129 | Vibe Coding & the New Pareto Principle (The 90/20 Rule)
What feels like 90% done may only be part of the way there... why finishing software is harder than starting it.
I was on fire. In less than 30 minutes, I had a spinning wheel app running in Cursor that looked decent and worked flawlessly. You click, the wheel spins, and it picks a random name from a list. I was vibing hard. I couldn’t help but think back to when I was writing Fortran on an IBM PC and debugging with Print statements. It was magical then, too—but this? This is rocket fuel by comparison.
But then, what started as a playful burst of creativity soon turned into a grinding crawl. That wheel app development? It’s a perfect example of what I call the 90/20 Rule.
What is Vibe Coding?
Vibe coding is what happens when building software feels less like assembling furniture (let’s not name the Scandinavian furniture giant, but you know the one—with the extra screws) and more like creating art. With the rise of generative AI tools like ChatGPT and Cursor, coding has become more intuitive, fluid, and dare I say…fun?
You don't start with a blueprint and measurements like a contractor building a house. Instead, you start with a spark, like a sculptor seeing the finished piece in the raw marble. You chat with the AI, tweak, test, build. It's less construction, more conversational. It’s improvisational. It’s exploratory. It’s fast. And a word of warning: it can be addictive!1
Cursor: Vibe Coding's Best Friend
Cursor is a development platform supercharged with AI. For professional software engineers, it’s like having a thought partner that understands your whole codebase. Ask it to explain a function, refactor some code, or generate a new component—Cursor gets it done. For vibe coders, Cursor is the dream tool: you stay in flow while it handles the heavy lifting.
In my case, it was perfect for that spinning wheel app. I had the visual, the logic, and the interactivity all live in under an hour. The power and speed were intoxicating.
Introducing the 90/20 Rule
Here’s the core idea: when you feel 90% done with your app, you’ve probably only invested about 20% of the total time it’s going to take.
It’s the vibe coder’s version of the Pareto Principle, but reversed. Instead of 80% of the value coming from 20% of the effort, this rule reminds you that the last 10% of what you want the app to do might take longer than all the exciting stuff that came before it.
Think of it like renovating a kitchen. Tearing out the old cabinets feels like huge progress—you can see the transformation! But then comes the tedious work: adjusting measurements for uneven walls, waiting for the electrician, finding the right hardware. Finishing the job? It can take months.
The Hidden Half
That last 10%—you know, the part you were sure would just take “a few more minutes”—includes things like:
UI polish so it doesn’t look like a demo
Little optional features that suddenly feel essential
Handling errors gracefully (or at all)
Dealing with edge cases you hadn’t thought of
Testing, testing, and more testing
Integrating with other systems
Deployment and user settings
And if you're feeling particularly optimistic, writing documentation that future-you will actually understand
This stuff isn’t glamorous. It’s where vibe meets the real world. And it’s slow.
Case Study: The Spinning Wheel App
The app was simple: a colorful wheel with names that spins on click. Cursor helped me get the basic version live in under 30 minutes. But here’s what followed:
Adding the ability to edit the name list dynamically
Making the pointer more prominent2
Deploying it so that others can run it
Improving the UI so it didn’t scream "prototype"
Each tweak took time. Sometimes, just one change took twice as much as building the original basic wheel. This experience was more like the 90/20 rule -- when I thought I was almost done, I still needed to invest a lot of time to make it into something others could access and use. Each tweak revealed another tweak that needed tweaking. The beautiful simplicity of vibe coding had collided with the messy reality of real-world use.
What Vibe Coders Need to Know
The 90/20 Rule isn’t a warning. It’s a lens. If you’re building a prototype to show off an idea, you’re golden. (If you've been around web design, think Figma.)
Ride the vibe, get to 70% or 90%, and stop.
But if you're building a volunteer sign-up tool or donor intake form—something your community will rely on—then brace yourself.. The real work begins when you think you’re almost done. The good news is that tools like Cursor, Bolt, Replit, and other low/no-code tools have general-purpose AI models built in, so you can continue to have a conversation to get help figuring out how to do that "last 10%" of the work. It's just that many of those tasks are outside the control of the AI coding app, so they may send you off into the complicated, fragile world of API keys, GitHub repos, and various other foreign lands of the geeks.
Know When to Stop (and When to Keep Going)
Vibe coding is a superpower for mission-driven organizations. It lets you quickly prototype a resource directory for your clients, test a new volunteer-matching workflow, or mock up a chatbot for common FAQs. But understanding the 90/20 Rule helps you make smarter decisions about where to invest your time.
If you're exploring whether AI can solve a problem for your team, ride the vibe to 70-90% completion. You'll learn what you need to know without getting stuck in the polish phase. But if you're building something your community will rely on, budget accordingly—and maybe partner with someone who enjoys crossing t's and dotting i's.
The goal isn't to avoid the grind, but to choose it consciously. Sometimes the prototype is enough. Ask yourself: is this tool just to show the ED what’s possible, or will staff be using it in the field every week? The 90/20 Rule helps you decide whether to keep going or to hand it off for more structured support. The 90/20 Rule simply helps you tell the difference.
And hey, maybe someday I’ll finish that spinning wheel app. Or maybe it’s done enough. In either case, I’ll be sure to follow my own advice to…
Make Good Choices
Lucky for me, I have a very understanding and indulgent husband, who has learned to live with my “just one more tweak” marathons!
After a half-dozen attempts over 10-15 minutes, I finally gave up!
I love the “this is not a warning, but a lens” vibe! I think that mentality strikes balance between extreme caution (which doesn’t feel true) and extreme voyeurism (which doesn’t feel wise) - love your content!