What Agentic Coding Actually Is

There is a real difference between chatting with a model and handing a task to an agent. A chat assistant answers the question in front of it and stops. A coding agent runs a loop: it reads your project, decides on an action, runs that action, looks at the result, and then decides what to do next. It keeps going until the job is done or it gets stuck.

That loop is the whole point. People often call it the agentic loop: gather context, take an action, verify the result, repeat. When you understand the loop, most of the advice in this course stops feeling like a list of tricks and starts feeling obvious.

The agentic loop

1. Gather context: read the relevant files, run a search, check the failing test.
2. Take action: edit a file, run a command, call a tool.
3. Verify: run the test, read the output, check the diff against the goal.
4. Repeat: use what it learned to decide the next action, or stop when done.

Chat assistant vs coding agent

A simple first session

Notice the agent did not need a clever prompt. It needed a clear goal and a way to confirm success. That is the foundation everything else is built on.