Field notes on AI in real software.
Plain-language writing on AI-native engineering, legacy modernization, and what we're learning shipping software in 2026. No hype. Subscribe via RSS →
2026
6 notesSpending more was never the AI strategy
For two years the answer to almost every AI question was to buy the biggest model and approve the budget. That era is closing. Five practices for teams that now have to get the same result for less, and the one number worth reporting.
AI adoption stalls when the team has no memory
Most teams hand out AI coding licenses, call it a transformation, and plateau at thirty percent adoption six months later. The tools are not the problem. The team has nowhere to keep what it learns.
Your AI specialist hire will not close the gap
One agentic engineer can now do the work of several. That is exactly why hiring an AI specialist into a system nobody designed concentrates risk instead of closing your strategy gap. Four things to build before the hire.
Six ways to tell an AI bet from an AI pilot
Most AI pilots don't fail loudly. They run alongside the old process until someone quietly stops renewing the licenses. Six structural differences between a pilot that dies and an org-level bet that compounds.
Three rules for adding AI to a codebase that already ships
Adding AI capabilities to a system already running your business is a different discipline than building a greenfield AI app. Three rules we keep coming back to.
AI-native engineering, defined
The phrase gets used a lot. Here's the working definition we use at CBSI and the test we apply to know whether a team has actually adopted it.