Asking for Fixes Without Spiraling

The Spiral Problem

When debugging with AI, conversations can spiral:

1. You report an error
2. AI suggests a fix
3. The fix creates a new error
4. AI suggests a bigger fix
5. Now your code is completely different
6. Nothing works

How to Avoid Spiraling

1. Describe the Problem Precisely

Bad prompt:

"It doesn't work"

Good prompt:

I'm getting an error in main.py.

**Expected behavior:** When I enter my name, I should see "Hello, [name]!"

**Actual behavior:** I see a TypeError

**Error message:**
TypeError: can only concatenate str (not "NoneType") to str

**Relevant code:**
name = input("What's your name?")
print("Hello, " + name + "!")

What's wrong and how can I fix it with a minimal change?

2. Ask for Minimal Fixes

Always include constraints:

• "Fix this with the smallest possible change"
• "Don't refactor other code"
• "Keep the same structure"
• "Just fix the bug, don't add features"

3. Request Explanations

Fix this error and explain:
1. What was wrong
2. Why your fix works
3. How to avoid this in the future

Keep the fix minimal - don't refactor.

4. One Fix at a Time

If AI suggests multiple changes:

• Apply only the first one
• Test it
• Then consider the next change

Escape Hatches

When things are spiraling, use these escape hatches:

The Reset

bash
Copy
# Discard all changes since last commit
git checkout .

Then start fresh with a clearer prompt.

The Constraint

Add to your prompt:

"Do not modify any other files. Only change [specific file/function]."

The Simplification

"Ignore my previous request. Let's start simpler: what is the minimal code needed to [core requirement]?"

The Explanation Request

"Before suggesting a fix, explain what you think the bug is. I want to understand before changing code."

Red Flags to Watch For

Stop and reassess if:

• The AI suggests installing new packages
• The fix involves rewriting large sections
• You're on your third "fix" without improvement
• The AI says "let's refactor this"

Practice: Intentional Bug

Let's practice with an intentional bug:

python
Copy
# buggy.py - This has a bug!
numbers = [1, 2, 3, 4, 5]
total = 0

for i in range(1, len(numbers)):  # Bug: starts at 1, not 0
    total += numbers[i]

print(f"Sum: {total}")  # Prints 14, should be 15

Try this prompt:

This code should sum all numbers in the list, but it's giving the wrong answer.

Expected: Sum should be 15 (1+2+3+4+5)
Actual: Sum is 14

What's the minimal fix? Don't refactor the code, just fix the bug.

Next: Ship something small, quickly.