Skip to content

Knowledge base

Effective beginner practices

Beginners get good with coding agents by practicing in a domain they can judge, keeping tasks and feedback loops small, and granting the agent more autonomy only once reviewing every small step turns into rubber-stamping.

Written by

Updated:

Getting Started 2

Judgment comes first, because a beginner who cannot tell right output from wrong learns from neither. A domain you already master lets you judge the output instead of trusting it blind (When AI writes almost all code (opens in a new tab)). Sandbox the environment before you widen the loop, so the extra autonomy is safe (Vom Coder zum Dirigenten (opens in a new tab)).

Proficiency grows with deliberate practice, not exposure. The people who get the most out of the tools treat a model like a musical instrument and keep playing after the first wrong note (deliberate intentional practice (opens in a new tab)). A daily habit works: an hour of real work with the best model available, not toy questions on a faster default (Something Big Is Happening (opens in a new tab)).

The tool makes it easy to stop thinking, so effective learners put the difficulty back. Letting the agent do all the thinking is how you learn least, and durable skill needs the productive struggle of effortful work (WEF – New Economy Skills: Unlocking the Human Advantage (opens in a new tab)). Practitioners remap the accept-suggestion hotkey or disable the agent’s file-write tool to keep critical thinking alive (building scalable apps with AI (opens in a new tab)). Reviewing is a different muscle from writing: lean only on evaluation and you slowly lose the capacity to create (AI’s impact on the state of the art in software engineering (opens in a new tab)).

Plan with the agent before it codes, describe tasks concretely and start new sessions often, because performance degrades as context grows. When you lose track of what the agent is doing, revert and approach the task again (The Pragmatic Engineer - Learnings from two years of using AI tools for software engineering (opens in a new tab)).

Start here

  • Deliberate practice: How beginners build judgment with familiar domains, tight feedback loops, active recall, and deliberate friction.

Resources

Articles

Papers and reports

Talks and videos

In this section

  • Deliberate practice

    Deliberate practice builds judgment with coding agents: start in a domain you can check, keep tasks and feedback loops small, practice daily on real work, and put friction back where the tool makes thinking optional.