ArclabLearn To Code In The AI Era.
AI writes code in seconds. Knowing whether it's right is the skill now. Foundations teaches Python, Java and Next.js the way that builds it: predict what code does, write it yourself, run it, then judge what AI wrote. In a bootcamp? Enter your cohort's join code.
Predict. Write. Run. Earn The AI.
Predict before you run
Every lesson starts with one idea and a question: what does this code do? You answer first, in your own words.
Write it and run it
main.py runs real Python inside your browser. Type python main.py in the terminal and see what happens.
Get checked on what ran
Your answer is checked against what the code actually did, then the next task opens: modify, explain back or find the bug.
Earn the AI
AI help opens in stages, and only after the earlier lessons are done. The rule is enforced on the server, so nobody skips it.
This Is What You Learn In.
A dark editor that works like VS Code, a lesson panel, a terminal, and a mentor that waits its turn.
Lesson 01 Is Waiting.
Make the machine do something, before the AI does. Nothing to install.
Start learning →AI Unlocks One Stage At A Time.
AI can write code in seconds, so the part worth learning is judging it. Foundations holds AI back until you've built the basics yourself, then opens it in four stages. Here's the Python track.
- 1No AIPython modules 1 to 7
You write every line yourself. Hints are there when you're stuck, answers are not.
- 2AI explainsmodules 8 and 9
Ask the mentor what a concept or an error means. It explains; it won't write the code for you.
- 3AI comparesmodules 10 to 13
Once your code passes, open an AI version beside yours and work out what it did better, worse or wrong.
- 4AI collaboratesmodules 14 to 16
Build with AI the way the job works now, then read and review everything it wrote.
Java opens AI at module 14 and Next.js at module 8. Task help opens only after you've made an attempt.
More Than Typing Code.
def add(a, b): return a + b x = add(2, 3) x holds 5. add() hands back a + b.
Predict. Read code and say what it will do before you run it. 83 of these in Python alone.
def area(width, height): return width * height room = area(4, 5) print(f"{room} square meters") 20 square meters
Modify. Change working code so it does something new, then run it to see if you were right.
# What's the difference between a # function that prints and one # that returns? print only shows it. return gives it back so I can use it.
Explain back. Say in your own words why it works. Gate tasks like this open the next lesson.
user_input = "wrong" if user_input = "admin": print("Welcome, admin!") else: print("Access denied")
Find the bug. A loop that never ends, = where == belongs. Find it, fix it, say why it broke.
total = bill + tip_amount each = total / people print(f"Each pays ${each:.2f}") Each pays $23.50
Build. A bill splitter, an expense tracker, your year wrapped. Projects you make and keep.
# yours best = nums[0] for n in nums: ... # AI draft print(max(nums)) # max([]) raises ValueError
Compare. Later modules put an AI draft beside your passing code. Spot what it got right and wrong.
Lesson 08, Task 1.
Predict the output before you run it.
nums = [3, 1, 2] print(sorted(nums)[-1])
Every Module, And When AI Unlocks.
Three tracks, each a straight line from the first lesson to a capstone. Each module tells you up front how much AI help you get there. Build modules are the projects you make along the way.
Python Foundations
- 01Make the machine do somethingno AI
- 02Data & variablesno AI
- 03Build: Vibe Check Quizno AI
- 04Making decisionsno AI
- 05Build: Smart Bill Splitterno AI
- 06Repetitionno AI
- 07Build: Your Year Wrappedno AI
- 08CollectionsAI explains
- 09Build: Expense TrackerAI explains
- 10FunctionsAI compares
- 11Build: Brainrot TrackerAI compares
- 12Thinking in objectsAI compares
- 13When things go wrongAI compares
- 14Coding with AIAI collaborates
- 15Reading other people's codeAI collaborates
- 16CapstoneAI collaborates
Java Foundations
- 01Make the machine do somethingno AI
- 02Data & variablesno AI
- 03Build: Vibe Checkno AI
- 04Making decisionsno AI
- 05Build: Bill Splitterno AI
- 06Loopsno AI
- 07Build: Year Wrappedno AI
- 08Collectionsno AI
- 09Build: Expense Trackerno AI
- 10Methodsno AI
- 11Build: Brainrot Trackerno AI
- 12Thinking in objectsno AI
- 13When things go wrongno AI
- 14Coding with AIAI collaborates
- 15Reading codeAI collaborates
- 16CapstoneAI collaborates
Junior Developer: Next.js
- 01From Python to the webno AI
- 02JavaScript fundamentalsno AI
- 03Your dev toolkitno AI
- 04HTML & JSXno AI
- 05Styling with Tailwindno AI
- 06Build: Profile Cardno AI
- 07TypeScript basicsno AI
- 08React components & propsAI explains
- 09State basicsAI explains
- 10Pages & routingAI explains
- 11Build: Personal PortfolioAI explains
- 12Advanced state patternsAI explains
- 13Build: Interactive QuizAI explains
- 14Async JavaScriptAI explains
- 15The server sideAI compares
- 16Connecting to SupabaseAI compares
- 17Build: Live DashboardAI compares
- 18Forms & full CRUDAI compares
- 19Build: Bookmark ManagerAI compares
- 20Authentication & protectionAI compares
- 21Build: Team PlannerAI compares
- 22Ship it & AI masteryAI collaborates
- 23Capstone: your app, your rulesAI collaborates
We Teach The Way The Research Says.
