Comment by bdangubic
3 days ago
the best way I found to deal with non-believers is to have claude run code reviews on their own work. I’ll point it to an older commit and get like 3-page markdown file :) works really, really well.
on one-shotting 3 minute prompt in 30 minutes though, software is a living organism and early gains can (and often result) in later pains. I do not use this type of argument as it relates to AI as the follow-up as the organism spreads its wings to production seldom makes its way to HN (if this 30 minute one-shot results in a huge security breach I doubt you would be back here with a follow-up, you will quietly handle it…)
You can get it to generate a 3-page markdown file for any random code, or its own code it just generated. If requested it will produce a seemingly plausible looking review with recommendations and possible issues.
How impressed someone get from that will depend on the recipient.
output, not recipient. try it on your own code. not everything on the example 3-page markdown you'll agree (much like you push back on the PR) but in significant number of occasions code changes were made based on the provided output
Recipient, as in the person who the output is intended for.
And I have seen what an AI do when it provides a code review, and it is very much like something that plausible looks like a code review. A lot of suggestion and nitpicks that at surface looks like plausible comments, but without any understanding. How much value a programmer get from that depend on the programmer. For me it reminds me of the value that teddy bears has on a support desk, or why some users are actually helped by being forced to go through layers of faq/ai suggested solutions before they are allowed to talk to a real person. Sometimes all that a person need to improve something is time to think about the code from a new perspective, and an AI code review can help the person find that time by throwing a bunch of shallow comments at them.