That's all good to restrict AI usage in codeforces, and since the blog restricting it has been recently up. I decided to post this.
Last round rk170 , qasxcvgwe, used AI. It is much obvious and he wasn't skipped even though Plag check is supposed to having been already finished ?
Obvious by excessive comments and even spamming many submissions, switching between C++, java and py, and last but not least, failing a constructive problem B-Mex (AI pretty sucks at those)
E1 which was a master rated problem (per Clist), solved by qasxcvgwe at 43rd minute, which is fairly early for such a problem.
Tbh E1 is not a master rated problem.
I agree (much easier than D), it had lower subs because placed at the end of the set. But at least it is rated as so on Clist, 2211.
the heck so AI solved a 2000 rated problem?
And it failed to solve 1122 rated problem. I think type is what matters not rate.
E could be brute-forced, known data structure. While B is constructive algo.
im a newbie so im not much aware of data structures but was E1 really that simple that it was a direct implementation of some data structure
Also i went through the guys contest problems he was switching between python and c++ continuously during the contest lol also funny how he got so many TLEs and WAs
I can't solve it. However given fact that you can afford n cubed means we can get the distance between any two nodes and still not fail—it seems plausible that a little thinking is what it needs.
what i can see here is an user directly paste AI code to submit, i think he is genius
https://mirror.codeforces.com/blog/entry/134829
Oh sorry, I didn't see this post. But the issue, even after it he remained rated for the round xD
Yeah, it is very unbalanced for legit users