Hacker News Reader

Comment by jstanley | Hacker News Reader

I found that when I tried to use Cursor for blogging I was totally infuriated by it. As happy as I am to have Cursor auto-complete my code all day long, the sensation of having Cursor try to put words into my mouth (or, worse, put thoughts into my head) was very unpleasant.

So I stopped trying to use Cursor for writing and went back to vim.

But I found the unpleasantness of the sensation very surprising! I love using Cursor for writing code, but for some reason, even though I am primarily a programmer, it's much more important to me that my words are a "true expression of my spirit" than that my code is. I couldn't quite figure out why.

View on HN

The reason I barely use copilot is because it tries to write comments for me. I don’t get why they don’t have a button that says “disable autocomplete within comment blocks” (since the IDE knows which things are comments), but it’s so yucky feeling when it tries to write something I wouldn’t say.

I find the comments Copilot proposes is better than average comment quality for the code base I routinely work on: maybe you are so great you don't need any help ever, but that's not true for the average software dev.

> I find the comments Copilot proposes is better than average comment quality for the code base I routinely work on

Maybe the average is just so bad? The completions I get for comments are document what the code is doing, which is not something that I ever put into comments. It's always:

a) A highlevel (prefixed to the function/block/scope definition) list of steps, input expectations and output expectations for the forthcoming function/block/scope

b) A note explaining why the code does what it does.

A comment repeating the code but in English is useless.

The best comments are the type of:

  Whatever bad thing you think of this part is probably true, but I can’t do better.

And

  Beware, weird optimization ahead.

And

  Editing this code requires arcane knowledge.

It's not that my comments are infallible, but if I write something wrong/silly it'll be caught during code review. similarly if there's a comment missing before some arcane nonsense nobody will remember in 3 years, then i'd expect a PR reviewer to tell the dev to add one.

Copilot just likes to puke very useless comments whenever I type "//", especially in autocomplete mode (I don't really use the chat mode).

https://docs.github.com/en/copilot/customizing-copilot/addin...

Not sure that would prevent it offering something as autocomplete in a comment block, but could ask it to write less/no comments generally

I've had the same experience with writing non-code text. I often use VSCode to write markdown documents on various things and having Copilot chime in to complete my mealplan for the week with 5 different kinds of chicken wraps is a weird feeling lol

I think code is purely functional - there might be a style component to it, but mostly what I care about is "does it do the thing I want it to do".

Writing, on the other hand, is much more about expressing one's inner self.

I write music and poetry as a hobby, and not once have I found chatGPT to be useful in this regard. Because the words aren't mine.

I've found that writing, then asking the chatbot for suggestions and selecting the ones I'm happy with is a reasonable approach. Treat it as an editor or a proof reader or a thesaurus on steroids. Not a pattern that really works with autocomplete, I guess.

Reply on news.ycombinator.com