It's also just inherently anti-intellectual, like "better the devil you know" - how about you actually justify the fence instead of lazily hiding behind "it must be there for a reason". Give it some thought, come on. Otherwise you can't blame people for concluding the fence is self-serving bs.

Argh, sorry about that! I can confirm your submission was received.

Congratulations Joseph!

I think you always need to be careful having the same string serve as 2 roles (argument in the function and column name)

if you ask it to write a whole paragraph or more from scratch, it's pretty bad. If you give it a sentence that you know sounds wonky, but you're not sure how to improve (eg because you're writing in a second or third language), it can be helpful.

It's easier to notice when a sentence is well written than to write a good sentence yourself. I don't personally use LLMs to write, but my ESL students do, not to draft from scratch but in cases when they aren't sure how to phrase something. LLMs can give them options, and they can pick their fave

My criterion for whether an LLM can be useful is when it is difficult to find a solution to a problem, but easy to verify that solution (like cryptography). If I am coding in a language I'm not familiar with, I can ask "how can I best achieve this outcome?" and then easily check whether it works

The graph says nothing about whether people are using it to improve at things they are not good at. If you assume that everyone is using it to replace their brain in each of these areas, then sure, that's bad - but that's not what was asked. The survey just asks in what domains people use it.

You can't ask google translate: "suggest other ways I might phrase this", and then pick your favourite. Very useful for non-english speakers, or even people who are learning to write well.

Happy birthday!

I would settle for academic types reliably making this distinction