You are an English major. A pre-med. A history student. A psychology undergrad. You looked at code once, decided it wasn't for you, moved on. This chapter argues you're wrong — not because you should switch majors, but because the calculus has changed and most non-CS students haven't noticed.
The argument
Every student should learn enough code to automate the boring parts of their life. Before AI, this was hard enough that opting out was reasonable. After AI, opting out is a mistake.
What "enough code" means
You do not need to become a software engineer. You don't need to learn data structures, algorithms, big-O notation, or any of the things that make CS majors miserable in their sophomore year. You need exactly two skills:
-
Read a script well enough to understand what it does. This means recognizing that a
forloop processes things one at a time, that a function takes inputs and returns outputs, that a variable is just a labeled container. That's the floor. -
Modify a script someone else wrote. Change the file path. Adjust the date range. Add an extra column to the output. This is where the leverage is. Most "small programs" you'll ever need already exist somewhere — the skill is reading them, understanding them, and bending them to your specific situation.
That's it. Two skills. AI handles everything in between.
What you can automate
Look at your week. Pick out the recurring 30-minute tasks. Almost every one of them can be automated.
- Class schedule into your calendar. Convert a weird CSV to a Google Calendar import.
- Reading list into Anki cards. Auto-generate flashcards from PDFs.
- Renaming 50 files at once. "Lecture Recording 2025-09-14 14:32:11.mp4" becomes "Bio101 Week 3 Lecture 1.mp4."
- Scraping a job board. Get a daily digest of new internship postings in your inbox.
- Tracking scholarship deadlines. Watch ten websites. Get notified when any update.
- Cleaning data for a research project. That spreadsheet with messy column names — clean it once, in code, never touch it again.
Each of these saves between 30 minutes and 5 hours per week. Compound that over four years.
The starter setup: Cursor
Don't battle with package managers. Use an AI-native code editor.
Cursor is VS Code with AI baked in. Describe what you want in plain English and watch the code appear. Highlight any line and ask "what does this do?" It removes the 40% of coding pain that comes from typos, imports, and syntax errors. The setup walkthrough lives in Cursor AI IDE Workflows.
If you want to skip code entirely, Build AI Apps with No Code covers that path.
Your first project: scope it tiny
The single biggest mistake first-time coders make is picking a project that's too big. They try to build "an app" instead of a script. They get lost in week one. They quit. Don't do this.
Here's a tiny first project that's perfect for a non-CS student.
Goal: Take your class schedule from your school portal and turn it into Google Calendar events.
Why it's a good first project:
- The input is real (your actual schedule).
- The output is real (it goes in your calendar, you use it Monday).
- The scope is small (maybe 50 lines of Python).
- It teaches the core concepts (read a file, transform data, write a file).
- It saves you actual time forever.
The workflow.
Open Cursor. Create a new folder for the project. Create a file called schedule.py. Put your prompt to the AI sidebar:
I have a CSV exported from my school's class portal. Each row has a
class name, day of the week, start time, end time, and location. I
want to convert this to an .ics file I can import into Google Calendar.
The semester runs from Sept 2 to Dec 15, 2025. Write me a Python script
that does this. Walk me through what each section does as you write it.
The AI writes the code. You read it. You ask follow-up questions where you don't understand a line. You run it. It probably fails the first time because your CSV column names don't match what the AI assumed. You paste the error back:
Got this error: [paste error]. Here's what my CSV header row actually
looks like: [paste row]. Adjust the script.
The AI fixes it. You run it again. It works. Three hours of effort. Now your schedule is in Google Calendar with reminders, and you have a script you can re-use every semester.
You also just learned Python. Not all of it — but enough to read the next script and modify it. That's the floor. Everything else builds from there.
What to do when stuck
You will get stuck. Code is unforgiving in a way prose is not — a missing colon breaks everything, and the error messages can be cryptic.
The two prompts that solve 80% of stuckness:
I'm getting this error: [paste full error]. Here's the relevant code:
[paste]. What's going on, what did I do wrong, and how do I fix it?
Explain in plain English.
I want to do [thing] but I don't know what's possible. Don't write
code yet. First tell me: is this a normal thing to do? What are the
common ways people do it? What library would I use? What are the
gotchas?
The second prompt — asking before you code — is the move that separates non-CS students who get good at this from the ones who flail. Don't dive in. Plan first.
What to learn after the first project
Pick another tiny project. Build one a week for a month. By the end you'll have shipped four small tools and know more practical Python than 80% of non-CS undergrads.
For chaining multiple AI calls — "summarize an article, extract key points, generate flashcards, email them to me" — see AI Prompt Chaining Workflows.
Why this is a career move even if you never become an engineer
Every field is becoming computational. A history student who writes 50 lines of Python to clean an 18th-century shipping dataset has a research edge. A biology student who automates lab-data parsing graduates with a portfolio consulting firms notice. A psychology student who can analyze survey data without paying for SPSS owns their own pipeline.
You don't need to be a CS major. You need to be able to glue things together when no off-the-shelf tool exists. AI just made that achievable in a weekend.
Open Cursor tonight. Pick a tiny problem. Ship it.