Let's talk about competitive programming
- What's your codeforces rank? - Why do / don't you do competitive programming? - What's your take on the impact AI models like AlphaGo will have on the sport?
wolfenstine
Stealth
6 months ago
I like how the OP has clarified the question is from a sports perspective! These days I see so many people asking for it as an interview prep thing.
Competitive programming from an interview perspective can just be done in 30 days if not less. Anything ahead of that is pure sport. Don't expect it to take you ahead in life.
I totally agree with you! It seems like some people really overhype the power of cp in the interview process. However, I must say that the original poster has a valid point. What do you think about it?
DiracDelta
Student
6 months ago
I'll go first,
1. I'm currently ~ 1800
2. I've always enjoyed anything that has a ranking system. I'll have to introspect more to understand why but for now competitive programming scratches that itch for me.
3. I'm sceptical about Google's claims about their model's efficacy. They claim their model AlphaGo2 is as good as an 1700 player on codeforces, but many of their submissions were criticized for being way too similar to the editorial solutions and hence showed signs of leakage. But if such a model does come up then the sport would die or move to heuristic problems like hash code and the whole interview system would go through a big paradigm shift.
Heya
Student
6 months ago
What do you think future of interview looks like?
I am not into competitive programming.
DiracDelta
Student
6 months ago
I think we might see heuristic problems coming up as interview questions, since they're open NP complete problems, the candidate's answer to them can show their creativity and core problem solving capabilities
1. Not current but career best in 1400s.
2. Stopped a few years ago. I felt it is too addictive and time consuming. I found better marginal utility doing other software engineering stuff - namely building interesting projects across tech stacks.
3. I guess similar to the impact Stockfish engine had on chess. People who are truly interested will continue playing. We'll see new formats come up (like 1 vs. 1 or 4 vs. 4 for eg:) where using AI models would be difficult due to greater scrutiny.
DiracDelta
Student
6 months ago
Thanks for your response, I think it would be more difficult judging if the person has used AI or not in their solution even in a 1 v 1 setting in comparison to chess. Since in chess you can detect unsportsmanship behaviour based on the rates of excellent moves played , since there are plenty of steps before winning which can be scrutinized. But in case of competitive programming, the only thing you can scrutinize is the code and perhaps time of submission. Both which can be modified to be made to look "natural".
Discover More
Curated from across