Yesterday, Furioso_Slient authored a blog article, in which he said that his submissions in Codeforces Round 1063 (Div. 2) were incorrectly skipped, and insisted he didn't cheat in the contest. I disagreed, and posted a comment to that blog post, showing some evidence that his submissions were indeed suspected to be AI-generated.
Today, he apparently made a reply to my comment, but I couldn't read it, because he deleted his original blog post for some reason. Now my comment is gone as the parent blog post is deleted. In this blog post, I recover what I wrote in my comment earlier with massive rewrites and additions, and try to explain why this person is a cheater. I usually don't write this kind of blogs, but I'm doing this time because (1) I don't want to waste my effort to investigate his submissions, and (2) I want to make him responsible for his own actions.
Evidence
I want to highlight his in-contest submissions to 2163B - Siga ta Kymata. He made six submissions in total; here are links to each of them in the chronological order:
- A: 348317920, submitted at 18:29:22, WA2.
- B: 348320121, submitted at 18:32:54, WA2.
- C: 348324641, submitted at 18:40:18, WA2.
- D: 348340077, submitted at 19:07:48, WA2.
- E: 348341711, submitted at 19:11:01, WA2.
- F: 348352209, submitted at 19:32:37, WA2.
I want to encourage you to look at the differences between A and B, and the differences between D and E (use the "Compare" button in the submission page). Those two pairs of submissions are only about 4 minutes apart each, yet there are surprisingly many differences between them (each pair).
Regarding A->B. This should be very clear; A is only 24 LOC but B suddenly becomes 161 LOC. It's very surprising if a human could write this amount of code in this short time period, in response to the WA verdict.
Regarding D->E. This is interesting in a different way. If you look at the diffs, you can see a lot of code churns that aren't fundamentally necessary. Examples include:
- A space after the keywords
if,for, andwhileis consistently deleted. - Variable names are renamed for no good reason —
get_min->qm,get_max->qx,L->l,R->r, andl->len. - Prefix
++operators are rewritten to postfix ones for no reason. - Arguments passed to
ios::sync_with_stdio()andcin.tie()are changed tofalseandnullptr, because they are cool, I suppose? - When a vector of pairs is printed, a structured binding is used for no reason:
- for (auto& pr : o) {
- cout << pr.first << ' ' << pr.second << '\n';
+ for(auto [l, r] : o) {
+ cout << l << " " << r << '\n';
At this point, it should be clear: a sane human contestant doesn't make such code changes during a contest. It's undeniably clear that he used AI for writing his submissions. The admins skipping his submissions was a right call.
That's it. Opinions welcome.








https://cf-cheater-database.vercel.app/reportCheaters
It's so obvious that even my great grandmother can see that he is cheating like come on bro.
Iron_china ballast666
still not banned nor marked as cheater in the database also sevenseas
I respect your exploration of my code and the inquiries you have raised. I am also willing to provide you with relevant evidence:
Some code is just a facelift, having appeared in previous related problems, so I directly imported the original code and made simple modifications
I admit this. I used AI for compilation and debugging, which is in line with Mike's usage guidelines for AI. Therefore, the changes and variations you perceive as too frequent are in fact Mike's original words:
Code Completion Tools (e.g., Copilot): AI-based code completion systems may be used, but only for syntax and minor coding suggestions. Using them to generate the core logic or algorithms for solving problems is not permitted.
At this point, the modification of AI meets the requirements, as it did not involve modifying the core logic.
All core logics and algorithms are not formed by AI. The AI available in China cannot and is incapable of achieving this.
3.There is a problem with the code submission for question B:
First, I successfully submitted the solution for B (161 LOC). However, I misread the problem and thought it could be solved directly by giving 1-n. Obviously, this was incorrect, so I debugged the 161 LOC code and spent the rest of the time submitting it.
If you have any questions, feel free to ask. We also welcome additional perspectives and opinions.
The reason for deleting the post was that I thought the administrator had already handled it。
Since you all think I'm cheating, then fine, I'll choose to stop using this account and start a new one. Let the truth speak for itself. My abilities don't need to be proven by you people.
Dear:MikeMirzayanov KAN
If you still think my reasons are unreasonable and unreasonable, you can choose to ban my account
Similarly, if you think my reasons are correct and feasible,
I think this farce should end here。
How can you prove that the core logic and algorithms are yours?
