Ah, yes, the problem isn't that AI produces bad programs: it's the fault of us for being fuddy-duddies and not accepting mediocre software.

An assistant who makes a lot of mistakes and needs to be corrected all the time is a bad assistant. More, LLMs are so far good at simple stuff that a competent programmer would get right anyway, and just terrible at anything bigger or more sophisticated.

Starting off with a big disclaimer about how you aren't interested in quality software but software that is "good enough" - a term you never actually define, by the way - gives the whole thing away.

Don't get me wrong - I work in AI myself, and I think this is a very promising field, but looking at the current results and saying it is "good enough" says a lot more about your work than anything else.

And yes, dozens of articles like yours talking down to others have left a bad taste in my mouth. The same article, written with less dogmatism, would have gotten claps from me.