INSIGHTS

Clear thinking.
Practical AI.

A few principles behind my approach to AI, development, and knowledge systems.

01 / MODEL STRATEGY

Every task doesn’t need your most powerful model.

Begin with what the task requires. A simple, repeatable step may suit a smaller model; a difficult planning decision may justify a more capable one. Evaluate the result, the cost, and the time together.

02 / DEVELOPMENT

Use AI to explore. Use code where the rules are clear.

AI can help test an idea before every detail is known. As a workflow settles, explicit code can take over predictable steps. Keep judgement with AI where it adds value, and give routine operations a stable structure.

03 / RELIABILITY

A process you can inspect is a process you can improve.

Record the inputs, decisions, tool calls, and outcomes that matter. When something fails, that record helps you and the AI locate the problem and restore the workflow with less guesswork.

04 / DOCUMENTATION

Better knowledge starts before the prompt.

Reliable sources, consistent terminology, and a clear document structure make information easier to retrieve and interpret. Treat the quality of the knowledge as part of the quality of the AI system.

Clear thinking.Practical AI.

Every task doesn’t need your most powerful model.

Use AI to explore. Use code where the rules are clear.

A process you can inspect is a process you can improve.

Better knowledge starts before the prompt.

Clear thinking.
Practical AI.