what a fake app taught me about building real ones
you wouldn't remodel a house you haven't designed yet
You wouldn't remodel a house you haven't designed yet.
But that's exactly what most vibe coders do with their projects.
Back in 2017-ish, I was “livin’ the dream” working at a tech startup… it was one of those places with beer on tap, kegs of cold brew, paid gym memberships… you weren't even allowed to wear shoes in the office :D
I'd bounced between different tech roles: sales engineer, technical account management… eventually head of tech partnerships (look at me now mom!)
Anyway… at one point, the product team was doing a massive overhaul of our reporting tools.
At the time, we were the largest Facebook Marketing Partner in the world by ad spend. About $2bn (with a b) in annual ad money flowed through the platform.
Which meant the tracking and reporting had to be:
#1 - able to handle a shit ton of data
#2 - able to surface insights across tons of data
One day the lead Product Designer pulled me in for a “power user research” session.
He'd drop prompts on me like:
"How would you find individual ad performance?"
"How would you compare metrics across two different time periods?"
"How would you use the array functions to create custom reporting columns?"
Then he'd watch what I did.
But here's the cool part: I wasn't touching the real product.
It was an interactive Figma mockup. Fully clickable, but totally fake.
I could drill into menus, toggle filters, explore dashboards — all before a single line of frontend code existed.
And based on what me (and some others) did? They changed a LOT.
Turned out, some features they thought were obvious… weren't.
Now imagine if they'd skipped that step.
Imagine a full frontend build. Designer hours. Developer hours. QA hours.
All to ship something that confused the first real user who touched it.
This ain't a hypothetical! It happens to 99% of the vibe-code-universe.
It's what happens when you build before you design.
Here's why I'm telling you this.
Most vibe coders go straight to prompting Claude or Cursor to build the UI.
They think: “hey, I've got [SpecKit/GSD/PlanMode/Whatever] - let's get buildin’!”
It feels like progress. The app is taking shape. Code is being written.
But they're renovating a house they've never blueprinted.
Then they act like a sad puppy dog when the kitchen is in the bathroom and the new Golden toilet they ordered off Temu is in the guest room.
So they rebuild, rebuild, rebuild… never quite hitting something that feels right.
Well… Google Stitch just did a massive overhaul that fixes problems like this.
There's actually 5 new feature sets, and they're pretty sweet.
If you want to see exactly how it works and how to fold it into your workflow before your next build...
Then you'll want to check out my latest YouTube video where I break down all the new updates.
It might just save you from rebuilding the same screen six times… unless of course, you're into that?
https://www.youtube.com/watch?v=Seub7NNBF8g
WHAT ELSE I'VE BEEN READING ABOUT OR DOING
Things I’ve Learned
Thing 1
I had one of the most productive days of the last two years yesterday. You know those days where for some reason you wake up 30 minutes before your alarm and you're lit up like a Christmas tree? You've crushed all your work for the day before the normies have wiped the morning crusties out of their eyes? And today is no different, or else this email might never have existed. So yes... imagine me actually YELLING THIS EMAIL EMPHATICALLY through the screen!
Why do I tell you this?
I started taking some supplements to optimize my sleep & daily focus and it's got me wired. And since you're an amazing human being that decided to join my group/newsletter, I thought I'd share a nugget. I don't want you thinkin' I'm trying to sell you supplements, so I'll type out what I'm taking (obviously this shit ain't medical advice):
Night-time: L-theanine, Magnesium Glycinate, & Magnesium L-Threonate
Day-time: Mars Men, Vitamin-D/K2
Thing 2
I found a pretty handy agent skill for creating premium-feeling frontends.
I've used it for landing pages and websites mostly, but it's worth checking out if you want to roll something kinda custom but still avoid the “AI SLOP” aesthetic: https://github.com/Leonxlnx/taste-skill/
Thing 3
#3 - Some food for thought for when you decide to finally slap a pricing page on one of those amazing projects sitting in the graveyard:
The Anchoring Bias - Users rely heavily on the first piece of information they see when making a decision.
What does this mean for you? If you're in the US - readers read from left-to-right.
Placing a high-priced subscription option on the left and then placing a lesser-priced valuable option to the right of it increases the perceived value of the second option.
If you've ever been on a sales call where they tell you something costs $10,000 but then drop the price to $7,000 because they really like your spunk - that's the anchoring bias at work.
Biases like this are powerful because it's an inescapable force of gravity… so use it respectfully!
- Sean