Phase 0: Vibe Coding Fundamentals · 15 min · Claude Code, Cursor, Lovable, v0, Bolt.new, Replit Agent, or similar

Why This Course Makes You Do Some of the Work

If you never struggle, you never learn. If you never learn, you can't debug — and you can't ship safely.

Real-world application: What changes when you use this: you can debug when AI-generated code breaks, not just demo when it works — which is the difference between someone who can ship and someone who can't.

What you will learn

Understand the struggle-before-automation principle and why it matters
Recognize the sycophant problem — AI agrees with your framing instead of challenging it
See how this course sequences struggle before automation in every phase
Understand what 'doing some of the work' actually means in practice

Objective

Your mission

You'll identify where the AI's "sycophant problem" has affected you, find a skill gap you're relying on AI to fill, and commit to the predict-before-reveal habit for the rest of this course. Output: struggle_commitment.md.

You need: A text editor. Honesty.

Time: 15 minutes.

Step 1: recall a time AI agreed with you — and you were wrong (5 min)

Lane Wagner (creator of Boot.dev) calls this the "sycophant problem": AI loves to agree. You say "I think the bug is in the database" → the AI agrees and looks at the database (even if the bug is in auth). You say "this is ready to ship" → the AI agrees (even if security is missing).

Think of a specific time this happened to you. Not hypothetical — real.

Write in your file:

## When the Sycophant Got Me
Situation: [what were you building/doing?]
What I said: [what did you tell the AI?]
What AI did: [how did it agree with you?]
What actually happened: [what was the real issue?]

Step 2: find your blind spot (5 min)

Where are you relying entirely on AI and couldn't explain what it did if someone asked?

Examples:

"The AI set up authentication and I have no idea how it works"
"The AI configured the database connection and I'd be lost if it broke"
"The AI wrote the deploy config and I don't know what any of it does"

Write in your file:

## My Blind Spot
The thing I rely on AI for but couldn't explain: [be specific]
Why it scares me: [what would happen if it broke and the AI couldn't fix it?]

Step 3: commit to the predict-before-reveal habit (5 min)

This course's core principle: struggle before automation. Before the AI acts, you think. Before the AI debugs, you predict. Before the AI reviews, you check.

This isn't "learn to code so you don't need AI." It's "build enough understanding to direct the AI and catch it when it's wrong."

Write your commitment:

## My Predict-Before-Reveal Commitment
For the rest of this course, before I paste any error to the AI, I will:
[Write a specific action — e.g., "write down my prediction of what's wrong before pasting the error"]

Before I accept any AI build, I will:
[Write a specific action — e.g., "check it against my spec feature by feature before accepting"]

Before I ask the AI to choose an approach, I will:
[Write a specific action — e.g., "write down what matters for my project first"]

The struggle-before-automation principle
This course sequences struggle before automation in every phase. Early reps ask you to predict, guess, or decide BEFORE the AI acts. Later reps let the AI accelerate what you already understand. This is the answer to "how is this different from just letting AI do it" — you build judgment first, then speed.

You tell the AI "I think the bug is in the database" and it starts editing database code. What should you suspect?

The sycophant problem: AI tends to agree with your framing. Predict first, then make the AI show evidence for where the bug is instead of following your guess.

Key Terms & Next

You're done

Save your struggle_commitment.md. Revisit it whenever you catch yourself pasting errors without thinking first.

That's the whole lesson. Now do the lab.

Its hands-on lab, and the lessons after it, open with a free account — the opening phases of 29 of 35 courses are free, no credit card.