BEGINNER LEARNING TRACK

From zero to shipping
real code with AI

Start from absolute zero. Set up your environment, learn coding fundamentals with AI assistance, build real mini projects, and ship your first portfolio app with confidence.

NO EXPERIENCE REQUIREDPRACTICE TERMINAL BUILT INCOSTS & MODELS EXPLAINED
14 daysstructured path
5 stepsto first shipped app
0 experiencerequired to start
app.py — first project
# ask AI, then run it yourself
import requests

name = input("Your name: ")
r = requests.get("https://api.chucknorris.io/jokes/random")
joke = r.json()["value"]

print(f"Hi {name}! Here is a joke:")
print(joke)
$ python app.py✓ ran successfully — you just built your first app
Beginner friendly
Runs in minutes
QUICK START

Start here if you feel overwhelmed

Follow this exact first-hour sequence. Do not skip steps. This gives you momentum fast and keeps learning straightforward.

0-15 mins: Tool setup

Install an editor and Python with exact copy-paste commands for Windows, Mac, or Linux.

Jump to setup

15-35 mins: Drive the tools

Take the interactive editor tour, then practise real commands in the built-in sandbox terminal.

Try the practice terminal

35-60 mins: First real win

Walk through your first program line by line, tick the day-1 checklist, and check honest costs.

Understand your first program
14-DAY PLAN

Your beginner roadmap

Follow this progression to avoid overwhelm. Each phase is intentionally small, practical, and designed to build confidence through shipped outcomes.

Day 1-2

Environment setup

Install the editor, terminal tools, Python, and optional Rust. Learn how to run files and understand common setup errors.

  • Editor ready
  • Terminal confidence
  • First script runs
Day 3-5

Core coding habits

Use AI the right way: ask for explanations, run code often, read errors line by line, and make small edits in tight loops.

  • Prompt patterns
  • Debug workflow
  • Safe iteration
Day 6-9

Useful mini projects

Build tiny but real tools: input validator, API fetch script, file organizer, and a simple CLI helper you can reuse daily.

  • 4 mini projects
  • API basics
  • File and data handling
Day 10-14

Ship your first portfolio app

Plan, build, test, and polish one beginner app with AI support while understanding every moving part before moving on.

  • Portfolio-ready app
  • Project structure
  • Next learning path
HANDS-ON SETUP

Environment setup and first coding loop

This is your practical launchpad. Complete these steps in order and run each command yourself. Avoid copy-paste only workflows. The goal is understanding plus execution.

01
EDITOR

Install your editor

Start with Cursor or VS Code. Both work. Cursor is AI-first, VS Code has the biggest extension ecosystem.

  • Install Cursor or VS Code
  • Enable autosave and format on save
  • Learn 3 shortcuts: open file, terminal, command palette
02
PY

Install Python

Python is the best first language for AI-assisted learning because syntax is readable and tooling is simple.

  • Install from python.org
  • Tick the Add Python to PATH option
  • Verify with python --version
03
RS

Optional: install Rust

Rust is not required for day one, but it is excellent for performance and systems thinking once you are comfortable.

  • Install via rustup.rs
  • Verify with rustc --version
  • Create first project with cargo new hello_rust
04
AI

Optional: local AI with Ollama

Run models locally for private, no-API-cost experimentation while learning prompt and debugging workflows.

  • Install from ollama.com
  • Run ollama run llama3
  • Use it as your local coding explainer
05
BUILD

Your first app workflow

Open your editor, create app.py, and ask AI for a tiny script with user input plus one API call. Then run, inspect, and improve it in three small iterations.

Prompt to try: Create a beginner Python script that asks for my name, fetches one safe joke from a public API, and prints a friendly message. Explain each line in plain language after the code.

Iteration 1

Run the script and confirm it executes successfully.

Iteration 2

Add error handling for failed network requests.

Iteration 3

Refactor into small functions and rename unclear variables.

Exact commands for your computer

We detected your system — but you can switch. Copy each command into your terminal, press Enter, read the reply.

Where to type: PowerShell (built in — search 'PowerShell' in Start, or Ctrl+` in VS Code)
  1. 1. Install Python (official installer — TICK 'Add python.exe to PATH')winget install Python.Python.3.12
  2. 2. Close and reopen the terminal, then verify Python answerspython --version
  3. 3. Make a project folder and enter itmkdir my-first-project; cd my-first-project
  4. 4. Open that folder in VS Code from the terminalcode .
TRY IT RIGHT NOW

Practice terminal — no install needed

The terminal is the scariest part for most beginners, so practise here first. This sandbox simulates your first session — including the classic first error and how to fix it.

practice terminal — my-first-project
Practice terminal — safe sandbox, nothing touches your real computer.
Try: python --version · dir · python app.py · pip install requests · help
$

Your mission (2 minutes)

  1. Check Python:
  2. Look around:
  3. Run the script — watch it fail:
  4. Fix it:
  5. Run again — success:

That fail → read → fix → rerun loop is the single most important skill in programming. You just did it before installing anything.

