← Back to context

Comment by omoikane

9 hours ago

> reviewing code is very different from producing it, and surely teaches you less

Maybe he meant "reviewing code from coding agents"? Reviewing code from other humans is often a great way to learn.

I interpreted this as not as good a way to learn.

I learn the most from struggling through a problem, and reading someone’s code doesn’t teach me all the wrong ways they attempted before it looked like the way it now does.

  • Exactly. And vice versa, one of the biggest benefits of code review is calling out pitfalls you, the reviewer have ran into that the reviewee isn't aware of. LLM addicts won't have any experience with what works/doesn't work, so their reviewing will be pretty useless

  • I was thinking in situations where a coworker might send me something to review, and I might have thought "hmm, I wouldn't have done it like that, but this is a great way to do it too". Also, a good source of teachable code is to participate in a programming contest, and then review the repositories of the teams who scored better than me after the contest.

    I agree that if I don't already know how to implement something, seeing a solution before trying it myself is not great, that's like skipping the homework exercises and copying straight from the answer books.

    • Yeah I learn from reading other work too, but it doesn’t stick as well as when I work through it.

      The problem now is the pressure to use llms means creating more code but understanding so much less.

  • This is why tutorials in programming don't really teach much because you get the finished version. Not all the wrong steps that were taken, why they failed, what else was tried.

    These steps are what help you solve other issues in the future.