Hey everyone!
I’ve been thinking—what if we used AI to translate code between languages on codeforces. It sounds attractive: you write a solution in Python, but it's too slow, and AI magically converts it to C++ for better performance. Cool, right?
But here’s where I’m confused. If AI does all the converting, am I really learning? Sure, it saves time, but could it lead to mistakes or make things too easy? Isn’t part of the fun (and frustration) of competitive programming solving things manually and improving by switching languages ourselves?
On the flip side, using AI could actually help me learn new tricks by comparing the code it generates. So I’m torn—should we rely on AI for this, or stick to doing it ourselves?
What do you think? Would AI make life easier, or would it take away from the experience?
Auto comment: topic has been updated by het20122005 (previous revision, new revision, compare).
I'm sticking on the conservative side: just convert codes between languages by your own.
It's not about the ethical AI uses to me. It's that, programming languages are once again just human constructs, and susceptible to inconsistency between designs. Having different speed (like between C++ and Python) is one thing, but another is sometimes code behavior might be flipped up-side-down if the original one isn't strictly weaved, thus causing unexpected responses.
Or sometimes, talking in a simpler analogy, you see Google Translate still struggling in perfectly conveying the context and intonation from one language to another? Same thing with this. And to me human intervention is required to decide what would our output code should turn out to be.
I think it's a good idea to do that outside of rated competitions if you want to learn new language. Especially transitioning from C\C++ to Rust: it's pain in the ass to learn it by yourself without proper guidance. As for Python -> C++ it's not particularly great, you won't learn all the cool lower level features and caveats of C++ like that, you'll just get used to writing suboptimal code in Python and using LLMs as a crutch.