Community & learning / WORKING NOTE

Learn in small loops. Share what changes.

Learning a new technology is easier to make concrete when there is a small question to answer. A useful experiment gives that question a boundary and leaves a trace for the next person.

3 min read · Open to revision

Pick a question small enough to answer

“Learn agentic systems” is an area of interest. “What happens when this tool call fails?” is an experiment. A narrow question makes it possible to compare an expectation with an observed result. It also gives you a natural stopping point.

Build the smallest observable example

Remove the parts that do not help answer the question. Keep inputs inspectable and outputs easy to compare. When the experiment fails, preserve enough context to explain what happened without asking someone else to recreate your entire environment.

Write down what surprised you

A command history is not the same as an explanation. Record the initial assumption, what happened, and what you would change next. Include the limit of the observation: one example does not establish a general rule.

Make it useful to another person

Share a small reproduction, an annotated diagram, or a short note with a clear question. Invite someone to try a different input or challenge the conclusion. That exchange turns an individual learning exercise into a resource that can improve over time.

Find a thread

Search across projects, notes, and ideas.