You didn't choose your stack, the Vibe did
Most people vibe coding right now are building on a stack they never actually chose.
They opened Cursor, described what they wanted to build, and the AI scaffolded something that looked reasonable. So they went with it.
Or they watched a YouTube video where someone shipped a SaaS in 45 minutes using Next.js, Supabase, and Drizzle, so now that's just the default.
Or a tweet went viral about some new framework and suddenly it's in every tutorial.
Three months later, they're six levels deep in a bug they don't understand, in a part of the codebase they've never had to think about before.
Sound familiar?
The trap nobody talks about
Vibe coding is really good at one thing: removing friction.
The gap between having an idea and having something that runs has never been shorter. That's the whole point.
But that friction you're removing? Part of it was doing useful work.
When you had to set up a project manually, you were forced to briefly confront what you were pulling in. You'd hit an error, Google something, skim a Stack Overflow thread. Not exactly fun, but you were building a mental model whether you liked it or not.
Vibe coding skips that step entirely.
The scaffolding happens so fast that the decision-making gets bypassed.
The AI picks something sensible based on what it's seen a million times. The tutorial creator picks what they know. The viral tweet picks what got the most engagement that week.
None of them picked what's right for your project.
The gap between expectation and reality
There's a really consistent pattern here.
It's not that any of these tools or stacks are bad. It's that they weren't chosen for you. They were chosen for a generic project that doesn't exist, by someone who wasn't thinking about your constraints.
Three questions worth asking before you accept the scaffold
You don't need to slow down. You need about ten minutes at the start of a project to actually make a decision instead of inheriting one.
These aren't hard questions. But most people don't ask them because the AI already has an answer ready, and it's really easy to just say yes.
The vibe is great. The verification has to come from you.
AI is a brilliant builder. It'll scaffold, generate, and ship faster than you can think. That part isn't going anywhere and it shouldn't.
But the AI doesn't know your constraints. It doesn't know your team, your timeline, or what a debugging session at 2am actually costs you. It just knows what a reasonable project looks like in general.
You're the one who knows the specifics. So be the one who makes the call.
When was the last time you actually chose your stack, rather than just accepted it?
