Man, vibe coders really think highly of themselves and their AI outputs.

As someone who actually understands the language used in AI generated scripts, AI is shit at writing code. It sometimes gets decent wins and helps me figure out something quicker than without, but I can count on my fingers the number of times I’ve gotten a good, and usable, bit of code from it. Vastly more often than not, I have to edit the code to make it run (because it hallucinates functions, parameters, and constantly uses reserved variables even after being corrected dozens of times) only to find out it doesn’t even give the right output, or more often, outputs nothing at all.

You show the quality of your knowledge by the inverse of the trust you put in AI code. It’s decent at blocking out basic things, but anything past that is a crapshoot at best