Two people are using Claude to code.

Fable asks both: Database is insecure, proceed?

One says: “Screw it! Keep building new features!”

The other says: “You’re right. Let’s stop and fix all the security issues before we move on.”

And so, will both of these approaches yield equally secure apps?

Oftentimes, bad/buggy/bloated code is that way because no one took the time to prompt it better. Even without writing a line of code, spending time debugging, fixing security gaps, etc., is a chore.

So is AI to blame for a prompter who never focuses on optimizations, security, maintainability, and reducing code bloat?

Or is it simply a case of operator ignorance?