For a while now, I have been doing badly in contests, so when I finished today's div. 3, I was feeling pretty good about myself. I decided to look for people that placed near me to see how my rating would get affected.
I went to check some of the code people had submitted, and realized just how much of it was ai.
There are full pages of people with only ai code. People who started cp a week ago getting AC on F. People who only solved 800s suddenly solving 2000+ rated problems.
This scares me. How are we supposed to compete fairly in an environment where everybody and their grandma uses ChatGPT or Gemini to solve whichever problem they want for them? I hope that there will be days that in contests no one uses ai, rather than everyone starting to do so.
This is why I'm not sure about putting a lot of effort into improving. Depressing.
I think instead of focusing on ranking, focusing on improving your problem solving skills, would make the journey worthwhile.
What's fun about grinding complex problems in my lonesome without a way to accurately gauge how I stack up to my peers? I've already graduated from university. There's no ICPC or IOI for me. The online racket is all I have.
This is exactly where I'm at too.
I feel like a CodeForces fork that does some kind of offline/document verification (so that there's actual consequences if found cheating) is the only path forward here.
If the problem quality is S-tier (e.g. USACO, CF, AtCoder), I think it could work...
It does get hard to only focus on improvement where there isn't any way to measure yourself against others.
I agree with you. I think that there should be a way for submissions to be verified.
However for now, we can only hope.
what about once a user submits a file codeforces can check the version history or edit history to make sure there aren’t any copy-paste’s? then the cheaters will have to manually type it which really could be done a lot faster by actually trying.
It will be annoying for non-cheater, like me, who wrote all the codes in an ide and then copy paste it to codeforces
right, but you can always justify that you aren’t cheating by not doing suspicious things when the contest is running. Codeforces probably already uses the MOSS ai code detection thing, so the real structure is what matters.
Competition is just motivation, but brain is something you'll always need, but I think it would be good if codeforces could replace the current rating system with, rating system like number of solved problems + number of attempts vs problem rating, and problem rating can be determined by trusted people like setters/testers.
Aren't you a cheater yourself?
348368369 348340063 348302256 348310988 346693314
He is already marked as a cheater :fire:
Yeah I am marked a cheater and I did cheat, I am not hiding anything here. But I also am working day and night to become a better version of myself and working on every contest to become a better problem solver. I am not good person, I was a cheater but I have turned a new leaf and I am working to atone for my sins. The fact I am commenting here is not to gloat or to act like I know it all, I am doing this to be better and regain the trust I lost in my community.
You should create a new account instead of trying to justify this. Then you will have a clean slate and can work on your problem solving.
Ok I will do that. I think that is best.
MikeMirzayanov
improve your problem solving that's it offline contests (ICPC , Meta hacker cup, etc ) should be your goal the cheaters will struggle here
That's what I think as well. I try to not get demotivated by results in contests, rather I focus on results from face-to-face contests I enter.
Guys play it like a game for fun. Why run after ratings? I do CP because I don’t enjoy anything else in my life. I love coding and solving problems. Ignore the cheaters — the Codeforces team works behind the scenes to catch them. But it isn’t easy to detect them. What we can do is simply ignore those bastards.