Like everything else with AI. It's only impressive if you don't look at it too closely, or if you just don't care about being beholden to perpetually paying some other company for code maintenance as a service.

The AI code works but it's a mess. There's like 2 or 3 different coding styles mixed together, multiple functions that do almost exactly the same thing, and whole swaths of code that never gets called because it doesn't ever clean up anything outside of it's context window.

I worry about what's going to happen to the code bases people will have to maintain if AI goes bust. I've spent the past week refactoring largely AI generated code because it bugged me that I didn't intuitively know how my own software was working.

Probably easier from a grading standpoint. When we were allowed to make our own conclusions it didn't matter whether they were well supported. It only mattered that we had the "correct" conclusion according to the curriculum.

I was told by a teacher in high school that our papers should not contain any original ideas. We were merely collecting and summarizing other sources.

I saw the sign. And it opened up my eyes.

Send some of that to the East Coast. It's hasn't gotten above like 35 in weeks.