Is vibe coding bad? The honest answer is: it depends on what happens next

"Vibe coding is bad" is a take you'll see a lot, usually from people watching someone ship code they don't understand. There's a real point in there. But the criticism aims at the wrong target. Vibe coding - describing what you want and letting AI write it - isn't the problem. Trusting the result blindly is.

The fair criticism

Here's the version that's true. When you generate code you can't read and put it straight in front of users, you've built something you can't reason about. If it breaks, you don't know why. If it's insecure, you can't tell. The code compiled and looked done, so it shipped - and a build passing is not the same as it working. That's a genuine risk, and hand-waving it away is how vibe coding earns its bad reputation.

Vibe coding isn't bad. Shipping untested code you can't reason about is bad, and that's true no matter who wrote it.

Why the idea itself is fine

Now the other side. A human developer also writes code that others don't read line by line. The reason we trust it anyway is not that a person typed it - it's that it got tested, reviewed, and run before real users touched it. Those safeguards are what make code trustworthy, and none of them care whether a person or a model wrote the first draft. Vibe coding is only reckless when it skips them.

The version that isn't bad

So the fix isn't to stop vibe coding. It's to put the safeguards back. That's the whole idea behind pondas: you describe the app, but a team of agents builds it, another tests it by actually running the flows, and one reviews the code before it ships - and it lands in a repo you own so nothing is a black box you can't open.

Is vibe coding bad? Vibe coding without a net is. Vibe coding with testing, review, and ownership is just building software faster. The tools worth using are the ones that give you the second one.