Live

What an agent is and how it differs from a chat

Before installing anything, one single idea. If you hold on to this one, the rest of the session makes sense on its own.

The chat answers you. The agent acts.

When you use an AI chat, the deal is this: you write, the model answers with text, and you do something with that text. You copy the code, paste it into a file, run it, go back to the chat with the error. The model never touches your computer.

An agent uses the same kind of model, but with one difference: it has tools. It can read files, write files, and run commands on your machine. So the deal changes: you describe what you want, and the agent does the work of copying, pasting, and running.

ChatAgent
What it receivesWhat you paste into itWhat it reads from your folder by itself
What it deliversText for you to useFiles created or modified
Who runs thingsYouIt does, with your permission
The riskThat it gives you a bad answerThat it makes a bad change

The difference isn't the model. It's that this one can read and write files.

The agent loop

When you give it a job, an agent repeats a loop until it's done:

  1. It thinks about what it needs to do.
  2. It uses a tool: reads a file, searches for something, writes a change.
  3. It looks at the result of that tool.
  4. It thinks again. If it's done, it tells you; if not, it repeats.

That's why a single job can turn into several requests to the model, and that's why the free plan is measured in requests.

Why this matters to you

If the agent can write files, your job changes from "writing" to "asking well and reviewing". That's the muscle we train today:

  • Asking well: describe the what (what you want to achieve), not the how.
  • Reviewing: look at what the agent changed before accepting it, and check at least one fact yourself.

Today you'll do it with OpenCode, which is open source and runs on any laptop. What you learn carries over to any other terminal agent.

Summary

  • A chat gives you text. An agent reads, writes, and runs things on your machine.
  • The agent works in a loop: it thinks, uses a tool, looks at the result, repeats.
  • Your job becomes asking well and reviewing.