The Assist

Blog » At work we recently started experimenting with generative AI for assistance with programming. We have a new Visual Studio Code plugin which we can ask questions in English, and it spits back code. It was a really interesting piece of, well, mandatory training. I've formed some opinions.

The main thing I dislike about AI coding assistance is that I have to very carefully review the AI's code to make sure that it does the right thing and none of the wrong things. And I, personally, find that reviewing code is more difficult than writing equivalent code myself.

I don't know if anybody else has this experience, but I understand code better when I interact with it directly by hand. Either by bringing code into existence, or by refactoring code which was already there to do something new, or to do the same thing more effectively. I need to take things apart and put them back together. I need to play with the code.

Passively reading someone else's code in the dead form of a git diff ...
4 likes ·   •  2 comments  •  flag
Share on Twitter
Published on April 04, 2024 17:58
Comments Showing 1-2 of 2 (2 new)    post a comment »
dateUp arrow    newest »

message 1: by Jon (last edited Apr 05, 2024 06:07AM) (new)

Jon We've been encouraged to use it where I work when looking at new integrations and I always defer. I can't even trust it's doing things like using the right/smart API calls vs just using the first one it finds.

And I'm not going to do an MR review or maintain code THAT NO ONE WROTE


message 2: by k-rice (new)

k-rice You've hit upon a known maxim in the tech industry (possibly not well know) - writing code is much easier than reading code. This is why reviewing code written by colleagues can be such a challenge, and also why having something generate code, such as an LLM, will most likely be more trouble than it's worth.


back to top

qntm's Blog

qntm
Follow qntm's blog with rss.