Fundamentals first, AI later.
Start conventionally, then bring AI in through deliberate steps; learners without the basics are confused by AI answers.
So the first seven Python modules are AI free, and the mentor unlocks in stages after that.
02Make the task one AI can't just finish.
ChatGPT-4 solved 100% of the simplest intro Python labs and 35% of the hardest.
So you trace, predict and explain code, not just submit an answer.
03Help after you try, not instead of trying.
With hint-first mentoring, 90% finished their projects versus 60% without (200 participants).
So task help opens only after you have made an attempt.
None of these studies tested Arclab. The notes list what each one found, what we built because of it, and what it does not prove.
Not A Video Course. Not A Chatbot.
| Feature | Arclab Foundations | Video course | Ask a chatbot |
|---|---|---|---|
| You predict the output before you run it | rarely | - | |
| Real code runs in your browser, nothing to install | sometimes | - | |
| Checked on what your code actually did | quizzes | - | |
| AI waits until you've tried | - | - | |
| You explain it back in your own words | - | - | |
| Join your bootcamp cohort with one code | sometimes | - | |
| Cost to learners | Free | Varies | Free tier |
The other columns describe the usual format, not any one product.
For Anyone Starting From Zero.
- never written code
- bootcamp student
- career switcher
- self-taught and stuck
- tired of tutorials
- scared of the terminal
- learning between shifts
- pre-work for a bootcamp
- uses ChatGPT, wants to understand it
- Python first, web next
- Java for class
- wants a portfolio project
Tracks
Python
Start from zero: 16 modules, 72 lessons and five build projects plus a capstone. Python for a world where AI writes code too.
Java
The same 16-module path in Java: types, loops, methods, objects and exceptions, with five builds along the way.
Next.js
23 modules and six projects, from HTML and React to Supabase and auth. The same stack the Sim uses, so you can go straight in afterwards.
Questions.
What does learning to code in the AI era mean?
AI can write most beginner code, so the skill that matters is reading code, predicting what it does and spotting where AI gets it wrong. Every lesson trains that, and AI joins in stages once you have the basics.
Does it cost anything?
Not for learners. There is no credit card and nothing to install. Bootcamps pay for cohorts.
Do I need any experience?
No. Lesson 01 is your first line of code. If you already code, skip ahead to any module.
My bootcamp gave me a join code. Where does it go?
Sign in, open Cohort in your dashboard and enter the code. Your organizer then sees your lessons and tasks completed.
How is this different from tutorials or ChatGPT?
Tutorials show you code; chatbots write it. Foundations makes you predict and write first, then compare with an AI draft, so you build judgment.
When does the AI mentor help?
In stages. In Python, modules 1 to 7 have no AI. Modules 8 and 9 let it explain, 10 to 13 let you compare with it, and 14 to 16 let you build with it. Task help opens only after you've made an attempt.
How is my work checked?
Multiple-choice answers are checked instantly. Code and written answers are checked by an AI model that sees what your code actually printed, not just what you typed.
Does my code really run in the browser?
Yes. Python runs in your browser with Pyodide, and the Next.js track boots a real project with a live preview. Nothing is installed on your computer.
How long does a track take?
Python Foundations takes about three to five weeks at around three lessons a sitting.
Do I get a certificate?
Not yet. You leave with the projects you built. We think your Sim and uptime99 results will say more to employers.
