Start Here
Your on-ramp to working with AI agents
Get oriented, set up your tools, and complete a real task with an AI agent in your first 15 minutes. No experience required.
- Lessons
- 5
- Reading time
- 41 min
- Level
- Beginner
- Free lessons
- 5
Lessons in this module
- 0.15 minFree
Welcome: What You'll Be Able to Do
How this course is structured, who it's for, and the concrete skills you'll walk away with — from delegating work to agents to building your own.
- 0.29 minFree
What Is an AI Agent, Really?
A no-hype definition of AI agents: a model, a loop, and tools. Learn how agents differ from chatbots and why they suddenly got useful.
- 0.311 minFree
Setting Up Your Agent Toolkit
Install and configure the essentials — a chat assistant, a coding agent, a terminal, and an API key — so you're ready for every lesson.
- 0.48 minFree
Your First Agent Task in 15 Minutes
Hands-on: hand an agent a real multi-step task, watch it plan, use tools, and verify its own work — then learn what just happened.
- 0.58 minFree
The Agent Builder Mindset
The mental shifts that separate people who get great results from agents from those who get frustrated: delegation, verification, and iteration.
About this module
An AI agent is a language model that works in a loop. It reads a goal, decides on an action, uses a tool such as reading a file, running a command, or searching the web, looks at the result, and repeats until the job is done. That loop is what separates an agent from a chatbot. A chatbot answers one message at a time and waits for you to do the next step; an agent carries out the steps itself and checks its own progress along the way.
Agents became practical for everyday work fairly recently. Models got much better at choosing and using tools, context windows grew large enough to hold real projects, and standards for connecting tools made it easy to plug agents into files, code, and services. Coding agents were the first place this clearly paid off, but the same pattern now shows up in research, data analysis, and office automation.
This part of the course is the on-ramp. You will get a plain-English definition of an AI agent and a quick test for telling real agents from relabeled chatbots. You will install a working toolkit on macOS, Windows, or Linux: a chat assistant, a terminal, git, Node.js, a coding agent such as Claude Code or Codex, and an Anthropic API key. Then you will hand an agent a real multi-step task, review its plan, approve its actions, and check the result yourself.
The last lesson covers the habits that make the difference between frustration and good results: delegating outcomes rather than keystrokes, verifying in proportion to the stakes, and improving your instructions instead of patching output by hand. No programming experience is required, and every lesson in this part is free to read.
Questions about Start Here
What is an AI agent in simple terms?
An AI agent is a language model that runs in a loop and uses tools to reach a goal. It decides on a step, takes it with a tool, looks at the result, and keeps going until the task is done or it needs your input.
What is the difference between an AI agent and a chatbot?
A chatbot responds to one message at a time, and you carry out the next step yourself. An agent takes the steps itself: it can read files, run commands, or call services, check the results, and continue without you prompting each move.
Do I need to know how to code to use AI agents?
No. Many agents, including chat assistants with research and file tools, work entirely in plain language. Coding agents run in a terminal, but you direct them in plain English, and knowing a handful of terminal commands is enough to start.
What do I need to install to start using AI agents?
For this course you need a chat assistant, a terminal, git, Node.js, and a coding agent such as Claude Code, Codex, or Cursor. Building your own agents later also requires an API key from a model provider. The setup lesson walks through each step for macOS, Windows, and Linux.
Are AI agents safe to use on my computer?
They can be used safely with sensible habits. Most agents ask for approval before risky actions like running commands or editing files, and keeping your work in git means you can undo changes. Start with low-stakes tasks, read the plan before approving it, and check the results.