> These days, I'm fairly senior and don't touch code much anymore but I find it really really instructive to get my hands dirty and struggle through new code and ideas. I think the "just tweak the prompts bro" people are missing out on learning.

If you just use prompts and don't actually read the output, and figure out why it worked, and why it works, you will never get better. But if you take the time to understand why it works, you will be better for it, and might not even bother asking next time.

I've said it before, but when I first started using Firefox w/ autocorrect in like 2005, I made it a point to learn to spell from it, so that over time I would make less typos. English is my second language, so its always been an uphill battle for me despite having a native American English accent. Autocorrect on Firefox helped me tremendously.

I can use LLMs to plunge into things I'm afraid of trying out due to impostor syndrome and get more done sooner and learn on the way there. I think the key thing is to use tools correctly.

AI is like the limitless drug to a degree, you have an insane fountain of knowledge at your fingertips, you just need to use it wisely and learn from it.

If that held true, than reading lots of source code from random FOSS projects would make you into an amazing coder. But clearly that's not enough to internalize and learn all that knowledge, you need to experiment with it, see it running, debug it, write some, extend it, fix bugs and so on. Just reading it with very little context for the rest of the project is a recipe for mediocrity.

It's different to read and understand a prompt and actually produce the code. It's a different level of cognition.