[llama052]

100% this. I observe this all the time. LLMs can be a force multiplier but those who don’t ask the right questions or understand the nuance still produce bad code, it’s just amplified. I don’t think any of the current models can avoid that, especially when it’s based on data it’s been fed, which is historically human generated.

[tetha]

And interestingly, a few good people at work are learning faster with the LLMs at hand.

To a degree, it looks like LLMs help them overcome the blank page anxiety and help them with the grunt work of, well, actually writing code, which actually contains a lot of technical nuances to keep track of. But once that's down, I'm having very good discussions with them on what makes good, maintainable and sustainable code bases.

It's almost like the old writing advice my Mark Twain: “Writing is easy. All you have to do is cross out the wrong words.” It's easier to dislike a part of something existing and fix that, than trying to create the perfect thing at once.