DRIVING LESSONS

Learn to operate VS Code (and Cursor)

This is the part most guides skip. Click each area of the editor below to learn what it does, the shortcut that opens it, and how a beginner should actually use it.

my-first-project — Visual Studio Code

Explorer — your project's file tree

The left sidebar shows every file and folder in the project you opened. You always open a FOLDER (File → Open Folder), not individual files — that's what makes the terminal, AI context, and search all work from the right place.

  • Ctrl+Shift+E jumps to the Explorer
  • Right-click → New File to create app.py
  • The top-level folder = your project root — keep one project per folder
UNDERSTAND, DON'T MEMORISE

Your first program, explained line by line

Six lines that take input, call a live API, and answer back. Click each line — if you understand these six ideas, you understand the skeleton of most beginner programs.

app.py — click any line
Line 1import requests

Borrows a toolbox. 'requests' is a library — code someone else wrote — that knows how to talk to websites. Importing it means 'let me use those tools in this file'. You installed it once with pip install requests.

DAY-1 CHECKLIST

Track your setup — it saves as you go

Tick each item off as you complete it. Progress is stored in your browser, so you can leave and come back. Finish all eight and you are genuinely ready to build.

0/8 complete
MODELS EXPLAINED

Which AI model should you actually use?

Every tool advertises different models — Claude, GPT, Gemini, Llama. As a beginner the differences matter less than the marketing suggests. Here is the honest map.

Claude (Anthropic)

Sonnet & Opus tiers

Explains code patiently, strong at long files and careful step-by-step reasoning. Powers Claude Code and many editor agents.

For beginners: Great default for 'explain this to me like I'm new' — answers teach rather than dump code.
claude.ai free tier · Pro subscription · inside Cursor/Copilot

GPT (OpenAI)

GPT-5.x family

Fast, versatile all-rounder with the biggest ecosystem of tutorials, and built into ChatGPT which you may already use.

For beginners: If you already chat with ChatGPT, staying here removes one new tool from your plate.
chatgpt.com free tier · Plus subscription · inside Copilot

Gemini (Google)

Gemini Pro & Flash tiers

Generous free access through AI Studio, huge context windows for reading long documents and codebases.

For beginners: The most generous free tier for experimenting without a card on file.
aistudio.google.com free · Gemini app · inside Cursor

Local models (Ollama)

Llama, Qwen, DeepSeek & more

Run entirely on your own machine: private, offline, zero per-use cost. Quality is below frontier cloud models but improving fast.

For beginners: Perfect judgement-free practice partner — ask 'dumb' questions all day for free.
ollama.com — free, needs ~8GB+ RAM for small models
The rule that matters: pick ONE assistant and use it for two weeks. Model-hopping is procrastination wearing a productivity costume. Every frontier model is more than good enough for beginner Python — your workflow (run, read, fix, repeat) is what makes you improve.
COSTS & SUBSCRIPTIONS

What this actually costs (spoiler: $0 to start)

The honest pricing map, from completely free to the first upgrade actually worth paying for. You do not need a credit card for anything on day one.

Tool / planCostWhat you getWhen it fits
VS Code + Copilot Free$0Full editor + limited AI completions & chats per monthDay-1 default — start here
Cursor Hobby$0AI-first editor with limited completions & agent requestsTrying the AI-native editor feel
ChatGPT / Claude / Gemini free tiers$0Browser chat for explanations, planning, debugging helpYour out-of-editor tutor
Ollama (local models)$0Unlimited private local AI — your hardware is the limitUnlimited practice, offline
GitHub Copilot Pro~$10/moUnlimited completions, more chat/agent usage, model pickerFirst paid upgrade for most beginners
Cursor Pro~$20/moExtended Tab autocomplete + agent usage poolIf Cursor's flow clicked for you
ChatGPT Plus / Claude Pro~$20/moPriority access, stronger models, higher limits in the chat appsHeavy chat-tutor users
Raw API keys (pay-as-you-go)per tokenYou pay per million tokens processed — powerful but meterless spending is realSkip until you ship apps that call AI

The sensible spending path

  1. Weeks 1-2 — spend $0. Free editor + one free chat assistant + optionally Ollama. Prices change; free tiers are the constant.
  2. Hit a real limit? If you run out of free completions mid-flow, Copilot Pro (~$10/mo) is the highest-value first upgrade.
  3. Loving the agent workflow? Cursor Pro or a chat subscription (~$20/mo) — pick the ONE you use daily, not all three.
  4. Avoid API keys for now. Pay-per-token has no monthly ceiling — a runaway loop can bill while you sleep. Subscriptions cap your downside.
CHECK YOURSELF

Quick knowledge check

Four questions covering the workflow, the editor, and the money. Answer honestly — explanations appear after each pick.

