Build and Fix Things
How do I find what is really wrong, and build things that keep working?
Five steps from the ideal of the philosopher-engineer to troubleshooting by evidence, honest arithmetic, systems thinking and a disciplined way to work with AI coding agents.
-
The Philosopher-Engineer: Study and Workshop Together Liberal Arts & Reasoning
Start with the ideal: the working habits shared by people who both think hard and build well, as a model any learner can adopt.
-
Evidence-Based Troubleshooting: Separating Explanations Engineering & Career
Then the core skill: ask what an observation actually establishes, and choose the next test that splits the competing causes.
-
Quantitative Reasoning with Stated Assumptions Engineering & Career
Put numbers on it honestly: worked examples in rates, packet sizes, recovery time and costs show that arithmetic is only as good as the assumptions you state.
-
Systems Thinking: Feedback, Queues, Information and Dynamics Mathematics, Systems & Languages
Most stubborn problems are systems: stocks, flows, feedback loops and delays explain why fixes overshoot and queues explode.
-
Working with AI Coding Agents: A Disciplined Loop Engineering & Career
Finish with today's tools: a workflow for coding agents that puts the spec and plan first, reviews every diff, verifies, and commits known-good states.