AI coding broke the Ballmer Peak

(adriankrebs.ch)

1 points | by hubraumhugo 2 hours ago

1 comments

  • davydm 2 hours ago
    Overly optimistic this relies entirely on the assumptions that you are able to architect and debug (two rather brain-powered things) better than writing plain code that at least a compiler would error on if it was completely bonkers wrong.

    I think there's soon going to be an analogy to the dunning-kruger effect, but for vibe coders, that would reveal just what a self-deception this is.