gartner's 2026 Data Says Before You Cut
Cutting headcount didn’t separate high-ROI AI adopters from the rest, just 3% of organizations exceed cost-savings goals via AI productivity, and Gartner expects 30% of AI-replaced roles to be refilled by 2029. What to check before any AI-attributed cut.
Code Is Cheap, Judgment Is Scarce: Engineering Roles in the AI Era
AI moved the software delivery bottleneck from writing code to trusting code. Learn how modern engineering teams automate routine checks, tier review by risk, and cultivate real engineering judgment.
Redesign the Work Before You Reduce the Team
DORA's 2025 research links AI adoption with higher delivery throughput and product performance, but also greater instability. Measure the full workflow before using individual productivity gains to size a team.
Counting the Wrong Thing: Why Developer Performance Metrics Break Down When AI Changes the Unit of Work
Velocity, PR count, and lines of code were built to measure a comparable unit of role-scoped work in a queue. AI breaks that two different ways—and DORA's own 2025 data already shows the first failure happening at industry scale.
Rethinking the Sprint: Why Small-Batch Task Decomposition Breaks Down for AI-Native Builders
Sprints chop work into tickets small enough for a queue of specialists to hand off between them. When one accountable builder ships a feature in days, most of that structure stops adding value—this works through one feature run both ways to see what to keep, cut, or replace.