You can try using AI systems that are allowed in China and see if they can pass the same questions as I did. You will find that Chinese AI systems are completely incapable of solving the questions that I can pass.
Currently known AI that can be used: Doubao, DeepSeek, Xinghuo, ERNIE Bot, Keen I don't know about the rest
But I can tell you clearly that I used Doubao to compile and debug the code several times when writing B. And it did not involve the core logic
I'm not in China so i don't know this. But next time you should really compile the code locally like I don't know why you use AI to compile the code it's just slower.
Because there are too many errors
If I go to fix the grammar problems one by one and RE
There will be no time to deal with the remaining issues.
You keep harping on about 'AI models currently available in China,' but let’s be real: most Chinese competitive programmers have no trouble accessing international models like Gemini, which can breeze through Div.2 problems.
I find it impossible to believe that a Master-rated contestant like yourself wouldn't realize this.
The most critical point—as the blog already highlighted—is that you consistently make modifications in places that don't need them, following a logic that is completely alien to human thought. You haven't offered a single explanation for this.
You are the real clown here. It's time for you to end this circus.
So I am not use the AI tool again.
I think you also knew that.
If use AI in a contest may lead to the question.
I just not to use it again.
that is the best way to solve it.
Hey, You all should realize that, this body who you accuse on cheating doesn't cheat i have some real reasons. Let's look at them step-by-step.
1) Firstly, if sb cheat in the contest and has some codes like that (348320121)!. Just bcs of if they cheat on the contest, their codes are just should be like that (if i == 0 // it is an operation for i)(just for example from me) it consists, comments, some unnecassariy things that is make code mor longer or sth like that.
2) BUT WHY THIS CONTESTANT IS NOT CHEATING???
BCS if he/she cheats and is really proffesional on cheating like real human and cp coders this man should write the code very very late and submit it so far BCS if he/she takes code from Genarative AI they should modify some action on code to become it realistic for example(Changing, variable names, or just write code in his/her style) And as u kniw it modifying takes so time and also submission queue is also takes so ling time. That's WHY THIS CONTESTANT IS NOT CHEATING FOR MY OPINION!
3) Also codes and as you said variable names that u think is replaced by AI, can be sth in his/her own language and brief form f this.
4) AND WHY I'M TRYIN TO SAVE HIM/HER?
It is definetly not about i know him or h/she is my firend etc//. That's why i think YOU ALL SHOULD NOT ACCUSE SB ON UNREALISTIC CHEATING. AND IT IS THE REAL REASON I AM IN SIDE WITH HIS/HER(as you said "cheater")
AGAIN IF YOU HAVE SOME WORDS TO SAY ME PLS SAY ME IN THESE COMMENTS OR DIRECTLY MESSAGE TO ME PLS!
Firstly, I must admit that I used AI-assisted programming in that competition. (Because my algorithm ability was not strong at that time, and there were persistent issues with Problem B, I asked AI to help review the issues and provide a solution to correct the code RE)
This is a major issue for me. Although AI did not directly participate in logical modifications, it violated the competition rules.
Thank you for your understanding, but making a mistake is still making a mistake.
I will ensure that I do not use AI in subsequent competitions. Because my current abilities are sufficient to tackle problems that I was unable to handle in the past.
For me YOu are not cheater whatever all they people say. I can provide and dedicate them on this. Furthermore as you said it is not forbidden to use AI on compile or make errors correct. Also did they bann ur account? they complain you?
They didn't ban any accounts.
But once a code is modified by AI, it is unfair to others no matter what.
So even if the rules allow it, don't do it.
A large number of people, including me, have already suffered.
Finally, I hope the matter of whether to use AI in that game can come to a close.
Three months ago, I indeed lacked sufficient capability, which led me to use AI to modify the code It should also be clarified that I do not possess the capability to use foreign AI. Firstly, those AI require foreign phone number/email verification, which I do not have. Secondly, most foreign AI do not allow Chinese users.
Then, three months have passed since this incident, and I have reflected on the issue. I think it's meaningless to keep dwelling on this matter. (Actually, at my current level, I don't need that game's score to enhance my reputation.)
We will all make progress, and we will all make mistakes.
I recognized my mistake and promptly took corrective action to prevent further damage.
Meanwhile, I hope everyone can forgive me and allow me to continue to be active in this community.
Thank you.
Just consider this as a joke or sth is really uncrucial for u