Q1 Your script fails with a long red traceback. What do you do first?

Q2 What does 'open a folder' (not a file) give you in VS Code?

Q3 What's the smartest day-1 budget for AI coding tools?

Q4 What is a 'token' in AI pricing?

AVOID THESE

Beginner mistakes to avoid

These are the exact patterns that slow down first-time learners. If you avoid these, your progress becomes dramatically faster.

CHECKPOINT

Copying full AI output without running tests after each change

CHECKPOINT

Trying to build a huge app before completing one tiny app end to end

CHECKPOINT

Skipping terminal practice and relying only on editor buttons

CHECKPOINT

Ignoring error messages instead of reading the first relevant line

CHECKPOINT

Changing 20 things at once so you cannot isolate what broke

SPEAK THE LANGUAGE

Jargon decoder — every word that confused you

Tutorials assume you know these. Nobody is born knowing them. Click any card — plain-language meaning plus a concrete example.

QUESTIONS

Beginner FAQ

Simple answers to common concerns before you begin your first proper coding cycle.

I have never coded before. Is this still for me?

Yes. The page is designed for true beginners. Start with setup, run one command at a time, and do not skip the mini projects.

Should I learn Python or Rust first?

Start with Python first, then add Rust later. Python gets you productive quickly while Rust teaches deeper systems skills after your basics are stable.

How do I use AI without becoming dependent on it?

Ask AI to explain and scaffold, but always run, edit, and debug yourself. Your skill grows from verifying outputs and understanding failures.

How long until I can build something useful?

Most beginners can build useful scripts in the first week if they keep scope small and practice daily in short focused sessions.

How much does it cost to start coding with AI?

Nothing. VS Code and Cursor have free tiers, GitHub Copilot has a free plan, and ChatGPT, Claude, and Gemini all offer free browser chat. A typical first paid upgrade is GitHub Copilot Pro at roughly $10/month, and only once you hit real limits.

Which AI model should a beginner use for coding?

Any frontier assistant works: Claude is excellent at patient explanations, GPT is a fast all-rounder, and Gemini has a generous free tier. If you want free unlimited private practice, run a local model with Ollama. Pick one, stick with it for two weeks, and focus on the workflow rather than the model.

Do I need Visual Studio or Visual Studio Code?

Visual Studio Code (VS Code) - the free, lightweight editor - is what you want as a beginner. Visual Studio is a separate, heavier IDE aimed mainly at .NET and C++ development on Windows. Cursor is a VS Code fork with AI built in, and everything you learn in one transfers to the other.

What are tokens and why do AI tools talk about them?

Models read and write text in tokens - chunks of roughly four characters. Subscriptions like Copilot or ChatGPT Plus hide tokens behind a flat monthly price, while raw API access bills per million tokens. As a beginner, stick to free tiers and subscriptions so costs stay predictable.

What is vibe coding and is it a real way to learn?

Vibe coding means describing what you want in plain language and letting AI write the code, then iterating by feel. It is a real and fast way to build - but to learn, you must run and read everything the AI produces. Use the loop: prompt, run, read, tweak, repeat. If you skip the reading step you accumulate code you cannot fix.

Do I need math to start coding with AI?

No. Beginner programming is logic and reading, not math. You need arithmetic at most. Math only becomes relevant later in specific fields like machine learning research, graphics, or scientific computing - and even then AI assistants help you through it.

What laptop or computer do I need to start coding?

Almost any computer from the last 8 years works. VS Code, Python, and cloud AI assistants are lightweight - 8GB RAM is comfortable. You only need more power for running local AI models with Ollama (8GB+ RAM for small models). Do not buy hardware to start; start with what you have.

How do I practice the terminal without breaking anything?

Use the practice terminal on this page first - it simulates the real first-session commands including the classic ModuleNotFoundError and its fix. On your real machine, commands like python --version, dir, ls, and cd are read-only and completely safe to experiment with.

Is GitHub Copilot free for beginners?

Yes - GitHub Copilot has a free plan with a monthly allowance of completions and chat messages, which is plenty for the first weeks. Students get Copilot Pro free through the GitHub Student Developer Pack. Cursor also has a free Hobby tier, and Claude, ChatGPT, and Gemini all offer free browser chat.

Next learning paths

After this beginner path, choose your direction: autonomous agents, prompt engineering, or practical free tools.